2026-07-31 18:32:07 +08:00
|
|
|
//
|
|
|
|
|
// Copyright 2026 The InfiniFlow Authors. All Rights Reserved.
|
|
|
|
|
//
|
|
|
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
|
// you may not use this file except in compliance with the License.
|
|
|
|
|
// You may obtain a copy of the License at
|
|
|
|
|
//
|
|
|
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
|
//
|
|
|
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
|
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
|
// See the License for the specific language governing permissions and
|
|
|
|
|
// limitations under the License.
|
|
|
|
|
//
|
|
|
|
|
|
|
|
|
|
package models
|
|
|
|
|
|
feat(go-models): migrate batch 3 model drivers to unified handlers (#17698)
## Summary
Relate to #17284
Migrate 10 OpenAI-compatible drivers (`minimax`, `mistral`,
`modelscope`, `moonshot`, `n1n`, `novita`, `ollama`, `openai`,
`openai_api_compatible`, `openrouter`) to use the unified response
handlers (`HandleNonStreamingResponse` / `HandleStreamingResponse`),
following the same pattern established by `deepseek` in #17634.
- Cut ~150 lines per driver (1692 lines removed, 172 added across 10
files).
- `openai_api_compatible` gains `ChatWithMessages` +
`ChatStreamlyWithSender` required by the unified handler infrastructure.
- `openai` driver preserves `reasoning_content` extraction for o-series
models.
- All drivers: pure deduplication of HTTP plumbing.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-03 15:08:55 +08:00
|
|
|
// ParserConfig maps a protocol to its usage and reasoning parsers. Drivers
|
|
|
|
|
// select a ParserConfig instead of implementing extraction individually.
|
2026-07-31 18:32:07 +08:00
|
|
|
type ParserConfig struct {
|
|
|
|
|
// Protocol is the protocol identifier (e.g. "openai", "claude").
|
|
|
|
|
Protocol string
|
|
|
|
|
// ResponseParser extracts usage from a non-streaming response body.
|
|
|
|
|
ResponseParser func(map[string]any) (*TokenUsage, bool)
|
|
|
|
|
// StreamParser extracts usage from one streaming event.
|
|
|
|
|
StreamParser func(map[string]any) (*TokenUsage, bool)
|
feat(go-models): migrate batch 3 model drivers to unified handlers (#17698)
## Summary
Relate to #17284
Migrate 10 OpenAI-compatible drivers (`minimax`, `mistral`,
`modelscope`, `moonshot`, `n1n`, `novita`, `ollama`, `openai`,
`openai_api_compatible`, `openrouter`) to use the unified response
handlers (`HandleNonStreamingResponse` / `HandleStreamingResponse`),
following the same pattern established by `deepseek` in #17634.
- Cut ~150 lines per driver (1692 lines removed, 172 added across 10
files).
- `openai_api_compatible` gains `ChatWithMessages` +
`ChatStreamlyWithSender` required by the unified handler infrastructure.
- `openai` driver preserves `reasoning_content` extraction for o-series
models.
- All drivers: pure deduplication of HTTP plumbing.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-03 15:08:55 +08:00
|
|
|
// ExtractStreamReasoning extracts the reasoning text from a parsed
|
|
|
|
|
// delta map (the delta field of one streaming event), if any.
|
|
|
|
|
// Defaults to reading delta.reasoning_content.
|
|
|
|
|
ExtractStreamReasoning func(delta map[string]any) string
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// extractDefaultStreamReasoning reads the reasoning text from a parsed
|
|
|
|
|
// delta (delta.reasoning_content).
|
|
|
|
|
func extractDefaultStreamReasoning(delta map[string]any) string {
|
|
|
|
|
if r, ok := delta["reasoning_content"].(string); ok {
|
|
|
|
|
return r
|
|
|
|
|
}
|
|
|
|
|
return ""
|
2026-07-31 18:32:07 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// OpenAIParserConfig is the ParserConfig for OpenAI-compatible APIs
|
feat(go-models): migrate batch 3 model drivers to unified handlers (#17698)
## Summary
Relate to #17284
Migrate 10 OpenAI-compatible drivers (`minimax`, `mistral`,
`modelscope`, `moonshot`, `n1n`, `novita`, `ollama`, `openai`,
`openai_api_compatible`, `openrouter`) to use the unified response
handlers (`HandleNonStreamingResponse` / `HandleStreamingResponse`),
following the same pattern established by `deepseek` in #17634.
- Cut ~150 lines per driver (1692 lines removed, 172 added across 10
files).
- `openai_api_compatible` gains `ChatWithMessages` +
`ChatStreamlyWithSender` required by the unified handler infrastructure.
- `openai` driver preserves `reasoning_content` extraction for o-series
models.
- All drivers: pure deduplication of HTTP plumbing.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-03 15:08:55 +08:00
|
|
|
// (NVIDIA NIM, DeepSeek, Aliyun, Moonshot, xAI, ...).
|
2026-07-31 18:32:07 +08:00
|
|
|
var OpenAIParserConfig = &ParserConfig{
|
feat(go-models): migrate batch 3 model drivers to unified handlers (#17698)
## Summary
Relate to #17284
Migrate 10 OpenAI-compatible drivers (`minimax`, `mistral`,
`modelscope`, `moonshot`, `n1n`, `novita`, `ollama`, `openai`,
`openai_api_compatible`, `openrouter`) to use the unified response
handlers (`HandleNonStreamingResponse` / `HandleStreamingResponse`),
following the same pattern established by `deepseek` in #17634.
- Cut ~150 lines per driver (1692 lines removed, 172 added across 10
files).
- `openai_api_compatible` gains `ChatWithMessages` +
`ChatStreamlyWithSender` required by the unified handler infrastructure.
- `openai` driver preserves `reasoning_content` extraction for o-series
models.
- All drivers: pure deduplication of HTTP plumbing.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-03 15:08:55 +08:00
|
|
|
Protocol: "openai",
|
|
|
|
|
ResponseParser: extractOpenAIUsage,
|
|
|
|
|
StreamParser: extractOpenAIStreamUsage,
|
|
|
|
|
ExtractStreamReasoning: extractDefaultStreamReasoning,
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// extractOpenRouterStreamReasoning reads the reasoning text from an
|
|
|
|
|
// OpenRouter streaming event. OpenRouter uses delta.reasoning (not
|
|
|
|
|
// delta.reasoning_content) for its reasoning content.
|
|
|
|
|
func extractOpenRouterStreamReasoning(delta map[string]any) string {
|
|
|
|
|
if r, ok := delta["reasoning"].(string); ok {
|
|
|
|
|
return r
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// OpenRouterParserConfig is the ParserConfig for OpenRouter, which emits
|
|
|
|
|
// reasoning under delta.reasoning instead of delta.reasoning_content.
|
|
|
|
|
var OpenRouterParserConfig = &ParserConfig{
|
|
|
|
|
Protocol: "openai",
|
|
|
|
|
ResponseParser: extractOpenAIUsage,
|
|
|
|
|
StreamParser: extractOpenAIStreamUsage,
|
|
|
|
|
ExtractStreamReasoning: extractOpenRouterStreamReasoning,
|
2026-07-31 18:32:07 +08:00
|
|
|
}
|