mirror of
https://github.com/CJackHwang/ds2api.git
synced 2026-05-07 01:45:27 +08:00
refactor: remove legacy compatibility configuration and UI components
This commit is contained in:
@@ -65,7 +65,7 @@ ds2api/
|
||||
│ ├── textclean/ # Text cleanup
|
||||
│ ├── toolcall/ # Tool-call parsing and repair
|
||||
│ ├── toolstream/ # Go streaming tool-call anti-leak and delta detection
|
||||
│ ├── translatorcliproxy/ # Cross-protocol translation bridge
|
||||
│ ├── translatorcliproxy/ # Vercel/fallback/test protocol translation bridge
|
||||
│ ├── util/ # Shared utility helpers
|
||||
│ ├── version/ # Version query/compare
|
||||
│ └── webui/ # WebUI static hosting logic
|
||||
@@ -187,12 +187,12 @@ flowchart LR
|
||||
|
||||
- `internal/server`: router tree + middlewares (health, protocol routes, Admin/WebUI).
|
||||
- `internal/httpapi/openai/*`: OpenAI HTTP surface split into chat, responses, files, embeddings, history, and shared packages; chat/responses share the promptcompat, stream, and toolcall semantics.
|
||||
- `internal/httpapi/{claude,gemini}`: protocol adapters that normalize into the same prompt compatibility semantics; direct paths share DeepSeek session/PoW/completion execution through `completionruntime`, while Vercel/proxy paths can still translate through `translatorcliproxy` into the OpenAI handler.
|
||||
- `internal/httpapi/{claude,gemini}`: protocol adapters that normalize into the same prompt compatibility semantics; normal direct paths must share DeepSeek session/PoW/completion execution through `completionruntime`, while `translatorcliproxy` is reserved for Vercel prepare/release, missing-backend fallback, and regression tests.
|
||||
- `internal/httpapi/requestbody`: shared HTTP body reading, JSON pre-validation, and UTF-8 error helpers across protocol adapters.
|
||||
- `internal/promptcompat`: compatibility core for turning OpenAI/Claude/Gemini requests into DeepSeek web-chat plain-text context.
|
||||
- `internal/assistantturn`: Go output-side canonical semantics, converting DeepSeek SSE collection results and stream finalization state into assistant turns and centralizing thinking, tool call, citation, usage, stop/error behavior.
|
||||
- `internal/completionruntime`: shared Go completion execution helpers for DeepSeek session/PoW/call startup, non-stream collection, and empty-output retry; streaming paths use it to start upstream requests, continue to use `internal/stream` for real-time consumption, and use `assistantturn` during finalization.
|
||||
- `internal/translatorcliproxy`: structure translation between Claude/Gemini and OpenAI.
|
||||
- `internal/translatorcliproxy`: bridge compatibility layer for Claude/Gemini and OpenAI shape translation; it is not the main business protocol conversion center.
|
||||
- `internal/deepseek/{client,protocol,transport}`: upstream requests, sessions, PoW adaptation, protocol constants, and transport details.
|
||||
- `internal/js/chat-stream` + `api/chat-stream.js`: Vercel Node streaming bridge; Go prepare/release owns auth, account lease, and completion payload assembly, while Node relays real-time SSE with Go-aligned finalization and tool sieve semantics.
|
||||
- `internal/stream` + `internal/sse`: Go stream parsing and incremental assembly.
|
||||
|
||||
@@ -65,7 +65,7 @@ ds2api/
|
||||
│ ├── textclean/ # 文本清洗
|
||||
│ ├── toolcall/ # 工具调用解析与修复
|
||||
│ ├── toolstream/ # Go 流式 tool call 防泄漏与增量检测
|
||||
│ ├── translatorcliproxy/ # 多协议互转桥
|
||||
│ ├── translatorcliproxy/ # Vercel/fallback/测试用协议互转桥
|
||||
│ ├── util/ # 通用工具函数
|
||||
│ ├── version/ # 版本查询/比较
|
||||
│ └── webui/ # WebUI 静态托管相关逻辑
|
||||
@@ -187,12 +187,12 @@ flowchart LR
|
||||
|
||||
- `internal/server`:路由树和中间件挂载(健康检查、协议入口、Admin/WebUI)。
|
||||
- `internal/httpapi/openai/*`:OpenAI HTTP surface,按 chat、responses、files、embeddings、history、shared 拆分;chat/responses 共享 promptcompat、stream、toolcall 等核心语义。
|
||||
- `internal/httpapi/{claude,gemini}`:协议输入输出适配,归一到同一套 prompt compatibility 语义;直连路径通过 `completionruntime` 共享 DeepSeek session/PoW/completion 调用,Vercel/代理路径仍可经 `translatorcliproxy` 转到 OpenAI handler。
|
||||
- `internal/httpapi/{claude,gemini}`:协议输入输出适配,归一到同一套 prompt compatibility 语义;正常直连路径必须通过 `completionruntime` 共享 DeepSeek session/PoW/completion 调用,`translatorcliproxy` 仅保留给 Vercel prepare/release、后端缺失 fallback 和回归测试。
|
||||
- `internal/httpapi/requestbody`:跨协议复用的请求体读取、JSON 解码前置校验与 UTF-8 错误处理辅助。
|
||||
- `internal/promptcompat`:OpenAI/Claude/Gemini 请求到 DeepSeek 网页纯文本上下文的兼容内核。
|
||||
- `internal/assistantturn`:Go 输出侧统一语义层,把 DeepSeek SSE 收集结果和流式收尾状态归一成 assistant turn,集中处理 thinking、tool call、citation、usage、stop/error 语义。
|
||||
- `internal/completionruntime`:Go surface 共享的 completion 执行辅助,负责 DeepSeek session/PoW/call 启动、非流式 collect 和 empty-output retry;流式路径复用它启动上游请求,继续用 `internal/stream` 做实时消费,并在最终收尾阶段接入 `assistantturn`。
|
||||
- `internal/translatorcliproxy`:Claude/Gemini 与 OpenAI 结构互转。
|
||||
- `internal/translatorcliproxy`:Claude/Gemini 与 OpenAI 结构互转的桥接兼容层,不作为主业务协议转换中心。
|
||||
- `internal/deepseek/{client,protocol,transport}`:上游请求、会话、PoW 适配、协议常量与传输层。
|
||||
- `internal/js/chat-stream` + `api/chat-stream.js`:Vercel Node 流式桥;Go prepare/release 管理鉴权、账号租约和 completion payload,Node 侧负责实时 SSE 转发并保持 Go 对齐的终结态和 tool sieve 语义。
|
||||
- `internal/stream` + `internal/sse`:Go 流式解析与增量处理。
|
||||
|
||||
@@ -68,7 +68,7 @@ gofmt -w <changed-go-files>
|
||||
3. 请求归一化:`internal/promptcompat` 或协议转换包。
|
||||
4. 上游请求:`internal/deepseek/client`。
|
||||
5. 流式输出:`internal/stream`、`internal/sse`、`internal/toolstream`。
|
||||
6. 响应格式:`internal/format/*` 或 `internal/translatorcliproxy`。
|
||||
6. 响应格式:主路径看 `internal/assistantturn` 与 `internal/format/*`;`internal/translatorcliproxy` 只用于 Vercel/fallback/test 桥接。
|
||||
|
||||
对话记录页面问题优先检查:
|
||||
|
||||
|
||||
@@ -108,7 +108,7 @@ DS2API 当前的核心思路,不是把客户端传来的 `messages`、`tools`
|
||||
- 对外返回给客户端的 `prompt_tokens` / `input_tokens` / `promptTokenCount` 不再按“最后一条消息”或字符粗估近似返回,而是基于**完整上下文 prompt**做 tokenizer 计数;为了避免上下文实际超限但客户端误以为还能塞下,请求侧上下文 token 会额外保守上浮一点,宁可略大也不低估。
|
||||
- 当前 `/v1/chat/completions` 业务路径仍是“每次请求新建一个远端 `chat_session_id`,并默认发送 `parent_message_id: null`”;因此 DS2API 对外默认表现为“新会话 + prompt 拼历史”,而不是复用 DeepSeek 原生会话树。
|
||||
- 但 DeepSeek 远端本身支持同一 `chat_session_id` 的跨轮次持续对话。2026-04-27 已用项目内现有 DeepSeek client 做过一次不改业务代码的双轮实测:同一 `chat_session_id` 下,第 1 轮返回 `request_message_id=1` / `response_message_id=2` / 文本 `SESSION_TEST_ONE`;第 2 轮重新获取一次 PoW,并发送 `parent_message_id=2` 后,成功返回 `request_message_id=3` / `response_message_id=4` / 文本 `SESSION_TEST_TWO`。这说明“同远端会话持续聊天”能力存在,且每轮需要携带正确的 parent/message 链接信息,同时重新获取对应轮次可用的 PoW。
|
||||
- OpenAI Chat / Responses 原生走统一 OpenAI 标准化与 DeepSeek payload 组装;Claude / Gemini 会尽量复用 OpenAI prompt/tool 语义,其中 Gemini 直接复用 `promptcompat.BuildOpenAIPromptForAdapter`。Go 主服务新增 `completionruntime` 启动层,统一执行 DeepSeek session/PoW/call;输出侧新增 `assistantturn` 语义层:非流式 OpenAI Chat / Responses / Claude / Gemini 会把 DeepSeek SSE 收集结果先归一成同一份 assistant turn,再分别渲染成各协议原生外形;流式 OpenAI Chat / Responses / Claude / Gemini 继续保持各协议实时 SSE framing,但最终收尾的 tool fallback、schema 归一、usage、empty-output / content-filter 错误语义同样由 `assistantturn` 判定。Claude / Gemini 的常规 Go 主路径不再依赖内部 `httptest` 转发到 OpenAI handler;`translatorcliproxy` 仍保留用于 Vercel bridge、兼容工具和回归测试。
|
||||
- OpenAI Chat / Responses 原生走统一 OpenAI 标准化与 DeepSeek payload 组装;Claude / Gemini 会尽量复用 OpenAI prompt/tool 语义,其中 Gemini 直接复用 `promptcompat.BuildOpenAIPromptForAdapter`。Go 主服务新增 `completionruntime` 启动层,统一执行 DeepSeek session/PoW/call;输出侧新增 `assistantturn` 语义层:非流式 OpenAI Chat / Responses / Claude / Gemini 会把 DeepSeek SSE 收集结果先归一成同一份 assistant turn,再分别渲染成各协议原生外形;流式 OpenAI Chat / Responses / Claude / Gemini 继续保持各协议实时 SSE framing,但最终收尾的 tool fallback、schema 归一、usage、empty-output / content-filter 错误语义同样由 `assistantturn` 判定。Claude / Gemini 的常规 Go 主路径不再依赖内部 `httptest` 转发到 OpenAI handler;`translatorcliproxy` 仅保留用于 Vercel bridge、后端缺失 fallback 和回归测试,不作为主业务协议转换中心。
|
||||
- Vercel Node 流式路径本轮不迁移,仍使用现有 Node bridge / stream-tool-sieve 实现;后续若变更 Node 流式语义,需要按 `assistantturn` 的 Go canonical 输出语义同步对齐。
|
||||
- 客户端传入的 thinking / reasoning 开关会被归一到下游 `thinking_enabled`。Gemini `generationConfig.thinkingConfig.thinkingBudget` 会翻译成同一套 thinking 开关;关闭时即使上游返回 `response/thinking_content`,兼容层也不会把它当作可见正文输出。若最终解析出的模型名带 `-nothinking` 后缀,则会无条件强制关闭 thinking,优先级高于请求体中的 `thinking` / `reasoning` / `reasoning_effort`。未显式关闭时,各 surface 会按解析后的 DeepSeek 模型默认能力开启 thinking,并用各自协议的原生形态暴露:OpenAI Chat 为 `reasoning_content`,OpenAI Responses 为 `response.reasoning.delta` / `reasoning` content,Claude 为 `thinking` block / `thinking_delta`,Gemini 为 `thought: true` part。
|
||||
- 对 OpenAI Chat / Responses 的非流式收尾,如果最终可见正文为空,兼容层会优先尝试把思维链中的独立 DSML / XML 工具块当作真实工具调用解析出来。流式链路也会在收尾阶段做同样的 fallback 检测,但不会因为思维链内容去中途拦截或改写流式输出;真正的工具识别始终基于原始上游文本,而不是基于“已经做过可见输出清洗”的版本,因此即使最终可见层会剥离完整 leaked DSML / XML `tool_calls` wrapper、并抑制全空参数或无效 wrapper 块,也不会影响真实工具调用转成结构化 `tool_calls` / `function_call`。补发结果会作为本轮 assistant 的结构化 `tool_calls` / `function_call` 输出返回,而不是塞进 `content` 文本;如果客户端没有开启 thinking / reasoning,思维链只用于检测,不会作为 `reasoning_content` 或可见正文暴露。只有正文为空且思维链里也没有可执行工具调用时,才继续按空回复错误处理。
|
||||
|
||||
Reference in New Issue
Block a user