mirror of
https://github.com/infiniflow/ragflow.git
synced 2026-07-29 04:08:12 +08:00
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>
This commit is contained in:
@@ -273,6 +273,8 @@ func (c *RAGFlowClient) ExecuteUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
return c.ASRUserCommand(cmd)
|
||||
case "ocr_user_command":
|
||||
return c.OCRUserCommand(cmd)
|
||||
case "parse_file_user_command":
|
||||
return c.ParseFileUserCommand(cmd)
|
||||
case "check_provider_connection":
|
||||
return c.CheckProviderConnection(cmd)
|
||||
case "use_model":
|
||||
@@ -285,6 +287,10 @@ func (c *RAGFlowClient) ExecuteUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
return c.ResetDefaultModel(cmd)
|
||||
case "list_user_default_models":
|
||||
return c.ListDefaultModels(cmd)
|
||||
case "list_tasks_user_command":
|
||||
return c.ListTasksUserCommand(cmd)
|
||||
case "show_task_user_command":
|
||||
return c.ShowTaskUserCommand(cmd)
|
||||
// Dataset, metadata commands
|
||||
case "create_dataset_table":
|
||||
return c.CreateDatasetInDocEngine(cmd)
|
||||
|
||||
@@ -437,6 +437,8 @@ func (l *Lexer) lookupIdent(ident string) Token {
|
||||
return Token{Type: TokenRegion, Value: ident}
|
||||
case "URL":
|
||||
return Token{Type: TokenURL, Value: ident}
|
||||
case "TASK":
|
||||
return Token{Type: TokenTask, Value: ident}
|
||||
case "TASKS":
|
||||
return Token{Type: TokenTasks, Value: ident}
|
||||
case "LOG":
|
||||
|
||||
@@ -319,6 +319,44 @@ func (r *EmbeddingsResponse) PrintOut() {
|
||||
}
|
||||
}
|
||||
|
||||
type SegmentResponse struct {
|
||||
Segments []map[string]interface{} `json:"segments"`
|
||||
}
|
||||
|
||||
type TaskResponse struct {
|
||||
Code int `json:"code"`
|
||||
Data map[string]interface{} `json:"data"`
|
||||
Message string `json:"message"`
|
||||
Duration float64
|
||||
OutputFormat OutputFormat
|
||||
}
|
||||
|
||||
func (r *TaskResponse) Type() string {
|
||||
return "task"
|
||||
}
|
||||
|
||||
func (r *TaskResponse) TimeCost() float64 {
|
||||
return r.Duration
|
||||
}
|
||||
|
||||
func (r *TaskResponse) SetOutputFormat(format OutputFormat) {
|
||||
r.OutputFormat = format
|
||||
}
|
||||
|
||||
func (r *TaskResponse) PrintOut() {
|
||||
if r.Code == 0 {
|
||||
segmentsRaw := r.Data["segments"].([]interface{})
|
||||
segments := make([]map[string]interface{}, len(segmentsRaw))
|
||||
for i, v := range segmentsRaw {
|
||||
segments[i] = v.(map[string]interface{})
|
||||
}
|
||||
PrintTableSimpleByFormat(segments, r.OutputFormat)
|
||||
} else {
|
||||
fmt.Println("ERROR")
|
||||
fmt.Printf("%d, %s\n", r.Code, r.Message)
|
||||
}
|
||||
}
|
||||
|
||||
// ==================== ContextEngine Commands ====================
|
||||
|
||||
// ContextListResponse represents the response for ls command
|
||||
|
||||
@@ -152,6 +152,7 @@ const (
|
||||
TokenTag
|
||||
TokenRegion
|
||||
TokenURL
|
||||
TokenTask
|
||||
TokenTasks
|
||||
TokenLog
|
||||
TokenLevel
|
||||
|
||||
@@ -2284,6 +2284,179 @@ func (c *RAGFlowClient) OCRUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
return &result, nil
|
||||
}
|
||||
|
||||
func (c *RAGFlowClient) ParseFileUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
if c.HTTPClient.APIToken == "" && c.HTTPClient.LoginToken == "" {
|
||||
return nil, fmt.Errorf("API token not set. Please login first")
|
||||
}
|
||||
|
||||
if c.ServerType != "user" {
|
||||
return nil, fmt.Errorf("this command is only allowed in USER mode")
|
||||
}
|
||||
|
||||
var providerName, instanceName, modelName string
|
||||
|
||||
// Check if composite_model_name is provided in command
|
||||
if compositeModelName, ok := cmd.Params["composite_model_name"].(string); ok && compositeModelName != "" {
|
||||
names := strings.Split(compositeModelName, "@")
|
||||
if len(names) != 3 {
|
||||
return nil, fmt.Errorf("model name must be in format 'model@instance@provider'")
|
||||
}
|
||||
providerName = names[2]
|
||||
instanceName = names[1]
|
||||
modelName = names[0]
|
||||
} else if c.CurrentModel != nil {
|
||||
// Use current model if set
|
||||
providerName = c.CurrentModel.Provider
|
||||
instanceName = c.CurrentModel.Instance
|
||||
modelName = c.CurrentModel.Model
|
||||
} else {
|
||||
return nil, fmt.Errorf("model name not provided and no current model set. Use 'use model' command first")
|
||||
}
|
||||
|
||||
var filename string
|
||||
var fileURL string
|
||||
var ok bool
|
||||
var fileContent []byte
|
||||
|
||||
filename, ok = cmd.Params["file"].(string)
|
||||
if ok {
|
||||
// read file and convert to base64
|
||||
var err error
|
||||
fileContent, err = os.ReadFile(filename)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to read file: %w", err)
|
||||
}
|
||||
} else {
|
||||
fileURL, ok = cmd.Params["url"].(string)
|
||||
if !ok {
|
||||
return nil, fmt.Errorf("file or url not provided")
|
||||
}
|
||||
}
|
||||
|
||||
payload := map[string]interface{}{
|
||||
"provider_name": providerName,
|
||||
"instance_name": instanceName,
|
||||
"model_name": modelName,
|
||||
}
|
||||
|
||||
if fileContent != nil {
|
||||
payload["content"] = fileContent
|
||||
} else {
|
||||
payload["url"] = fileURL
|
||||
}
|
||||
|
||||
url := "/file/parse"
|
||||
|
||||
resp, err := c.HTTPClient.Request("POST", url, "web", nil, payload)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to PARSE document: %w", err)
|
||||
}
|
||||
if resp.StatusCode != 200 {
|
||||
return nil, fmt.Errorf("failed to PARSE document: HTTP %d, body: %s", resp.StatusCode, string(resp.Body))
|
||||
}
|
||||
var result CommonDataResponse
|
||||
if err = json.Unmarshal(resp.Body, &result); err != nil {
|
||||
return nil, fmt.Errorf("PARSE document failed: invalid JSON (%w)", err)
|
||||
}
|
||||
if result.Code != 0 {
|
||||
return nil, fmt.Errorf("%s", result.Message)
|
||||
}
|
||||
result.Duration = resp.Duration
|
||||
|
||||
return &result, nil
|
||||
}
|
||||
|
||||
func (c *RAGFlowClient) ListTasksUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
if c.HTTPClient.APIToken == "" && c.HTTPClient.LoginToken == "" {
|
||||
return nil, fmt.Errorf("API token not set. Please login first")
|
||||
}
|
||||
|
||||
if c.ServerType != "user" {
|
||||
return nil, fmt.Errorf("this command is only allowed in USER mode")
|
||||
}
|
||||
|
||||
var providerName, instanceName string
|
||||
|
||||
// Check if composite_instance_name is provided in command
|
||||
if compositeModelName, ok := cmd.Params["composite_instance_name"].(string); ok && compositeModelName != "" {
|
||||
names := strings.Split(compositeModelName, "@")
|
||||
if len(names) != 2 {
|
||||
return nil, fmt.Errorf("model name must be in format 'instance@provider'")
|
||||
}
|
||||
providerName = names[1]
|
||||
instanceName = names[0]
|
||||
} else {
|
||||
return nil, fmt.Errorf("no provider name or instance name")
|
||||
}
|
||||
|
||||
url := fmt.Sprintf("/providers/%s/instances/%s/tasks", providerName, instanceName)
|
||||
|
||||
resp, err := c.HTTPClient.Request("GET", url, "web", nil, nil)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to list tasks: %w", err)
|
||||
}
|
||||
if resp.StatusCode != 200 {
|
||||
return nil, fmt.Errorf("failed to list tasks: HTTP %d, body: %s", resp.StatusCode, string(resp.Body))
|
||||
}
|
||||
var result CommonResponse
|
||||
if err = json.Unmarshal(resp.Body, &result); err != nil {
|
||||
return nil, fmt.Errorf("list tasks failed: invalid JSON (%w)", err)
|
||||
}
|
||||
if result.Code != 0 {
|
||||
return nil, fmt.Errorf("%s", result.Message)
|
||||
}
|
||||
result.Duration = resp.Duration
|
||||
return &result, nil
|
||||
}
|
||||
|
||||
func (c *RAGFlowClient) ShowTaskUserCommand(cmd *Command) (ResponseIf, error) {
|
||||
if c.HTTPClient.APIToken == "" && c.HTTPClient.LoginToken == "" {
|
||||
return nil, fmt.Errorf("API token not set. Please login first")
|
||||
}
|
||||
|
||||
if c.ServerType != "user" {
|
||||
return nil, fmt.Errorf("this command is only allowed in USER mode")
|
||||
}
|
||||
|
||||
var providerName, instanceName string
|
||||
|
||||
// Check if composite_instance_name is provided in command
|
||||
if compositeModelName, ok := cmd.Params["composite_instance_name"].(string); ok && compositeModelName != "" {
|
||||
names := strings.Split(compositeModelName, "@")
|
||||
if len(names) != 2 {
|
||||
return nil, fmt.Errorf("model name must be in format 'instance@provider'")
|
||||
}
|
||||
providerName = names[1]
|
||||
instanceName = names[0]
|
||||
} else {
|
||||
return nil, fmt.Errorf("no provider name or instance name")
|
||||
}
|
||||
|
||||
taskID, ok := cmd.Params["task_id"].(string)
|
||||
if !ok {
|
||||
return nil, fmt.Errorf("task id not provided")
|
||||
}
|
||||
|
||||
url := fmt.Sprintf("/providers/%s/instances/%s/tasks/%s", providerName, instanceName, taskID)
|
||||
|
||||
resp, err := c.HTTPClient.Request("GET", url, "web", nil, nil)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to get task: %w", err)
|
||||
}
|
||||
if resp.StatusCode != 200 {
|
||||
return nil, fmt.Errorf("failed to get task: HTTP %d, body: %s", resp.StatusCode, string(resp.Body))
|
||||
}
|
||||
var result TaskResponse
|
||||
if err = json.Unmarshal(resp.Body, &result); err != nil {
|
||||
return nil, fmt.Errorf("get task failed: invalid JSON (%w)", err)
|
||||
}
|
||||
if result.Code != 0 {
|
||||
return nil, fmt.Errorf("%s", result.Message)
|
||||
}
|
||||
result.Duration = resp.Duration
|
||||
return &result, nil
|
||||
}
|
||||
|
||||
func (c *RAGFlowClient) CheckProviderConnection(cmd *Command) (ResponseIf, error) {
|
||||
if c.HTTPClient.APIToken == "" && c.HTTPClient.LoginToken == "" {
|
||||
return nil, fmt.Errorf("API token not set. Please login first")
|
||||
|
||||
@@ -163,6 +163,8 @@ func (p *Parser) parseListCommand() (*Command, error) {
|
||||
return NewCommand("list_user_chats"), nil
|
||||
case TokenFiles:
|
||||
return p.parseListFiles()
|
||||
case TokenQuotedString:
|
||||
return p.parseListQuotedStringCommand()
|
||||
default:
|
||||
return nil, fmt.Errorf("unknown LIST target: %s", p.curToken.Value)
|
||||
}
|
||||
@@ -280,9 +282,57 @@ func (p *Parser) parseListFiles() (*Command, error) {
|
||||
return cmd, nil
|
||||
}
|
||||
|
||||
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()
|
||||
@@ -333,6 +383,10 @@ func (p *Parser) parseShowCommand() (*Command, error) {
|
||||
return p.parseShowInstance()
|
||||
case TokenBalance:
|
||||
return p.parseShowBalance()
|
||||
case TokenTask:
|
||||
return p.parseShowTask()
|
||||
case TokenQuotedString:
|
||||
return p.parseShowQuotedStringCommand()
|
||||
default:
|
||||
return nil, fmt.Errorf("unknown SHOW target: %s", p.curToken.Value)
|
||||
}
|
||||
@@ -1454,6 +1508,27 @@ func (p *Parser) parseShowBalance() (*Command, error) {
|
||||
return cmd, nil
|
||||
}
|
||||
|
||||
// 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
|
||||
@@ -2900,6 +2975,45 @@ func (p *Parser) parseOCRCommand() (*Command, error) {
|
||||
return cmd, nil
|
||||
}
|
||||
|
||||
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
|
||||
|
||||
@@ -2964,11 +3078,14 @@ func (p *Parser) parseUseCommand() (*Command, error) {
|
||||
func (p *Parser) parseParseCommand() (*Command, error) {
|
||||
p.nextToken() // consume PARSE
|
||||
|
||||
if p.curToken.Type == TokenDataset {
|
||||
switch p.curToken.Type {
|
||||
case TokenDataset:
|
||||
return p.parseParseDataset()
|
||||
case TokenWith:
|
||||
return p.parseModelParseCommand()
|
||||
default:
|
||||
return p.parseParseDocs()
|
||||
}
|
||||
|
||||
return p.parseParseDocs()
|
||||
}
|
||||
|
||||
func (p *Parser) parseParseDataset() (*Command, error) {
|
||||
|
||||
Reference in New Issue
Block a user