Files
ragflow/internal/cli/user_parser.go

3885 lines
87 KiB
Go
Raw Normal View History

package cli
import (
"fmt"
"strconv"
"strings"
)
// Command parsers
func (p *Parser) parseLogout() (*Command, error) {
cmd := NewCommand("logout")
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseLoginUser() (*Command, error) {
cmd := NewCommand("login_user")
p.nextToken() // consume LOGIN
if p.curToken.Type != TokenUser {
return nil, fmt.Errorf("expected USER after LOGIN")
}
p.nextToken()
email, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["email"] = email
p.nextToken()
// Optional: PASSWORD 'password'
if p.curToken.Type == TokenPassword {
p.nextToken()
password, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["password"] = password
p.nextToken()
}
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parsePingServer() (*Command, error) {
cmd := NewCommand("ping")
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseRegisterCommand() (*Command, error) {
cmd := NewCommand("register_user")
if err := p.expectPeek(TokenUser); err != nil {
return nil, err
}
p.nextToken() // consume USER
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["user_name"] = userName
p.nextToken() // consume Email
if p.curToken.Type != TokenAs {
return nil, fmt.Errorf("expected AS")
}
p.nextToken() // consume AS
nickname, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["nickname"] = nickname
p.nextToken() // consume nickname
if p.curToken.Type != TokenPassword {
return nil, fmt.Errorf("expected PASSWORD")
}
p.nextToken() // consume PASSWORD
password, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["password"] = password
p.nextToken() // consume 'password'
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListCommand() (*Command, error) {
p.nextToken() // consume LIST
switch p.curToken.Type {
case TokenVars:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_variables"), nil
case TokenConfigs:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_configs"), nil
case TokenEnvs:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_environments"), nil
case TokenDatasets:
return p.parseListDatasets()
case TokenDocuments:
return p.parseListDatasetDocuments()
case TokenMetadata:
return p.parseListMetadata()
case TokenAgents:
return p.parseListAgents()
case TokenTokens:
return p.parseListTokens()
case TokenModel:
return p.parseListModelProviders()
case TokenSupported:
return p.parseListModelsOfProvider()
case TokenModels:
return p.parseListModelsOfProvider()
case TokenProviders:
return p.parseListProviders()
case TokenInstances:
return p.parseListInstances()
case TokenDefault:
return p.parseListDefaultModels()
case TokenAvailable:
return p.parseCommonListProviders()
case TokenChats:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_user_chats"), nil
case TokenFiles:
return p.parseListFiles()
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
case TokenQuotedString:
return p.parseListQuotedStringCommand()
default:
return nil, fmt.Errorf("unknown LIST target: %s", p.curToken.Value)
}
}
func (p *Parser) parseListDatasets() (*Command, error) {
cmd := NewCommand("list_datasets")
p.nextToken() // consume DATASETS
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListDatasetDocuments() (*Command, error) {
p.nextToken() // consume DOCUMENTS
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
datasetID, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("list_dataset_documents")
cmd.Params["dataset_id"] = datasetID
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListMetadata() (*Command, error) {
p.nextToken() // consume METADATA
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF after METADATA")
}
p.nextToken()
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET after OF")
}
p.nextToken()
// Parse dataset names (space-separated)
var datasetNames []string
for {
name, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset name: %w", err)
}
datasetNames = append(datasetNames, name)
p.nextToken()
// Stop at semicolon or non-quoted (dataset name must be quoted)
if p.curToken.Type == TokenSemicolon {
break
}
// If next token is not a quoted string, stop parsing dataset names
if p.curToken.Type != TokenQuotedString {
break
}
}
cmd := NewCommand("list_metadata")
cmd.Params["dataset_names"] = datasetNames
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListAgents() (*Command, error) {
p.nextToken() // consume AGENTS
if p.curToken.Type == TokenSemicolon {
return NewCommand("list_user_agents"), nil
}
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF")
}
p.nextToken()
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("list_agents")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListTokens() (*Command, error) {
p.nextToken() // consume TOKENS
cmd := NewCommand("list_tokens")
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseListModelProviders() (*Command, error) {
p.nextToken() // consume MODEL
if p.curToken.Type != TokenProviders {
return nil, fmt.Errorf("expected PROVIDERS")
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_user_model_providers"), nil
}
// parseListProviders parses LIST PROVIDERS command
func (p *Parser) parseListProviders() (*Command, error) {
p.nextToken() // consume PROVIDERS
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_providers"), nil
}
func (p *Parser) parseListDefaultModels() (*Command, error) {
p.nextToken() // consume DEFAULT
if p.curToken.Type != TokenModels {
return nil, fmt.Errorf("expected MODELS")
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("list_user_default_models"), nil
}
func (p *Parser) parseListFiles() (*Command, error) {
p.nextToken() // consume FILES
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF")
}
p.nextToken()
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET")
}
p.nextToken()
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("list_user_dataset_files")
cmd.Params["dataset_name"] = datasetName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
func (p *Parser) parseListQuotedStringCommand() (*Command, error) {
str, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken() // consume str
switch p.curToken.Type {
case TokenTasks:
p.nextToken() // consume TASKS
cmd := NewCommand("list_tasks_user_command")
cmd.Params["composite_instance_name"] = str
return cmd, nil
default:
return nil, fmt.Errorf("unknown command: %s", str)
}
}
func (p *Parser) parseShowQuotedStringCommand() (*Command, error) {
str, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken() // consume str
switch p.curToken.Type {
case TokenTask:
p.nextToken() // consume TASK
var taskID string
taskID, err = p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected string: %w", err)
}
p.nextToken()
cmd := NewCommand("show_task_user_command")
cmd.Params["task_id"] = taskID
cmd.Params["composite_instance_name"] = str
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
default:
return nil, fmt.Errorf("unknown command: %s", str)
}
}
func (p *Parser) parseShowCommand() (*Command, error) {
p.nextToken() // consume SHOW
switch p.curToken.Type {
case TokenVersion:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("show_version"), nil
case TokenToken:
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("show_token"), nil
case TokenCurrent:
p.nextToken()
if p.curToken.Type == TokenUser {
p.nextToken()
// Semicolon is optional for SHOW CURRENT USER
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("show_current_user"), nil
} else if p.curToken.Type == TokenModel {
p.nextToken()
// Semicolon is optional for SHOW CURRENT MODEL
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("show_current_model"), nil
}
return nil, fmt.Errorf("expected USER or MODEL after CURRENT")
case TokenUser:
return p.parseShowUser()
case TokenRole:
return p.parseShowRole()
case TokenVar:
return p.parseShowVariable()
case TokenService:
return p.parseShowService()
case TokenProvider:
return p.parseShowProvider()
case TokenModel:
return p.parseShowModel()
case TokenInstance:
return p.parseShowInstance()
case TokenBalance:
return p.parseShowBalance()
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
case TokenTask:
return p.parseShowTask()
case TokenQuotedString:
return p.parseShowQuotedStringCommand()
default:
return nil, fmt.Errorf("unknown SHOW target: %s", p.curToken.Value)
}
}
func (p *Parser) parseShowUser() (*Command, error) {
p.nextToken() // consume USER
// Check for PERMISSION
if p.curToken.Type == TokenPermission {
p.nextToken()
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("show_user_permission")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("show_user")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseShowRole() (*Command, error) {
p.nextToken() // consume ROLE
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("show_role")
cmd.Params["role_name"] = roleName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseShowVariable() (*Command, error) {
p.nextToken() // consume VAR
varName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("show_variable")
cmd.Params["var_name"] = varName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseShowService() (*Command, error) {
p.nextToken() // consume SERVICE
serviceNum, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd := NewCommand("show_service")
cmd.Params["number"] = serviceNum
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseShowModel() (*Command, error) {
p.nextToken() // consume model
modelName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected model name: %w", err)
}
cmd := NewCommand("show_model")
cmd.Params["model_name"] = modelName
p.nextToken() // consume model_name
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken() // consume from
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
cmd.Params["provider_name"] = providerName
p.nextToken() // consume provider name
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseShowProvider parses SHOW PROVIDER <name> command
func (p *Parser) parseShowProvider() (*Command, error) {
p.nextToken() // consume PROVIDER
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
cmd := NewCommand("show_provider")
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseCreateCommand() (*Command, error) {
p.nextToken() // consume CREATE
switch p.curToken.Type {
case TokenUser:
return p.parseCreateUser()
case TokenRole:
return p.parseCreateRole()
case TokenModel:
return p.parseCreateModelProvider()
case TokenDataset:
return p.parseCreateDataset()
case TokenChat:
return p.parseCreateChat()
case TokenToken:
return p.parseCreateToken()
case TokenChunkStore:
return p.parseCreateChunkStore()
case TokenMetadata:
return p.parseCreateMetadataStore()
case TokenProvider:
return p.parseCreateProviderInstance()
default:
return nil, fmt.Errorf("unknown CREATE target: %s", p.curToken.Value)
}
}
func (p *Parser) parseAddCommand() (*Command, error) {
p.nextToken() // consume ADD
switch p.curToken.Type {
case TokenProvider:
return p.parseAddProvider()
case TokenModel:
return p.parseAddModel()
default:
return nil, fmt.Errorf("unknown ADD target: %s", p.curToken.Value)
}
}
func (p *Parser) parseCreateToken() (*Command, error) {
p.nextToken() // consume TOKEN
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("create_token"), nil
}
// Internal CLI for GO
// parseCreateChunkStore parses: CREATE CHUNK STORE for Dataset 'name' VECTOR SIZE N
func (p *Parser) parseCreateChunkStore() (*Command, error) {
p.nextToken() // consume CHUNK STORE compound token
// Expect FOR
if p.curToken.Type != TokenFor {
return nil, fmt.Errorf("expected FOR after CHUNK STORE, got %s", p.curToken.Value)
}
p.nextToken()
// Expect Dataset
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected Dataset after FOR, got %s", p.curToken.Value)
}
p.nextToken()
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset name, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type != TokenVector {
return nil, fmt.Errorf("expected VECTOR after dataset name, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type != TokenSize {
return nil, fmt.Errorf("expected SIZE after VECTOR, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type != TokenInteger {
return nil, fmt.Errorf("expected vector size number, got %s", p.curToken.Value)
}
vectorSize, err := strconv.Atoi(p.curToken.Value)
if err != nil {
return nil, fmt.Errorf("invalid vector size: %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("create_chunk_store")
cmd.Params["dataset_name"] = datasetName
cmd.Params["vector_size"] = vectorSize
return cmd, nil
}
// Internal CLI for GO
// parseCreateMetadataStore parses: CREATE METADATA STORE
func (p *Parser) parseCreateMetadataStore() (*Command, error) {
// CREATE METADATA STORE
p.nextToken() // consume METADATA
if p.curToken.Type != TokenStore {
return nil, fmt.Errorf("expected STORE after METADATA, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("create_metadata_store"), nil
}
func (p *Parser) parseCreateUser() (*Command, error) {
p.nextToken() // consume USER
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
password, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("create_user")
cmd.Params["user_name"] = userName
cmd.Params["password"] = password
cmd.Params["role"] = "user"
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseCreateRole() (*Command, error) {
p.nextToken() // consume ROLE
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("create_role")
cmd.Params["role_name"] = roleName
p.nextToken()
if p.curToken.Type == TokenDescription {
p.nextToken()
description, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["description"] = description
p.nextToken()
}
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseCreateModelProvider() (*Command, error) {
p.nextToken() // consume MODEL
if p.curToken.Type != TokenProvider {
return nil, fmt.Errorf("expected PROVIDER")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
providerKey, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("create_model_provider")
cmd.Params["provider_name"] = providerName
cmd.Params["provider_key"] = providerKey
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseAddProvider parses ADD PROVIDER commands
// ADD PROVIDER <name>
// ADD PROVIDER <name> <api_key>
func (p *Parser) parseAddProvider() (*Command, error) {
p.nextToken() // consume PROVIDER
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
cmd := NewCommand("add_provider")
cmd.Params["provider_name"] = providerName
p.nextToken()
// Check if api_key is provided (optional)
if p.curToken.Type == TokenQuotedString {
apiKey, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected api key: %w", err)
}
cmd.Params["api_key"] = apiKey
p.nextToken()
}
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// syntax: add model 'xxx' to provider 'vllm' instance 'test' with tokens 1024 chat think vision;
func (p *Parser) parseAddModel() (*Command, error) {
p.nextToken() // consume MODEL
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected model name")
}
modelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken() // consume model name
if p.curToken.Type != TokenTo {
return nil, fmt.Errorf("expected TO")
}
p.nextToken()
if p.curToken.Type != TokenProvider {
return nil, fmt.Errorf("expected PROVIDER")
}
p.nextToken()
// provider name
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected provider name")
}
providerName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenInstance {
return nil, fmt.Errorf("expected INSTANCE")
}
p.nextToken()
// instance name
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected provider name")
}
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
var modelTypes []string
var supportThink *bool = nil
maxTokens := 0
if p.curToken.Type == TokenWith {
p.nextToken() // pass WITH
optionsLoop:
for {
switch p.curToken.Type {
case TokenThink:
if supportThink != nil {
return nil, fmt.Errorf("think model is already set")
}
supportThink = new(bool)
p.nextToken()
*supportThink = true
case TokenVision:
p.nextToken()
modelTypes = append(modelTypes, "vision")
case TokenChat:
p.nextToken()
modelTypes = append(modelTypes, "chat")
case TokenEmbedding:
p.nextToken()
modelTypes = append(modelTypes, "embedding")
case TokenRerank:
p.nextToken()
modelTypes = append(modelTypes, "rerank")
case TokenOCR:
p.nextToken()
modelTypes = append(modelTypes, "ocr")
Go: implement provider: MinerU_Local (#15051) ### What problem does this PR solve? 1. Add model types when add model --- ``` RAGFlow(user)> add model 'pipeline' to provider 'mineru_local' instance 'test' with tokens 131072 doc_parse; SUCCESS ``` 2. implement provider: MinerU_Local --- **Verified from CLI** ``` RAGFlow(user)> parse with 'pipeline@test@mineru_local' file './internal/test.pdf' +--------------------------------------+ | task_id | +--------------------------------------+ | c7260e31-b6e2-4b36-955d-e9c60510c669 | +--------------------------------------+ RAGFlow(user)> show 'test@mineru_local' task 'c7260e31-b6e2-4b36-955d-e9c60510c669' +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation Bingxin Ke Anton Obukhov Shengyu Huang Nando Metzger Rodrigo Caye Daudt Konrad Schindler Photogrammetry and Remote Sensing, ETH Zurich ¨ ![](images/ae256101419715b544d13722... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-20 19:21:57 +08:00
case TokenDocParse:
p.nextToken()
modelTypes = append(modelTypes, "doc_parse")
case TokenTTS:
p.nextToken()
modelTypes = append(modelTypes, "tts")
case TokenASR:
p.nextToken()
modelTypes = append(modelTypes, "asr")
case TokenTokens:
p.nextToken() // pass TOKENS
if maxTokens != 0 {
return nil, fmt.Errorf("max tokens is already given %d", maxTokens)
}
if p.curToken.Type != TokenInteger {
return nil, fmt.Errorf("expected integer")
}
maxTokens, err = p.parseNumber()
if err != nil {
return nil, err
}
p.nextToken() // consume
case TokenSemicolon:
p.nextToken()
break optionsLoop // done
default:
// No more options to process
break optionsLoop
}
}
}
cmd := NewCommand("add_custom_model")
cmd.Params["model_name"] = modelName
cmd.Params["model_types"] = modelTypes
cmd.Params["provider_name"] = providerName
cmd.Params["instance_name"] = instanceName
if supportThink != nil {
cmd.Params["support_think"] = *supportThink
}
cmd.Params["max_tokens"] = maxTokens
return cmd, nil
}
func (p *Parser) parseCreateDataset() (*Command, error) {
p.nextToken() // consume DATASET
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenWith {
return nil, fmt.Errorf("expected WITH")
}
p.nextToken()
if p.curToken.Type != TokenEmbedding {
return nil, fmt.Errorf("expected EMBEDDING")
}
p.nextToken()
embedding, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("create_user_dataset")
cmd.Params["dataset_name"] = datasetName
cmd.Params["embedding"] = embedding
if p.curToken.Type == TokenParser {
p.nextToken()
parserType, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["parser_type"] = parserType
p.nextToken()
} else if p.curToken.Type == TokenPipeline {
p.nextToken()
pipeline, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["pipeline"] = pipeline
p.nextToken()
} else {
return nil, fmt.Errorf("expected PARSER or PIPELINE")
}
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseCreateChat() (*Command, error) {
p.nextToken() // consume CHAT
chatName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("create_user_chat")
cmd.Params["chat_name"] = chatName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseDropCommand() (*Command, error) {
p.nextToken() // consume DROP
switch p.curToken.Type {
case TokenUser:
return p.parseDropUser()
case TokenRole:
return p.parseDropRole()
case TokenDataset:
return p.parseDropDataset()
case TokenChat:
return p.parseDropChat()
case TokenToken:
return p.parseDropToken()
case TokenChunkStore:
return p.parseDropChunkStore()
case TokenMetadata:
return p.parseDropMetadataStore()
case TokenInstance:
return p.parseDropInstance()
case TokenModel:
return p.parseDropInstanceModel()
default:
return nil, fmt.Errorf("unknown DROP target: %s", p.curToken.Value)
}
}
func (p *Parser) parseDeleteCommand() (*Command, error) {
p.nextToken() // consume DELETE
switch p.curToken.Type {
case TokenProvider:
return p.parseDeleteProvider()
case TokenMetadata:
return p.parseDeleteMeta()
default:
return nil, fmt.Errorf("unknown DELETE target: %s", p.curToken.Value)
}
}
func (p *Parser) parseRemoveCommand() (*Command, error) {
p.nextToken() // consume RM
switch p.curToken.Type {
case TokenTag:
return p.parseRemoveTags()
case TokenChunks, TokenAll:
return p.parseRemoveChunk()
default:
return nil, fmt.Errorf("unknown REMOVE target: %s", p.curToken.Value)
}
}
func (p *Parser) parseDropToken() (*Command, error) {
p.nextToken() // consume TOKEN
tokenValue, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF")
}
p.nextToken()
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("drop_token")
cmd.Params["token"] = tokenValue
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// Internal CLI for GO
// parseDropChunkStore parses: DROP CHUNK STORE for Dataset 'name'
func (p *Parser) parseDropChunkStore() (*Command, error) {
p.nextToken() // consume CHUNK STORE
// Expect FOR
if p.curToken.Type != TokenFor {
return nil, fmt.Errorf("expected FOR after CHUNK STORE, got %s", p.curToken.Value)
}
p.nextToken()
// Expect Dataset
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected Dataset after FOR, got %s", p.curToken.Value)
}
p.nextToken()
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset name, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("drop_chunk_store")
cmd.Params["dataset_name"] = datasetName
return cmd, nil
}
// parseDropMetadataStore parses: DROP METADATA STORE
func (p *Parser) parseDropMetadataStore() (*Command, error) {
// DROP METADATA STORE
p.nextToken() // consume METADATA
if p.curToken.Type != TokenStore {
return nil, fmt.Errorf("expected STORE after METADATA, got %s", p.curToken.Value)
}
p.nextToken()
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("drop_metadata_store")
return cmd, nil
}
func (p *Parser) parseDropUser() (*Command, error) {
p.nextToken() // consume USER
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("drop_user")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseDropRole() (*Command, error) {
p.nextToken() // consume ROLE
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("drop_role")
cmd.Params["role_name"] = roleName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseDeleteProvider parses DELETE PROVIDER <name> command
func (p *Parser) parseDeleteProvider() (*Command, error) {
p.nextToken() // consume PROVIDER
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
cmd := NewCommand("delete_provider")
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseDropDataset() (*Command, error) {
p.nextToken() // consume DATASET
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("drop_user_dataset")
cmd.Params["dataset_name"] = datasetName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseDropChat() (*Command, error) {
p.nextToken() // consume CHAT
chatName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("drop_user_chat")
cmd.Params["chat_name"] = chatName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseAlterCommand() (*Command, error) {
p.nextToken() // consume ALTER
switch p.curToken.Type {
case TokenUser:
return p.parseAlterUser()
case TokenRole:
return p.parseAlterRole()
case TokenProvider:
return p.parseAlterProvider()
case TokenInstance:
return p.parseAlterInstance()
default:
return nil, fmt.Errorf("unknown ALTER target: %s", p.curToken.Value)
}
}
func (p *Parser) parseAlterUser() (*Command, error) {
p.nextToken() // consume USER
if p.curToken.Type == TokenActive {
return p.parseActivateUser()
}
if p.curToken.Type == TokenPassword {
p.nextToken()
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
password, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("alter_user")
cmd.Params["user_name"] = userName
cmd.Params["password"] = password
p.nextToken()
// Semicolon is optional for SHOW TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenSet {
return nil, fmt.Errorf("expected SET")
}
p.nextToken()
if p.curToken.Type != TokenRole {
return nil, fmt.Errorf("expected ROLE")
}
p.nextToken()
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("alter_user_role")
cmd.Params["user_name"] = userName
cmd.Params["role_name"] = roleName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseActivateUser() (*Command, error) {
p.nextToken() // consume ACTIVE
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
// Accept 'on' or 'off' as identifier
status := p.curToken.Value
if status != "on" && status != "off" {
return nil, fmt.Errorf("expected 'on' or 'off', got %s", p.curToken.Value)
}
cmd := NewCommand("activate_user")
cmd.Params["user_name"] = userName
cmd.Params["activate_status"] = status
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseAlterRole() (*Command, error) {
p.nextToken() // consume ROLE
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenSet {
return nil, fmt.Errorf("expected SET")
}
p.nextToken()
if p.curToken.Type != TokenDescription {
return nil, fmt.Errorf("expected DESCRIPTION")
}
p.nextToken()
description, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("alter_role")
cmd.Params["role_name"] = roleName
cmd.Params["description"] = description
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseAlterProvider parses ALTER PROVIDER <name> NAME <new_name> command
func (p *Parser) parseAlterProvider() (*Command, error) {
p.nextToken() // consume PROVIDER
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenName {
return nil, fmt.Errorf("expected NAME")
}
p.nextToken()
newName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected new provider name: %w", err)
}
cmd := NewCommand("alter_provider")
cmd.Params["provider_name"] = providerName
cmd.Params["new_name"] = newName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseCreateProviderInstance parses CREATE PROVIDER <name> INSTANCE <instance_name> KEY <api_key> URL <base_url> REGION <region> command
// instance_name cannot be "default"
func (p *Parser) parseCreateProviderInstance() (*Command, error) {
p.nextToken() // consume PROVIDER
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenInstance {
return nil, fmt.Errorf("expected INSTANCE after provider name")
}
p.nextToken()
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenKey {
return nil, fmt.Errorf("expected KEY after instance name")
}
p.nextToken()
apiKey, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected API key: %w", err)
}
p.nextToken()
baseURL := ""
region := ""
optionsLoop:
for {
switch p.curToken.Type {
case TokenRegion:
p.nextToken()
region, err = p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected region: %w", err)
}
p.nextToken()
case TokenURL:
p.nextToken()
baseURL, err = p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected base URL: %w", err)
}
p.nextToken()
default:
break optionsLoop
}
}
cmd := NewCommand("create_provider_instance")
cmd.Params["provider_name"] = providerName
cmd.Params["instance_name"] = instanceName
cmd.Params["api_key"] = apiKey
if baseURL != "" {
// Only local model provider need to set URL
cmd.Params["base_url"] = baseURL
}
if region != "" {
cmd.Params["region"] = region
}
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseListInstances parses LIST INSTANCES FROM PROVIDER <name> command
func (p *Parser) parseListInstances() (*Command, error) {
p.nextToken() // consume INSTANCES
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
cmd := NewCommand("list_provider_instances")
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseShowInstance parses SHOW INSTANCE <name> FROM PROVIDER <name> command
func (p *Parser) parseShowInstance() (*Command, error) {
p.nextToken() // consume INSTANCE
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
cmd := NewCommand("show_provider_instance")
cmd.Params["instance_name"] = instanceName
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseShowBalance parses SHOW BALANCE FROM <provider_name> <instance_name>
func (p *Parser) parseShowBalance() (*Command, error) {
p.nextToken() // consume INSTANCE
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER")
}
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected instance name")
}
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
cmd := NewCommand("show_instance_balance")
cmd.Params["instance_name"] = instanceName
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
// parseShowTask parses SHOW TASK <task>
func (p *Parser) parseShowTask() (*Command, error) {
p.nextToken() // consume TASK
taskID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected string: %w", err)
}
p.nextToken()
cmd := NewCommand("show_task_user_command")
cmd.Params["task_id"] = taskID
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseAlterInstance parses ALTER INSTANCE <name> NAME <new_name> FROM PROVIDER <name> command
func (p *Parser) parseAlterInstance() (*Command, error) {
p.nextToken() // consume INSTANCE
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenName {
return nil, fmt.Errorf("expected NAME")
}
p.nextToken()
newName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected new instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
if p.curToken.Type != TokenProvider {
return nil, fmt.Errorf("expected PROVIDER after FROM")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
cmd := NewCommand("alter_provider_instance")
cmd.Params["instance_name"] = instanceName
cmd.Params["new_name"] = newName
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseDropInstance parses DROP INSTANCE <name> FROM PROVIDER <name> command
func (p *Parser) parseDropInstance() (*Command, error) {
p.nextToken() // consume INSTANCE
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
cmd := NewCommand("drop_provider_instance")
cmd.Params["instance_name"] = instanceName
cmd.Params["provider_name"] = providerName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseDropInstanceModel parses DROP MODEL <name> FROM <provider_name> <instance_name> command
// Only works for local deployed model
func (p *Parser) parseDropInstanceModel() (*Command, error) {
p.nextToken() // consume MODEL
modelName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name: %w", err)
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
providerName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected provider name after FROM PROVIDER: %w", err)
}
p.nextToken()
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected instance name after provider name: %w", err)
}
p.nextToken()
cmd := NewCommand("drop_instance_model")
cmd.Params["instance_name"] = instanceName
cmd.Params["provider_name"] = providerName
cmd.Params["model_name"] = modelName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseGrantCommand() (*Command, error) {
p.nextToken() // consume GRANT
if p.curToken.Type == TokenAdmin {
return p.parseGrantAdmin()
}
return p.parseGrantPermission()
}
func (p *Parser) parseGrantAdmin() (*Command, error) {
p.nextToken() // consume ADMIN
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("grant_admin")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseGrantPermission() (*Command, error) {
actions, err := p.parseIdentifierList()
if err != nil {
return nil, err
}
if p.curToken.Type != TokenOn {
return nil, fmt.Errorf("expected ON")
}
p.nextToken()
resource, err := p.parseIdentifier()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenTo {
return nil, fmt.Errorf("expected TO")
}
p.nextToken()
if p.curToken.Type != TokenRole {
return nil, fmt.Errorf("expected ROLE")
}
p.nextToken()
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("grant_permission")
cmd.Params["actions"] = actions
cmd.Params["resource"] = resource
cmd.Params["role_name"] = roleName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseRevokeCommand() (*Command, error) {
p.nextToken() // consume REVOKE
if p.curToken.Type == TokenAdmin {
return p.parseRevokeAdmin()
}
return p.parseRevokePermission()
}
func (p *Parser) parseRevokeAdmin() (*Command, error) {
p.nextToken() // consume ADMIN
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("revoke_admin")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseRevokePermission() (*Command, error) {
actions, err := p.parseIdentifierList()
if err != nil {
return nil, err
}
if p.curToken.Type != TokenOn {
return nil, fmt.Errorf("expected ON")
}
p.nextToken()
resource, err := p.parseIdentifier()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
if p.curToken.Type != TokenRole {
return nil, fmt.Errorf("expected ROLE")
}
p.nextToken()
roleName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
cmd := NewCommand("revoke_permission")
cmd.Params["actions"] = actions
cmd.Params["resource"] = resource
cmd.Params["role_name"] = roleName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseIdentifierList() ([]string, error) {
var list []string
ident, err := p.parseIdentifier()
if err != nil {
return nil, err
}
list = append(list, ident)
p.nextToken()
for p.curToken.Type == TokenComma {
p.nextToken()
ident, err := p.parseIdentifier()
if err != nil {
return nil, err
}
list = append(list, ident)
p.nextToken()
}
return list, nil
}
func (p *Parser) parseSetCommand() (*Command, error) {
p.nextToken() // consume SET
if p.curToken.Type == TokenVar {
return p.parseSetVariable()
}
if p.curToken.Type == TokenDefault {
return p.parseSetDefault()
}
if p.curToken.Type == TokenToken {
return p.parseSetToken()
}
if p.curToken.Type == TokenMetadata {
return p.parseSetMeta()
}
if p.curToken.Type == TokenLog {
return p.parseSetLog()
}
return nil, fmt.Errorf("unknown SET target: %s", p.curToken.Value)
}
func (p *Parser) parseSetVariable() (*Command, error) {
p.nextToken() // consume VAR
varName, err := p.parseIdentifier()
if err != nil {
return nil, err
}
p.nextToken()
varValue, err := p.parseVariableValue()
if err != nil {
return nil, err
}
cmd := NewCommand("set_variable")
cmd.Params["var_name"] = varName
cmd.Params["var_value"] = varValue
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseSetDefault() (*Command, error) {
p.nextToken() // consume DEFAULT
var modelType, compositeModelName string
var err error
switch p.curToken.Type {
case TokenChat:
modelType = "chat"
case TokenVision:
modelType = "vision"
case TokenEmbedding:
modelType = "embedding"
case TokenRerank:
modelType = "rerank"
case TokenASR:
modelType = "asr"
case TokenTTS:
modelType = "tts"
case TokenOCR:
modelType = "ocr"
default:
return nil, fmt.Errorf("unknown model type: %s", p.curToken.Value)
}
p.nextToken() // pass model type
if p.curToken.Type != TokenModel {
return nil, fmt.Errorf("expected MODEL")
}
p.nextToken() // pass MODEL
// Format: 'provider/instance/model' or just 'message'
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected quoted string with format provider/instance/model")
}
compositeModelName, err = p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("set_default_model")
cmd.Params["model_type"] = modelType
cmd.Params["composite_model_name"] = compositeModelName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseSetToken() (*Command, error) {
p.nextToken() // consume TOKEN
tokenValue, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("set_token")
cmd.Params["token"] = tokenValue
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseSetLog() (*Command, error) {
p.nextToken() // consume LOG
switch p.curToken.Type {
case TokenLevel:
return p.parseSetLogLevel()
default:
return nil, fmt.Errorf("unknown log target: %s", p.curToken.Value)
}
}
func (p *Parser) parseSetLogLevel() (*Command, error) {
p.nextToken() // consume LEVEL
cmd := NewCommand("set_log_level")
switch p.curToken.Type {
case TokenDebug:
cmd.Params["level"] = "debug"
case TokenInfo:
cmd.Params["level"] = "info"
case TokenWarn:
cmd.Params["level"] = "warn"
case TokenError:
cmd.Params["level"] = "error"
case TokenFatal:
cmd.Params["level"] = "fatal"
case TokenPanic:
cmd.Params["level"] = "panic"
default:
return nil, fmt.Errorf("unknown log target: %s", p.curToken.Value)
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseResetCommand() (*Command, error) {
p.nextToken() // consume RESET
if p.curToken.Type != TokenDefault {
return nil, fmt.Errorf("expected DEFAULT")
}
p.nextToken()
var modelType string
switch p.curToken.Type {
case TokenChat:
modelType = "chat"
case TokenVision:
modelType = "vision"
case TokenEmbedding:
modelType = "embedding"
case TokenRerank:
modelType = "rerank"
case TokenASR:
modelType = "asr"
case TokenTTS:
modelType = "tts"
case TokenOCR:
modelType = "ocr"
default:
return nil, fmt.Errorf("unknown model type: %s", p.curToken.Value)
}
cmd := NewCommand("reset_default_model")
cmd.Params["model_type"] = modelType
p.nextToken()
if p.curToken.Type != TokenModel {
return nil, fmt.Errorf("expected MODEL")
}
p.nextToken() // pass MODEL
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseGenerateCommand() (*Command, error) {
p.nextToken() // consume GENERATE
if p.curToken.Type != TokenToken {
return nil, fmt.Errorf("expected TOKEN")
}
p.nextToken()
if p.curToken.Type != TokenFor {
return nil, fmt.Errorf("expected FOR")
}
p.nextToken()
if p.curToken.Type != TokenUser {
return nil, fmt.Errorf("expected USER")
}
p.nextToken()
userName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("generate_token")
cmd.Params["user_name"] = userName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseImportCommand() (*Command, error) {
p.nextToken() // consume IMPORT
documentPaths, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenInto {
return nil, fmt.Errorf("expected INTO")
}
p.nextToken()
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET")
}
p.nextToken()
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("import_docs_into_dataset")
cmd.Params["document_paths"] = documentPaths
cmd.Params["dataset_name"] = datasetName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseInsertCommand parses INSERT command and dispatches to specific handler
func (p *Parser) parseInsertCommand() (*Command, error) {
p.nextToken() // consume INSERT
// Expect CHUNKS or METADATA
if p.curToken.Type == TokenChunks {
return p.parseInsertChunksFromFile()
}
if p.curToken.Type == TokenMetadata {
return p.parseInsertMetadataFromFile()
}
return nil, fmt.Errorf("expected CHUNKS or METADATA after INSERT, got %s", p.curToken.Value)
}
// Internal CLI for GO
// parseInsertChunksFromFile parses: INSERT CHUNKS FROM FILE "file_path"
func (p *Parser) parseInsertChunksFromFile() (*Command, error) {
p.nextToken() // consume CHUNKS
// Expect FROM
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM, got %s", p.curToken.Value)
}
p.nextToken()
// Expect FILE
if p.curToken.Type != TokenFile {
return nil, fmt.Errorf("expected FILE, got %s", p.curToken.Value)
}
p.nextToken()
// Get file path (quoted string)
filePath, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("insert_chunks_from_file")
cmd.Params["file_path"] = filePath
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// Internal CLI for GO
// parseInsertMetadataFromFile parses: INSERT METADATA FROM FILE "file_path"
func (p *Parser) parseInsertMetadataFromFile() (*Command, error) {
p.nextToken() // consume METADATA
// Expect FROM
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM, got %s", p.curToken.Value)
}
p.nextToken()
// Expect FILE
if p.curToken.Type != TokenFile {
return nil, fmt.Errorf("expected FILE, got %s", p.curToken.Value)
}
p.nextToken()
// Get file path (quoted string)
filePath, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("insert_metadata_from_file")
cmd.Params["file_path"] = filePath
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseSearchCommand() (*Command, error) {
p.nextToken() // consume SEARCH
var err error
var question string
if p.curToken.Type == TokenQuotedString {
question, err = p.parseQuotedString()
if err != nil {
return nil, err
}
} else if p.curToken.Type == TokenIdentifier {
question, err = p.parseIdentifier()
if err != nil {
return nil, err
}
} else {
return nil, fmt.Errorf("expected quoted string or identifier")
}
p.nextToken()
if p.curToken.Type == TokenOn {
p.nextToken() // skip on
if p.curToken.Type != TokenDatasets {
return nil, fmt.Errorf("expected DATASETS")
}
p.nextToken()
datasets, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("search_on_datasets")
cmd.Params["question"] = question
cmd.Params["datasets"] = datasets
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
cmd := NewCommand("ce_search")
cmd.Params["query"] = question
if p.curToken.Type == TokenEOF {
cmd.Params["path"] = "."
return cmd, nil
}
for p.curToken.Type != TokenEOF {
if p.curToken.Type == TokenDash {
p.nextToken() // skip dash
if p.curToken.Type != TokenIdentifier {
return nil, fmt.Errorf("expect identifier")
}
if strings.ToLower(p.curToken.Value) == "n" {
p.nextToken()
var err error
if p.curToken.Type != TokenInteger {
return nil, fmt.Errorf("expect number")
}
cmd.Params["number"], err = p.parseNumber()
if err != nil {
return nil, err
}
p.nextToken()
continue
}
//if strings.ToLower(p.curToken.Value) == "t" {
// p.nextToken()
// var err error
// if p.curToken.Type != TokenInteger {
// return nil, fmt.Errorf("expect number")
// }
// cmd.Params["threshold"], err = p.parseFloat()
// if err != nil {
// return nil, err
// }
// p.nextToken()
// continue
//}
return nil, fmt.Errorf("unknow parameter: %s", p.curToken.Value)
} else if p.curToken.Type == TokenIdentifier {
if cmd.Params["path"] == nil {
cmd.Params["path"] = p.curToken.Value
} else {
cmd.Params["path"] = fmt.Sprintf("%s%s", cmd.Params["path"], p.curToken.Value)
}
p.nextToken() // skip path
continue
} else if p.curToken.Type == TokenSlash {
if cmd.Params["path"] == nil {
cmd.Params["path"] = "/"
} else {
cmd.Params["path"] = fmt.Sprintf("%s/", cmd.Params["path"])
}
p.nextToken() // skip slash
if p.curToken.Type == TokenIdentifier {
cmd.Params["path"] = fmt.Sprintf("%s%s", cmd.Params["path"], p.curToken.Value)
p.nextToken()
}
continue
}
}
return cmd, nil
}
func (p *Parser) parseListModelsOfProvider() (*Command, error) {
if p.curToken.Type == TokenSupported {
// List supported models
p.nextToken()
cmd := NewCommand("list_supported_models")
if p.curToken.Type != TokenModels {
return nil, fmt.Errorf("expected MODELS")
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected quoted string for provider name")
}
firstName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected quoted string for instance name")
}
secondName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd.Params["provider_name"] = firstName
cmd.Params["instance_name"] = secondName
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
if p.curToken.Type != TokenModels {
return nil, fmt.Errorf("expected MODELS")
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
// Parse first quoted string (could be instance_name or provider_name)
firstName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
// Check if there's a second quoted string (provider_name)
// If so, format is: LIST MODELS FROM <instance_name> <provider_name>
// If not, format is: LIST MODELS FROM <provider_name>
if p.curToken.Type == TokenQuotedString {
// Two arguments: instance_name and provider_name
instanceName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
cmd := NewCommand("list_instance_models")
cmd.Params["instance_name"] = instanceName
cmd.Params["provider_name"] = firstName
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// Only one argument: provider_name
cmd := NewCommand("list_provider_models")
cmd.Params["provider_name"] = firstName
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseEnableCommand() (*Command, error) {
p.nextToken() // consume ENABLE
if p.curToken.Type != TokenModel {
return nil, fmt.Errorf("expected MODEL")
}
p.nextToken()
modelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
modelProvider, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
modelInstance, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("enable_model")
cmd.Params["model_name"] = modelName
cmd.Params["instance_name"] = modelInstance
cmd.Params["provider_name"] = modelProvider
return cmd, nil
}
func (p *Parser) parseDisableCommand() (*Command, error) {
p.nextToken() // consume DISABLE
if p.curToken.Type != TokenModel {
return nil, fmt.Errorf("expected MODEL")
}
p.nextToken()
modelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
modelProvider, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
modelInstance, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("disable_model")
cmd.Params["model_name"] = modelName
cmd.Params["instance_name"] = modelInstance
cmd.Params["provider_name"] = modelProvider
return cmd, nil
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
// CHAT 'model@instance@provider' 'hello world'
// CHAT WITH 'model@instance@provider' MESSAGE 'hello world' 'who are you' IMAGE 'url1' 'file0' VIDEO "url2.mov" "file1" FILE "url" "path file2" AUDIO "file.wav"
func (p *Parser) parseChatCommand() (*Command, error) {
p.nextToken() // consume CHAT
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
var err error
var compositeModelName string = ""
var messages []string
var images []string
var videos []string
var audios []string
var files []string
effort := "default"
verbosity := "low"
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
optionsLoop:
for {
switch p.curToken.Type {
case TokenWith:
p.nextToken()
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
// 'model@instance@provider'
if compositeModelName != "" {
return nil, fmt.Errorf("model name is already set")
}
compositeModelName, err = p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
case TokenMessage:
p.nextToken()
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
if len(messages) != 0 {
return nil, fmt.Errorf("message is already set")
}
messageLoop:
for {
if p.curToken.Type != TokenQuotedString {
break messageLoop
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
var message string
message, err = p.parseQuotedString()
if err != nil {
return nil, err
}
message = strings.TrimSpace(message)
messages = append(messages, message)
p.nextToken()
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
case TokenImage:
p.nextToken()
if len(images) != 0 {
return nil, fmt.Errorf("image is already set")
}
imageLoop:
for {
if p.curToken.Type != TokenQuotedString {
break imageLoop
}
var image string
image, err = p.parseQuotedString()
if err != nil {
return nil, err
}
images = append(images, image)
p.nextToken()
}
case TokenVideo:
p.nextToken()
if len(videos) != 0 {
return nil, fmt.Errorf("video is already set")
}
videoLoop:
for {
if p.curToken.Type != TokenQuotedString {
break videoLoop
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
var video string
video, err = p.parseQuotedString()
if err != nil {
return nil, err
}
videos = append(videos, video)
p.nextToken()
}
case TokenAudio:
p.nextToken()
if len(audios) != 0 {
return nil, fmt.Errorf("video is already set")
}
audioLoop:
for {
if p.curToken.Type != TokenQuotedString {
break audioLoop
}
var audio string
audio, err = p.parseQuotedString()
if err != nil {
return nil, err
}
audios = append(audios, audio)
p.nextToken()
}
case TokenFile:
p.nextToken()
if len(files) != 0 {
return nil, fmt.Errorf("video is already set")
}
fileLoop:
for {
if p.curToken.Type != TokenQuotedString {
break fileLoop
}
var file string
file, err = p.parseQuotedString()
if err != nil {
return nil, err
}
files = append(files, file)
p.nextToken()
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
case TokenEffort:
p.nextToken() // pass Effort
switch p.curToken.Type {
case TokenNone:
effort = "none"
case TokenMinimal:
effort = "minimal"
case TokenLow:
effort = "low"
case TokenMedium:
effort = "medium"
case TokenHigh:
effort = "high"
case TokenMax:
effort = "max"
default:
return nil, fmt.Errorf("invalid effort level")
}
p.nextToken()
break optionsLoop
case TokenVerbosity:
p.nextToken() // pass VERBOSITY
switch p.curToken.Type {
case TokenLow:
verbosity = "low"
case TokenMedium:
verbosity = "median"
case TokenHigh:
verbosity = "high"
default:
return nil, fmt.Errorf("invalid verbosity level")
}
p.nextToken()
break optionsLoop
case TokenSemicolon:
p.nextToken()
break optionsLoop // done
default:
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
// No more options to process
break optionsLoop
}
}
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
cmd := NewCommand("chat_to_model")
Go: CLI chat with text, image, video (#14573) ### What problem does this PR solve? ``` RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the pics talk about?' image 'https://cdn.bigmodel.cn/static/logo/register.png' 'https://cdn.bigmodel.cn/static/logo/api-key.png' Answer: The first picture shows a login/register modal with options for phone number login, account login, and WeChat QR code login, along with a prompt for new users to get a 20 million tokens experience package. The second picture displays the API keys management page of a platform, including a warning about API key security and a table listing existing API keys with details like creation time and usage history. Time: 31.600545 RAGFlow(user)> chat with 'glm-4.6v-flash@test@zhipu-ai' message 'What are the video talk about?' video 'https://cdn.bigmodel.cn/agent-demos/lark/113123.mov' Answer: Based on the sequence of frames provided, the video is a demonstration of a web search and navigation process. 1. The video starts with a blank Google search page. 2. The user types "智谱" (which is the Chinese name for the company Zhipu AI) into the search box. 3. The search is initiated and the page shows "About 0 results". 4. The search results load, showing information about Zhipu AI, including its website. 5. The user clicks on the main website link (www.zhipuai.cn). 6. The video ends by showing the homepage of Zhipu AI's website, titled "Z.ai GLM Large Model Open Platform". In summary, the video is about searching for the company "智谱" (Zhipu AI) on Google and then navigating to its official website. Time: 76.582520 ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-05 18:14:39 +08:00
cmd.Params["composite_model_name"] = compositeModelName
cmd.Params["messages"] = messages
cmd.Params["images"] = images
cmd.Params["videos"] = videos
cmd.Params["audios"] = audios
cmd.Params["files"] = files
cmd.Params["thinking"] = false
cmd.Params["stream"] = false
cmd.Params["effort"] = effort
cmd.Params["verbosity"] = verbosity
return cmd, nil
}
func (p *Parser) parseThinkCommand() (*Command, error) {
p.nextToken() // consume THINK
if p.curToken.Type != TokenChat {
return nil, fmt.Errorf("expected CHAT after THINK")
}
command, err := p.parseChatCommand()
if err != nil {
return nil, err
}
command.Params["thinking"] = true
return command, nil
}
func (p *Parser) parseStreamCommand() (*Command, error) {
p.nextToken() // consume STREAM
var command *Command
var err error
switch p.curToken.Type {
case TokenChat:
command, err = p.parseChatCommand()
if err != nil {
return nil, err
}
case TokenThink:
command, err = p.parseThinkCommand()
if err != nil {
return nil, err
}
case TokenASR:
command, err = p.parseASRCommand()
if err != nil {
return nil, err
}
case TokenTTS:
command, err = p.parseTTSCommand()
if err != nil {
return nil, err
}
default:
return nil, fmt.Errorf("expected CHAT, THINK, ASR, or TTS after STREAM")
}
command.Params["stream"] = true
return command, nil
}
func (p *Parser) parseEmbedCommand() (*Command, error) {
p.nextToken() // consume EMBED
if p.curToken.Type != TokenText {
return nil, fmt.Errorf("expected WITH after EMBED")
}
p.nextToken() // consume TEXT
var texts []string
textLoop:
for {
if p.curToken.Type != TokenQuotedString {
break textLoop
}
text, err := p.parseQuotedString()
if err != nil {
return nil, err
}
text = strings.TrimSpace(text)
texts = append(texts, text)
p.nextToken()
}
if p.curToken.Type != TokenWith {
return nil, fmt.Errorf("expected WITH after EMBED")
}
p.nextToken() // consume WITH
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenDimension {
return nil, fmt.Errorf("expected DIMENSION")
}
p.nextToken() // consume WITH
dimension, err := p.parseNumber()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("embed_user_text")
cmd.Params["composite_model_name"] = compositeModelName
cmd.Params["texts"] = texts
cmd.Params["dimension"] = dimension
return cmd, nil
}
func (p *Parser) parseRerankCommand() (*Command, error) {
p.nextToken() // consume RERANK
if p.curToken.Type != TokenQuery {
return nil, fmt.Errorf("expected WITH after EMBED")
}
p.nextToken() // consume QUERY
query, err := p.parseQuotedString()
if err != nil {
return nil, err
}
query = strings.TrimSpace(query)
p.nextToken() // consume query
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after query")
}
p.nextToken() // consume DOCUMENT
var documents []string
documentLoop:
for {
if p.curToken.Type != TokenQuotedString {
break documentLoop
}
var document string
document, err = p.parseQuotedString()
if err != nil {
return nil, err
}
document = strings.TrimSpace(document)
documents = append(documents, document)
p.nextToken()
}
if p.curToken.Type != TokenWith {
return nil, fmt.Errorf("expected WITH after EMBED")
}
p.nextToken() // consume WITH
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenTop {
return nil, fmt.Errorf("expected TOP after model")
}
p.nextToken()
topN, err := p.parseNumber()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("rarank_user_document")
cmd.Params["composite_model_name"] = compositeModelName
cmd.Params["query"] = query
cmd.Params["documents"] = documents
cmd.Params["top_n"] = topN
return cmd, nil
}
func (p *Parser) parseASRCommand() (*Command, error) {
p.nextToken() // consume ASR
if p.curToken.Type != TokenWith {
return nil, fmt.Errorf("expected WITH after ASR")
}
p.nextToken() // consume WITH
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
if p.curToken.Type != TokenAudio {
return nil, fmt.Errorf("expected AUDIO to ASR")
}
Go: implement TTS for fishaudio, openrouter and asr for fishaudio (#14926) ### What problem does this PR solve? This PR implement TTS for FishAudio and MiniMax provider and ASR for FishAudio **The following functionalities are now supported:** **FishAudio:** - [x] Text To Speech - [x] Stream Text To Speech - [x] Audio To Text **OpenRouter:** - [x] Text To Speech **Verified examples from the CLI:** ```plaintext **FishAudio** RAGFlow(user)> tts with 's1@test@fishaudio' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"reference_id": "90e65eaaf50e4470b8e6d43ee6afd7d5", "temperature": 0.7, "top_p": 0.7, "prosody": {"speed": 1, "volume": 0, "normalize_loudness": true}, "chunk_length": 300, "normalize": true, "sample_rate": 44100, "mp3_bitrate": 128, "latency": "normal", "max_new_tokens": 1024, "repetition_penalty": 1.2, "min_chunk_length": 50, "condition_on_previous_chunks": true, "early_stop_threshold": 1}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/s1_output.wav SUCCESS RAGFlow(user)> stream tts with 's1@test@fishaudio' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"reference_id": "90e65eaaf50e4470b8e6d43ee6afd7d5", "temperature": 0.7, "top_p": 0.7, "prosody": {"speed": 1, "volume": 0, "normalize_loudness": true}, "chunk_length": 300, "normalize": true, "sample_rate": 44100, "mp3_bitrate": 128, "latency": "normal", "max_new_tokens": 1024, "repetition_penalty": 1.2, "min_chunk_length": 50, "condition_on_previous_chunks": true, "early_stop_threshold": 1}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/s1_output.wav SUCCESS RAGFlow(user)> asr with 'transcribe-1@test@fishaudio' audio './internal/test.wav' param '{"language": "en", "ignore_timestamps": true}' +----------------------------------------------------------------------------------------------------------------------+ | text | +----------------------------------------------------------------------------------------------------------------------+ | The examination and testimony of the experts enabled the commission to conclude that five shots may have been fired. | +----------------------------------------------------------------------------------------------------------------------+ ``` ### Type of change - [x] Bug Fix (non-breaking change which fixes an issue) - [x] New Feature (non-breaking change which adds functionality) - [x] Refactoring
2026-05-14 18:58:00 +08:00
p.nextToken() // consume AUDIO
audioFile, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
Go: implement TTS for fishaudio, openrouter and asr for fishaudio (#14926) ### What problem does this PR solve? This PR implement TTS for FishAudio and MiniMax provider and ASR for FishAudio **The following functionalities are now supported:** **FishAudio:** - [x] Text To Speech - [x] Stream Text To Speech - [x] Audio To Text **OpenRouter:** - [x] Text To Speech **Verified examples from the CLI:** ```plaintext **FishAudio** RAGFlow(user)> tts with 's1@test@fishaudio' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"reference_id": "90e65eaaf50e4470b8e6d43ee6afd7d5", "temperature": 0.7, "top_p": 0.7, "prosody": {"speed": 1, "volume": 0, "normalize_loudness": true}, "chunk_length": 300, "normalize": true, "sample_rate": 44100, "mp3_bitrate": 128, "latency": "normal", "max_new_tokens": 1024, "repetition_penalty": 1.2, "min_chunk_length": 50, "condition_on_previous_chunks": true, "early_stop_threshold": 1}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/s1_output.wav SUCCESS RAGFlow(user)> stream tts with 's1@test@fishaudio' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"reference_id": "90e65eaaf50e4470b8e6d43ee6afd7d5", "temperature": 0.7, "top_p": 0.7, "prosody": {"speed": 1, "volume": 0, "normalize_loudness": true}, "chunk_length": 300, "normalize": true, "sample_rate": 44100, "mp3_bitrate": 128, "latency": "normal", "max_new_tokens": 1024, "repetition_penalty": 1.2, "min_chunk_length": 50, "condition_on_previous_chunks": true, "early_stop_threshold": 1}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/s1_output.wav SUCCESS RAGFlow(user)> asr with 'transcribe-1@test@fishaudio' audio './internal/test.wav' param '{"language": "en", "ignore_timestamps": true}' +----------------------------------------------------------------------------------------------------------------------+ | text | +----------------------------------------------------------------------------------------------------------------------+ | The examination and testimony of the experts enabled the commission to conclude that five shots may have been fired. | +----------------------------------------------------------------------------------------------------------------------+ ``` ### Type of change - [x] Bug Fix (non-breaking change which fixes an issue) - [x] New Feature (non-breaking change which adds functionality) - [x] Refactoring
2026-05-14 18:58:00 +08:00
cmd := NewCommand("asr_user_command")
cmd.Params["composite_model_name"] = compositeModelName
cmd.Params["audio_file"] = audioFile
for p.curToken.Type != TokenEOF && p.curToken.Type != TokenSemicolon {
switch p.curToken.Type {
case TokenParam:
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expect quoted string after 'param'")
}
cmd.Params["param_str"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
default:
return nil, fmt.Errorf("unexpected token in asr command: %s", p.curToken.Value)
}
}
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseTTSCommand() (*Command, error) {
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
p.nextToken()
cmd := NewCommand("tts_user_command")
if p.curToken.Type != TokenWith {
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
return nil, fmt.Errorf("expect 'with' after tts")
}
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
p.nextToken()
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
if p.curToken.Type != TokenQuotedString && p.curToken.Type != TokenIdentifier {
return nil, fmt.Errorf("expect model name after 'with'")
}
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
cmd.Params["composite_model_name"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
if p.curToken.Type != TokenText {
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
return nil, fmt.Errorf("expect 'text' parameter")
}
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
p.nextToken()
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expect quoted string after 'text'")
}
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
cmd.Params["text"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
Go: implement TTS for MiniMax provider and CLI testing for TTS (#14911) ### What problem does this PR solve? This PR implement TTS for MiniMax provider and CLI testing for TTS **The following functionalities are now supported:** **MiniMax:** - [x] Chat / Stream Chat - [x] Embedding - [x] Rerank - [x] Model listing - [x] Provider connection checking - [x] Text To Speech - [ ] OCRFile - [ ] ~~Audio To Text~~ - [ ] ~~Balance~~ **Verified examples from the CLI:** ```plaintext RAGFlow(user)> tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS RAGFlow(user)> stream tts with 'speech-2.8-hd@test@minimax' text 'He who desires but acts not, breeds pestilence.' play format 'wav' save './internal' param '{"voice_setting": {"voice_id": "English_radiant_girl", "speed": 1, "vol": 1, "pitch": 0}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "wav", "channel": 1}, "output_format": "hex"}' Saved to directory: /home/infiniflow/Documents/development/ragflow/internal/speech-2.8-hd_output.wav SUCCESS ``` Set `Play` to play audio in CLI Set `Save` `PATH_TO_SAVE` to save file Set `format` to save file in wav or mp3 Set `Param` align with official request body ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2026-05-14 13:19:31 +08:00
for p.curToken.Type != TokenEOF && p.curToken.Type != TokenSemicolon {
switch p.curToken.Type {
case TokenPlay:
p.nextToken()
cmd.Params["play"] = true
case TokenParam:
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expect quoted string after 'param'")
}
cmd.Params["param_str"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
p.nextToken()
case TokenSave:
p.nextToken()
if p.curToken.Type != TokenQuotedString && p.curToken.Type != TokenIdentifier {
return nil, fmt.Errorf("expect directory path after 'save'")
}
cmd.Params["save"] = true
cmd.Params["save_path"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
case TokenFormat:
p.nextToken()
if p.curToken.Type != TokenQuotedString && p.curToken.Type != TokenIdentifier {
return nil, fmt.Errorf("expect format string (e.g. 'wav') after 'format'")
}
cmd.Params["format"] = strings.Trim(p.curToken.Value, "\"'")
p.nextToken()
default:
return nil, fmt.Errorf("unexpected token: %s", p.curToken.Value)
}
}
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseOCRCommand() (*Command, error) {
p.nextToken() // consume OCR
if p.curToken.Type != TokenWith {
return nil, fmt.Errorf("expected WITH after OCR")
}
p.nextToken() // consume WITH
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("ocr_user_command")
switch p.curToken.Type {
case TokenFile:
p.nextToken()
var file string
file, err = p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["file"] = file
p.nextToken()
case TokenURL:
p.nextToken()
var url string
url, err = p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["url"] = url
p.nextToken()
default:
return nil, fmt.Errorf("expected FILE or URL")
}
cmd.Params["composite_model_name"] = compositeModelName
return cmd, nil
}
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
func (p *Parser) parseModelParseCommand() (*Command, error) {
p.nextToken() // consume WITH
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
cmd := NewCommand("parse_file_user_command")
switch p.curToken.Type {
case TokenFile:
p.nextToken()
var file string
file, err = p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["file"] = file
p.nextToken()
case TokenURL:
p.nextToken()
var url string
url, err = p.parseQuotedString()
if err != nil {
return nil, err
}
cmd.Params["url"] = url
p.nextToken()
default:
return nil, fmt.Errorf("expected FILE or URL")
}
cmd.Params["composite_model_name"] = compositeModelName
return cmd, nil
}
func (p *Parser) parseCheckCommand() (*Command, error) {
p.nextToken() // consume CHECK
switch p.curToken.Type {
case TokenInstance:
return p.parseCheckInstanceCommand()
case TokenProvider:
return p.parseCheckProviderByKeyCommand()
default:
return nil, fmt.Errorf("expected INSTANCE or PROVIDER after CHECK")
}
}
func (p *Parser) parseCheckInstanceCommand() (*Command, error) {
if p.curToken.Type != TokenInstance {
return nil, fmt.Errorf("expected INSTANCE after CHECK")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected instance name after INSTANCE")
}
instanceName := p.curToken.Value
p.nextToken()
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM after instance name")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected provider name after FROM")
}
providerName := p.curToken.Value
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("check_provider_connection")
cmd.Params["provider_name"] = providerName
cmd.Params["instance_name"] = instanceName
return cmd, nil
}
func (p *Parser) parseCheckProviderByKeyCommand() (*Command, error) {
if p.curToken.Type != TokenProvider {
return nil, fmt.Errorf("expected PROVIDER after CHECK")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected provider name after PROVIDER")
}
providerName := p.curToken.Value
p.nextToken()
if p.curToken.Type != TokenRegion {
return nil, fmt.Errorf("expected REGION after provider name")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected region name after REGION")
}
regionName := p.curToken.Value
p.nextToken()
if p.curToken.Type != TokenKey {
return nil, fmt.Errorf("expected KEY after region name")
}
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected API key after KEY")
}
apiKey := p.curToken.Value
p.nextToken()
baseURL := ""
if p.curToken.Type == TokenURL {
p.nextToken()
if p.curToken.Type != TokenQuotedString {
return nil, fmt.Errorf("expected base URL after URL")
}
baseURL = p.curToken.Value
p.nextToken()
}
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
if p.curToken.Type != TokenEOF {
return nil, fmt.Errorf("unexpected token: %s", p.curToken.Value)
}
cmd := NewCommand("check_provider_with_key")
cmd.Params["provider_name"] = providerName
cmd.Params["region"] = regionName
cmd.Params["api_key"] = apiKey
if baseURL != "" {
cmd.Params["base_url"] = baseURL
}
return cmd, nil
}
func (p *Parser) parseUseCommand() (*Command, error) {
p.nextToken() // consume USE
if p.curToken.Type != TokenModel {
return nil, fmt.Errorf("expected MODEL after USE")
}
p.nextToken() // consume MODEL
// Parse model identifier in format 'model@instance@provider'
compositeModelName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected model identifier in format 'model@instance@provider': %w", err)
}
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
cmd := NewCommand("use_model")
cmd.Params["composite_model_name"] = compositeModelName
return cmd, nil
}
func (p *Parser) parseParseCommand() (*Command, error) {
p.nextToken() // consume PARSE
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
switch p.curToken.Type {
case TokenDataset:
return p.parseParseDataset()
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
case TokenWith:
return p.parseModelParseCommand()
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
case TokenDocument:
Go: add file parse command (#14892) ### What problem does this PR solve? ``` RAGFlow(user)> ocr with 'hunyuanocr@test@gitee' file './picture.png' +----------------------------------------------------------+ | text | +----------------------------------------------------------+ | 生活不是等待风暴过去,而是学会在雨中翩翩起舞。 ——佚名 | +----------------------------------------------------------+ RAGFlow(user)> list 'test@gitee' tasks; +---------+----------------------------------+ | status | task_id | +---------+----------------------------------+ | success | C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5 | +---------+----------------------------------+ RAGFlow(user)> show 'test@gitee' task 'C3FX4MQNKY5MGC6ZFMIXIAMJKHCEBQB5'; +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | content | index | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ | # PDF 1: Purpose of RAGFlow RAGFlow is an open source Retrieval-Augmented Generation (RAG) engine designed to turn raw documents into reliable context for large language models.Its purpose is to make it practical to build an Al assistant that can ans... | 1 | +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-15 12:29:52 +08:00
return p.parseParseDocs()
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
default:
return nil, fmt.Errorf("expected DATASET, WITH, or DOCUMENT")
}
}
func (p *Parser) parseParseDataset() (*Command, error) {
p.nextToken() // consume DATASET
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
var method string
if p.curToken.Type == TokenSync {
method = "sync"
} else if p.curToken.Type == TokenAsync {
method = "async"
} else {
return nil, fmt.Errorf("expected SYNC or ASYNC")
}
cmd := NewCommand("parse_dataset")
cmd.Params["dataset_name"] = datasetName
cmd.Params["method"] = method
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseParseDocs() (*Command, error) {
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
p.nextToken() // consume document
documentsStr, err := p.parseQuotedString()
if err != nil {
return nil, err
}
p.nextToken()
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM")
}
p.nextToken()
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
datasetID, err := p.parseQuotedString()
if err != nil {
return nil, err
}
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
p.nextToken()
Go: add ingestion server (#15094) ### What problem does this PR solve? 1. Go ingestion server will connected with admin server with gRPC stream 2. Go ingestion server will be responsible for ingestion tasks ``` RAGFlow(admin)> list ingestors; +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | address | cpu_usage | id | last_heartbeat | name | process_id | rss_usage | status | task_count | vms_usage | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ | 127.0.0.1:58564 | 0 | bdd1870eea2646e0aacb8a2cd3307aa2 | 2026-05-24T18:16:17+08:00 | ingestor | 680152 | 212.72265625 | active | 0 | 2589.12109375 | +-----------------+-----------+----------------------------------+---------------------------+----------+------------+--------------+--------+------------+---------------+ RAGFlow(admin)> start ingestion 'abc'; +----------------------------------+ | task_id | +----------------------------------+ | e714777639ca4760ab427b5f211e81ad | +----------------------------------+ RAGFlow(admin)> stop ingestion 'f7bd39d0a724457eb5fdce6d81699776'; +----------------------------------+ | task_id | +----------------------------------+ | f7bd39d0a724457eb5fdce6d81699776 | +----------------------------------+ RAGFlow(admin)> list tasks; +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | ETA | assign_to | error | from | id | last_update | start_time | status | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | eae6431da72a40e796cff3a03008091b | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 6cccdd174bd049ecb05a774bbb47593f | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | ef360d777e57485799adb96b30f2b4b8 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bcc5c5448cb64de48b6b6171c36fb790 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | bfc25384c43a443294fe2da979a38ac2 | 2026-05-24T19:46:03+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 84960537b85d413b8990a9efd5952d67 | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 3d223c1b51e24b36861a3bfb2f1d58d4 | 2026-05-24T19:46:03+08:00 | | CANCELED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | e433b0e356b846c89c301621a3c54494 | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7c93a3880f074ebd8eca14e6b51bb7ef | 2026-05-24T19:46:03+08:00 | | COMPLETED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df2e4ef51aaf4390bff9a23f2692486e | 2026-05-24T19:46:04+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 7377c53010194ef7a83aa206698d66ff | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | df64d1a1f9d348e3a2f174c4d7d69e73 | 2026-05-24T19:46:05+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | b59834512e2847e1bdf13ace04b8a456 | 2026-05-24T19:46:06+08:00 | | DISPATCHED | | 0 | 17937da188b84f23a5c10bb87588944b | | CLI | 0064bb0ab69344028d1ecfda053826f4 | 2026-05-24T19:46:03+08:00 | | QUEUED | +-----+----------------------------------+-------+------+----------------------------------+---------------------------+------------+------------+ ``` ### Type of change - [x] New Feature (non-breaking change which adds functionality) --------- Signed-off-by: Jin Hai <haijin.chn@gmail.com>
2026-05-25 14:00:08 +08:00
cmd := NewCommand("parse_documents_user_command")
documents := strings.Split(documentsStr, " ")
cmd.Params["documents"] = documents
cmd.Params["dataset_id"] = datasetID
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseBenchmarkCommand() (*Command, error) {
cmd := NewCommand("benchmark")
p.nextToken() // consume BENCHMARK
concurrency, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd.Params["concurrency"] = concurrency
p.nextToken()
iterations, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd.Params["iterations"] = iterations
p.nextToken()
// Parse user_statement
nestedCmd, err := p.parseUserStatement() // Not only user statement
if err != nil {
return nil, err
}
cmd.Params["command"] = nestedCmd
return cmd, nil
}
func (p *Parser) parseUserStatement() (*Command, error) {
switch p.curToken.Type {
case TokenPing:
return p.parsePingServer()
case TokenDelete:
return p.parseDeleteCommand()
case TokenShow:
return p.parseShowCommand()
case TokenCreate:
return p.parseCreateCommand()
case TokenDrop:
return p.parseDropCommand()
case TokenSet:
return p.parseSetCommand()
case TokenUnset:
return p.parseUnsetCommand()
case TokenReset:
return p.parseResetCommand()
case TokenList:
return p.parseListCommand()
case TokenParse:
return p.parseParseCommand()
case TokenImport:
return p.parseImportCommand()
case TokenInsert:
return p.parseInsertCommand()
case TokenSearch:
return p.parseSearchCommand()
case TokenGet:
return p.parseGetCommand()
case TokenUpdate:
return p.parseUpdateCommand()
case TokenRemove:
return p.parseRemoveCommand()
default:
return nil, fmt.Errorf("invalid user statement: %s", p.curToken.Value)
}
}
func (p *Parser) parseStartupCommand() (*Command, error) {
p.nextToken() // consume STARTUP
if p.curToken.Type != TokenService {
return nil, fmt.Errorf("expected SERVICE")
}
p.nextToken()
serviceNum, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd := NewCommand("startup_service")
cmd.Params["number"] = serviceNum
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseShutdownCommand() (*Command, error) {
p.nextToken() // consume SHUTDOWN
if p.curToken.Type != TokenService {
return nil, fmt.Errorf("expected SERVICE")
}
p.nextToken()
serviceNum, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd := NewCommand("shutdown_service")
cmd.Params["number"] = serviceNum
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseRestartCommand() (*Command, error) {
p.nextToken() // consume RESTART
if p.curToken.Type != TokenService {
return nil, fmt.Errorf("expected SERVICE")
}
p.nextToken()
serviceNum, err := p.parseNumber()
if err != nil {
return nil, err
}
cmd := NewCommand("restart_service")
cmd.Params["number"] = serviceNum
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
func (p *Parser) parseUnsetCommand() (*Command, error) {
p.nextToken() // consume UNSET
if p.curToken.Type != TokenToken {
return nil, fmt.Errorf("expected TOKEN after UNSET")
}
p.nextToken()
// Semicolon is optional for UNSET TOKEN
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return NewCommand("unset_token"), nil
}
// parseGetCommand parses: GET CHUNK 'chunk_id'
func (p *Parser) parseGetCommand() (*Command, error) {
p.nextToken() // consume GET
if p.curToken.Type == TokenChunk {
return p.parseGetChunk()
}
return nil, fmt.Errorf("unknown GET target: %s", p.curToken.Value)
}
// parseGetChunk parses: GET CHUNK 'chunk_id' OF DOCUMENT 'doc_id' IN DATASET 'dataset_id'
func (p *Parser) parseGetChunk() (*Command, error) {
p.nextToken() // consume CHUNK
// Parse chunk_id
chunkID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected chunk_id: %w", err)
}
cmd := NewCommand("get_chunk")
cmd.Params["chunk_id"] = chunkID
p.nextToken()
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF after chunk_id")
}
p.nextToken()
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after OF")
}
p.nextToken()
// Parse doc_id
docID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected doc_id: %w", err)
}
cmd.Params["doc_id"] = docID
p.nextToken()
if p.curToken.Type != TokenIn {
return nil, fmt.Errorf("expected IN after doc_id")
}
p.nextToken()
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET after IN")
}
p.nextToken()
// Parse dataset_id
datasetID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset_id: %w", err)
}
cmd.Params["dataset_id"] = datasetID
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// Internal
// parseUpdateCommand parses: UPDATE CHUNK 'chunk_id' OF DATASET 'dataset_name' SET '{"content": "..."}'
func (p *Parser) parseUpdateCommand() (*Command, error) {
p.nextToken() // consume UPDATE
if p.curToken.Type == TokenChunk {
return p.parseUpdateChunk()
}
return nil, fmt.Errorf("unknown UPDATE target: %s", p.curToken.Value)
}
// Internal CLI for GO
// parseUpdateChunk parses: UPDATE CHUNK 'chunk_id' OF DOCUMENT 'doc_id' IN DATASET 'dataset_id' SET '{"content": "..."}'
func (p *Parser) parseUpdateChunk() (*Command, error) {
p.nextToken() // consume CHUNK
// Parse chunk_id
chunkID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected chunk_id: %w", err)
}
cmd := NewCommand("update_chunk")
cmd.Params["chunk_id"] = chunkID
p.nextToken()
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF after chunk_id")
}
p.nextToken()
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after OF")
}
p.nextToken()
// Parse doc_id
docID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected doc_id: %w", err)
}
cmd.Params["doc_id"] = docID
p.nextToken()
if p.curToken.Type != TokenIn {
return nil, fmt.Errorf("expected IN after doc_id")
}
p.nextToken()
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET after IN")
}
p.nextToken()
// Parse dataset_name
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset_name: %w", err)
}
cmd.Params["dataset_name"] = datasetName
p.nextToken()
if p.curToken.Type != TokenSet {
return nil, fmt.Errorf("expected SET after dataset_name")
}
p.nextToken()
// Parse JSON body
jsonBody, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected JSON body: %w", err)
}
cmd.Params["json_body"] = jsonBody
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseSetMeta parses: SET METADATA OF DOCUMENT 'doc_id' TO '{"key": "value"}'
func (p *Parser) parseSetMeta() (*Command, error) {
p.nextToken() // consume METADATA
// Expect OF
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF after SET METADATA")
}
p.nextToken()
// Expect DOCUMENT
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after SET METADATA OF")
}
p.nextToken()
// Parse doc_id
docID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected doc_id: %w", err)
}
cmd := NewCommand("set_meta")
cmd.Params["doc_id"] = docID
p.nextToken()
// Expect TO
if p.curToken.Type != TokenTo {
return nil, fmt.Errorf("expected TO after doc_id")
}
p.nextToken()
// Parse meta JSON
meta, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected meta JSON: %w", err)
}
cmd.Params["meta"] = meta
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseDeleteMeta parses: DELETE METADATA OF DOCUMENT 'doc_id' [KEYS '["key1", "key2"]']
// If KEYS is not provided, deletes entire document metadata
func (p *Parser) parseDeleteMeta() (*Command, error) {
p.nextToken() // consume METADATA
// Expect OF
if p.curToken.Type != TokenOf {
return nil, fmt.Errorf("expected OF after DELETE METADATA")
}
p.nextToken()
// Expect DOCUMENT
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after DELETE METADATA OF")
}
p.nextToken()
// Parse doc_id
docID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected doc_id: %w", err)
}
cmd := NewCommand("delete_meta")
cmd.Params["doc_id"] = docID
p.nextToken()
// KEYS is optional - if not provided, delete entire document metadata
if p.curToken.Type != TokenKeys {
if p.curToken.Type == TokenSemicolon {
p.nextToken()
return cmd, nil
}
if p.curToken.Type == TokenEOF {
return cmd, nil
}
return nil, fmt.Errorf("expected KEYS or end of command after doc_id")
}
// Parse keys JSON array
p.nextToken()
keys, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected keys JSON array: %w", err)
}
cmd.Params["keys"] = keys
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
return cmd, nil
}
if p.curToken.Type != TokenEOF {
return nil, fmt.Errorf("expected end of command after KEYS")
}
return cmd, nil
}
// parseRemoveTags parses: REMOVE TAGS 'tag1', 'tag2' from DATASET 'dataset_name';
func (p *Parser) parseRemoveTags() (*Command, error) {
p.nextToken() // consume TAGS
// Parse first tag
tag, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected tag: %w", err)
}
tags := []string{tag}
// Parse additional tags separated by commas
for {
p.nextToken()
if p.curToken.Type == TokenComma {
p.nextToken()
tag, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected tag after comma: %w", err)
}
tags = append(tags, tag)
} else {
break
}
}
cmd := NewCommand("rm_tags")
cmd.Params["tags"] = tags
// Expect from
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM after tags")
}
p.nextToken()
// Expect DATASET
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET after FROM")
}
p.nextToken()
// Parse dataset_name
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset_name: %w", err)
}
cmd.Params["dataset_name"] = datasetName
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}
// parseRemoveChunk parses:
// - REMOVE CHUNKS 'chunk_id1', 'chunk_id2' FROM DOCUMENT 'doc_id' IN DATASET 'dataset_name';
// - REMOVE ALL CHUNKS FROM DOCUMENT 'doc_id' IN DATASET 'dataset_name';
func (p *Parser) parseRemoveChunk() (*Command, error) {
cmd := NewCommand("remove_chunks")
// Check if ALL CHUNKS - if we came here from TokenAll case, curToken is already ALL
if p.curToken.Type == TokenAll {
p.nextToken() // consume ALL
if p.curToken.Type != TokenChunks {
return nil, fmt.Errorf("expected CHUNKS after ALL")
}
p.nextToken() // consume CHUNKS
cmd.Params["delete_all"] = true
} else {
// curToken is TokenChunks, consume it first
p.nextToken()
// Multiple chunks: REMOVE CHUNKS 'id1' 'id2' FROM DOCUMENT 'doc_id' IN DATASET 'dataset_name' (space-separated)
// Parse first chunk ID
chunkID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected chunk_id: %w", err)
}
chunkIDs := []string{chunkID}
// Parse additional chunk IDs separated by spaces (each quoted)
for {
p.nextToken()
// Stop if we hit FROM or non-quoted token
if p.curToken.Type == TokenFrom || p.curToken.Type != TokenQuotedString {
break
}
chunkID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected chunk_id: %w", err)
}
chunkIDs = append(chunkIDs, chunkID)
}
cmd.Params["chunk_ids"] = chunkIDs
}
// Expect FROM
if p.curToken.Type != TokenFrom {
return nil, fmt.Errorf("expected FROM after chunk(s)")
}
p.nextToken()
// Expect DOCUMENT
if p.curToken.Type != TokenDocument {
return nil, fmt.Errorf("expected DOCUMENT after FROM")
}
p.nextToken()
// Parse doc_id
docID, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected doc_id: %w", err)
}
cmd.Params["doc_id"] = docID
p.nextToken()
// Expect IN
if p.curToken.Type != TokenIn {
return nil, fmt.Errorf("expected IN after doc_id")
}
p.nextToken()
// Expect DATASET
if p.curToken.Type != TokenDataset {
return nil, fmt.Errorf("expected DATASET after IN")
}
p.nextToken()
// Parse dataset_name (quoted string)
datasetName, err := p.parseQuotedString()
if err != nil {
return nil, fmt.Errorf("expected dataset_name: %w", err)
}
cmd.Params["dataset_name"] = datasetName
p.nextToken()
// Semicolon is optional
if p.curToken.Type == TokenSemicolon {
p.nextToken()
}
return cmd, nil
}