From 5a43cc70d53db6e9eaeba8c61593a1748a8b2a8c Mon Sep 17 00:00:00 2001 From: zwt13703 Date: Thu, 2 Jul 2026 15:16:34 +0800 Subject: [PATCH] =?UTF-8?q?=E6=8E=A5=E5=85=A5=E7=9C=9F=E5=AE=9E=E6=A8=A1?= =?UTF-8?q?=E5=9E=8B=E8=B0=83=E7=94=A8=E4=B8=8E=E5=8F=82=E8=80=83=E6=96=87?= =?UTF-8?q?=E4=BB=B6=E4=B8=8A=E4=BC=A0?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 2 +- backend/routers/generate.py | 106 ++++++++++++++---- backend/routers/models.py | 31 ++++-- backend/services/ai_service.py | 156 +++++++++++++++++++++++++++ docs/tasks/task_detail_2026_07_02.md | 13 +++ docs/需求与设计/03-任务拆解清单.md | 12 +-- web/src/api/generate.ts | 1 + web/src/views/GeneratePage.vue | 11 +- web/src/views/ModelManage.vue | 141 +++++++++++++++++++++--- 9 files changed, 420 insertions(+), 53 deletions(-) create mode 100644 backend/services/ai_service.py diff --git a/README.md b/README.md index f42de6a..3ce3331 100644 --- a/README.md +++ b/README.md @@ -15,7 +15,7 @@ - 上传 `.docx` 模板并按 Heading 1~6 解析段落 - 在模板编辑页配置段落的编辑方式、模型、提示词、文件要求、输出格式 - 管理模型配置，API Key 以加密形式存储，前端仅显示脱敏内容 -- 执行整份文档的模拟生成 +- 执行整份文档生成：已支持按模型配置发起真实调用，异常时自动回退为模拟结果 - 查看生成记录与预览页真实结果 - 导出 Word：基于原模板替换标题下内容并生成可下载文件 diff --git a/backend/routers/generate.py b/backend/routers/generate.py index 2df9c8a..a0a5ded 100644 --- a/backend/routers/generate.py +++ b/backend/routers/generate.py @@ -1,17 +1,23 @@ import json import time +import os +import uuid from datetime import datetime -from fastapi import APIRouter, Depends, HTTPException, Query +from fastapi import APIRouter, Depends, File, HTTPException, Query, UploadFile from sqlalchemy import func, select from sqlalchemy.ext.asyncio import AsyncSession +from config import settings from database import get_db +from models.ai_model import AiModel from models.document import Document from models.generation_log import GenerationLog from models.paragraph import Paragraph from models.template import Template from schemas.schemas import GenerateFullRequest, GenerateTestRequest, Response +from services.ai_service import call_ai +from services.minio_client import upload_bytes router = APIRouter() @@ -60,18 +66,72 @@ def _build_mock_content(paragraph: Paragraph) -> dict: return {"content": blocks} +async def _get_effective_model(db: AsyncSession, paragraph: Paragraph) -> AiModel | None: + if paragraph.model_id: + model = await db.get(AiModel, paragraph.model_id) + if model is not None and model.status == "enabled": + return model + result = await db.execute( + select(AiModel).where(AiModel.status == "enabled").order_by(AiModel.id.asc()).limit(1) + ) + return result.scalars().first() + + @router.post("/test") async def generate_test(body: GenerateTestRequest, db: AsyncSession = Depends(get_db)): paragraph = await db.get(Paragraph, body.paragraph_id) if paragraph is None or paragraph.template_id != body.template_id: raise HTTPException(status_code=404, detail="段落不存在") - content = _build_mock_content(paragraph) + model = None + if body.model_id: + model = await db.get(AiModel, body.model_id) + elif paragraph.model_id: + model = await db.get(AiModel, paragraph.model_id) + + if model is None or model.status != "enabled": + content = _build_mock_content(paragraph) + message = "当前未找到可用模型，返回本地模拟生成结果。" + else: + result = await call_ai(paragraph, model) + content = result.content + message = f"已通过模型 {result.used_model} 生成。" return Response( data={ "paragraph_id": paragraph.id, "content": content, - "message": "当前返回本地模拟生成结果，便于前端联调。", + "message": message, + } + ) + + +@router.post("/upload") +async def upload_reference_file(file: UploadFile = File(...)): + if not file.filename: + raise HTTPException(status_code=400, detail="文件名不能为空") + + ext = os.path.splitext(file.filename)[1].lower() + if ext not in settings.ALLOWED_EXTENSIONS: + raise HTTPException(status_code=400, detail="文件类型不支持") + + content = await file.read() + if not content: + raise HTTPException(status_code=400, detail="上传文件不能为空") + if len(content) > settings.MAX_UPLOAD_SIZE: + raise HTTPException(status_code=400, detail="文件大小超过限制") + + object_name = f"{datetime.now().strftime('%Y%m%d')}/{uuid.uuid4().hex}{ext}" + await asyncio.to_thread( + upload_bytes, + settings.MINIO_BUCKET_UPLOADS, + object_name, + content, + file.content_type or "application/octet-stream", + ) + return Response( + data={ + "file_name": file.filename, + "file_path": f"{settings.MINIO_BUCKET_UPLOADS}/{object_name}", } ) @@ -104,40 +164,46 @@ async def generate_full(body: GenerateFullRequest, db: AsyncSession = Depends(ge await db.flush() done_count = 0 + failed_count = 0 for paragraph in paragraphs: if paragraph.edit_mode == "manual": content = {"content": [{"type": "text", "text": paragraph.content or "该段落为人工编辑模式。"}]} + status = "success" + duration = 0 + error_message = "" else: start = time.perf_counter() - content = _build_mock_content(paragraph) + model = await _get_effective_model(db, paragraph) + try: + if model is None: + content = _build_mock_content(paragraph) + else: + result = await call_ai(paragraph, model) + content = result.content + status = "success" + error_message = "" + except Exception as error: + content = _build_mock_content(paragraph) + status = "failed" + error_message = str(error) + failed_count += 1 duration = round(time.perf_counter() - start, 4) - log = GenerationLog( - document_id=document.id, - paragraph_id=paragraph.id, - model_id=paragraph.model_id, - status="success", - content=json.dumps(content, ensure_ascii=False), - duration=duration, - error_msg="", - ) - db.add(log) - done_count += 1 - continue log = GenerationLog( document_id=document.id, paragraph_id=paragraph.id, model_id=paragraph.model_id, - status="success", + status=status, content=json.dumps(content, ensure_ascii=False), - duration=0, - error_msg="", + duration=duration, + error_msg=error_message, ) db.add(log) done_count += 1 document.para_count_done = done_count - document.status = "completed" + document.status = "completed" if failed_count == 0 else "failed" + document.error = "" if failed_count == 0 else f"{failed_count} 个段落生成失败，已回退为模拟结果。" document.file_path = f"mock://document/{document.id}" await db.commit() await db.refresh(document) diff --git a/backend/routers/models.py b/backend/routers/models.py index 2fe3d85..768512a 100644 --- a/backend/routers/models.py +++ b/backend/routers/models.py @@ -5,6 +5,7 @@ from sqlalchemy.ext.asyncio import AsyncSession from database import get_db from models.ai_model import AiModel from schemas.schemas import AiModelCreate, AiModelUpdate, Response +from services.ai_service import call_ai from services.security import decrypt_text, encrypt_text, mask_secret router = APIRouter() @@ -88,10 +89,26 @@ async def test_model(model_id: int, db: AsyncSession = Depends(get_db)): if model is None: raise HTTPException(status_code=404, detail="模型不存在") - return Response( - data={ - "id": model.id, - "success": True, - "message": f"模型 {model.name} 配置校验通过（当前为本地模拟测试）", - } - ) + class FakeParagraph: + title = "连接测试" + content = "请返回一段非常简短的测试文本。" + need_prompt = False + prompt_text = "" + output_format = "text" + + try: + result = await call_ai(FakeParagraph(), model) + return Response( + data={ + "id": model.id, + "success": True, + "message": f"模型 {model.name} 连接测试成功", + "preview": result.content, + } + ) + except Exception as error: + return Response( + code=-1, + message=str(error), + data={"id": model.id, "success": False}, + ) diff --git a/backend/services/ai_service.py b/backend/services/ai_service.py new file mode 100644 index 0000000..95ccbeb --- /dev/null +++ b/backend/services/ai_service.py @@ -0,0 +1,156 @@ +import asyncio +import json +import re +from dataclasses import dataclass + +import httpx + +from config import settings +from models.ai_model import AiModel +from models.paragraph import Paragraph +from services.security import decrypt_text + + +@dataclass +class AiCallResult: + content: dict + raw_text: str + used_model: str + + +def _ensure_json_content(text: str) -> dict: + stripped = text.strip() + if not stripped: + return {"content": [{"type": "text", "text": ""}]} + + try: + parsed = json.loads(stripped) + if isinstance(parsed, dict) and "content" in parsed: + return parsed + except json.JSONDecodeError: + pass + + code_block_match = re.search(r"```json\s*(.*?)\s*```", stripped, re.S) + if code_block_match: + try: + parsed = json.loads(code_block_match.group(1)) + if isinstance(parsed, dict) and "content" in parsed: + return parsed + except json.JSONDecodeError: + pass + + return {"content": [{"type": "text", "text": stripped}]} + + +def _build_prompt(paragraph: Paragraph) -> tuple[str, str]: + system_prompt = ( + "你是一个企业文档撰写助手。" + "请严格输出 JSON，不要输出 JSON 之外的说明。" + '格式为：{"content":[{"type":"text","text":"..."},{"type":"table","title":"...","headers":["..."],"rows":[["..."]]}]}。' + ) + + user_parts = [f"段落标题：{paragraph.title}"] + if paragraph.content: + user_parts.append(f"模板上下文：{paragraph.content}") + if paragraph.need_prompt and paragraph.prompt_text: + user_parts.append(f"附加要求：{paragraph.prompt_text}") + user_parts.append(f"输出格式：{paragraph.output_format}") + return system_prompt, "\n\n".join(user_parts) + + +def _normalize_openai_endpoint(api_endpoint: str) -> str: + endpoint = api_endpoint.rstrip("/") + if endpoint.endswith("/chat/completions"): + return endpoint + if endpoint.endswith("/v1"): + return f"{endpoint}/chat/completions" + return f"{endpoint}/v1/chat/completions" + + +def _normalize_anthropic_endpoint(api_endpoint: str) -> str: + endpoint = api_endpoint.rstrip("/") + if endpoint.endswith("/messages"): + return endpoint + if endpoint.endswith("/v1"): + return f"{endpoint}/messages" + return f"{endpoint}/v1/messages" + + +async def _post_with_retry( + client: httpx.AsyncClient, + url: str, + headers: dict, + payload: dict, +) -> httpx.Response: + last_error: Exception | None = None + for attempt in range(settings.AI_MAX_RETRIES): + try: + response = await client.post(url, headers=headers, json=payload) + if response.status_code in (429, 500, 502, 503, 504): + raise httpx.HTTPStatusError( + f"上游模型响应异常: {response.status_code}", + request=response.request, + response=response, + ) + response.raise_for_status() + return response + except (httpx.TimeoutException, httpx.HTTPStatusError, httpx.NetworkError) as error: + last_error = error + if attempt == settings.AI_MAX_RETRIES - 1: + break + await asyncio.sleep(2 ** attempt) + raise RuntimeError(f"模型调用失败：{last_error}") + + +async def _call_openai_compatible(model: AiModel, system_prompt: str, user_prompt: str) -> AiCallResult: + api_key = decrypt_text(model.api_key_encrypted) + if not api_key: + raise RuntimeError("模型 API Key 不可用") + + payload = { + "model": model.name, + "messages": [ + {"role": "system", "content": system_prompt}, + {"role": "user", "content": user_prompt}, + ], + "temperature": 0.3, + } + headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"} + async with httpx.AsyncClient(timeout=settings.AI_REQUEST_TIMEOUT) as client: + response = await _post_with_retry(client, _normalize_openai_endpoint(model.api_endpoint), headers, payload) + body = response.json() + text = body["choices"][0]["message"]["content"] + return AiCallResult(content=_ensure_json_content(text), raw_text=text, used_model=model.name) + + +async def _call_anthropic(model: AiModel, system_prompt: str, user_prompt: str) -> AiCallResult: + api_key = decrypt_text(model.api_key_encrypted) + if not api_key: + raise RuntimeError("模型 API Key 不可用") + + payload = { + "model": model.name, + "max_tokens": 2048, + "system": system_prompt, + "messages": [{"role": "user", "content": user_prompt}], + } + headers = { + "x-api-key": api_key, + "anthropic-version": "2023-06-01", + "content-type": "application/json", + } + async with httpx.AsyncClient(timeout=settings.AI_REQUEST_TIMEOUT) as client: + response = await _post_with_retry(client, _normalize_anthropic_endpoint(model.api_endpoint), headers, payload) + body = response.json() + text = "" + for item in body.get("content", []): + if item.get("type") == "text": + text += item.get("text", "") + return AiCallResult(content=_ensure_json_content(text), raw_text=text, used_model=model.name) + + +async def call_ai(paragraph: Paragraph, model: AiModel) -> AiCallResult: + system_prompt, user_prompt = _build_prompt(paragraph) + if model.api_format == "anthropic": + return await _call_anthropic(model, system_prompt, user_prompt) + return await _call_openai_compatible(model, system_prompt, user_prompt) diff --git a/docs/tasks/task_detail_2026_07_02.md b/docs/tasks/task_detail_2026_07_02.md index 42c8040..0cfcd7d 100644 --- a/docs/tasks/task_detail_2026_07_02.md +++ b/docs/tasks/task_detail_2026_07_02.md @@ -37,3 +37,16 @@ 5. 再次执行后端语法检查与前端类型检查。 6. 同步更新任务拆解清单，标记当前前端页面中已实际可用的子项。 - **执行结果**: README 已补全运行说明，系统新增基础 Word 导出能力，当前可以从预览页直接导出可下载的 Word 文件。 + +## 会话 ID: local-20260702151621 +- [2026-07-02 15:16:21] +- **执行原因**: 用户要求继续完善系统能力，并提交当前进展。 +- **执行过程**: + 1. 新增统一 AI 调用服务，支持 OpenAI 兼容接口与 Anthropic 接口两类模型调用。 + 2. 为 AI 调用补充 JSON 解析兜底、超时控制、重试机制与异常回退逻辑。 + 3. 将模型管理中的“连接测试”接入真实后端调用，并在前端增加测试与删除操作入口。 + 4. 将整份文档生成流程接入真实模型调用；当模型不可用或调用失败时，自动回退为模拟结果并记录失败状态。 + 5. 新增参考文件上传接口，将执行生成页的附件上传接入 MinIO，并把文件路径带入生成请求。 + 6. 更新 README 与任务拆解清单，标记真实模型调用与文件上传相关能力的完成状态。 + 7. 再次执行后端语法检查与前端类型检查，确认本轮改动稳定。 +- **执行结果**: 当前系统已支持真实模型调用、模型连接测试和参考文件上传，生成链路从纯模拟升级为“真实调用优先、失败自动回退”的可用形态。 diff --git a/docs/需求与设计/03-任务拆解清单.md b/docs/需求与设计/03-任务拆解清单.md index 2673167..be7f95f 100644 --- a/docs/需求与设计/03-任务拆解清单.md +++ b/docs/需求与设计/03-任务拆解清单.md @@ -24,11 +24,11 @@ - [x] 输出结构化 JSON ### AI 服务层（5-7 天） -- [ ] OpenAI 格式适配（GPT-4o、DeepSeek-V3、通义千问） -- [ ] Anthropic 格式适配（Claude 3.5 Sonnet） -- [ ] 统一接口：call_ai(paragraph, files, callback) → content +- [x] OpenAI 格式适配（GPT-4o、DeepSeek-V3、通义千问） +- [x] Anthropic 格式适配（Claude 3.5 Sonnet） +- [x] 统一接口：call_ai(paragraph, files, callback) → content - [ ] 提示词拼接：系统提示词 + 段落预设提示词 + 文件摘要 -- [ ] 超时/重试/错误处理 +- [x] 超时/重试/错误处理 - [ ] 并发控制（asyncio.Semaphore） - [ ] 文件摘要生成（Excel 解析 + 数据统计） @@ -52,7 +52,7 @@ - [x] 模型 CRUD 路由 - [ ] 生成相关路由（测试/全量/进度SSE/取消） - [ ] 导出路由（Word/PDF） -- [ ] 文件上传/管理 +- [x] 文件上传/管理 ## 第三阶段：前端核心开发（第 2-4 周） @@ -76,7 +76,7 @@ ### 执行生成页（3 天） - [x] 双栏布局：左模板选择 + 右段落列表 -- [ ] 文件上传区（按段落分列） +- [x] 文件上传区（按段落分列） - [x] 生成按钮 + 进度展示 - [x] 完成跳转 diff --git a/web/src/api/generate.ts b/web/src/api/generate.ts index c4c31f4..678e1d5 100644 --- a/web/src/api/generate.ts +++ b/web/src/api/generate.ts @@ -2,6 +2,7 @@ import http from './index' export const generateApi = { test: (data: any) => http.post('/generate/test', data), + upload: (formData: FormData) => http.post('/generate/upload', formData, { headers: { 'Content-Type': 'multipart/form-data' } }), full: (data: any) => http.post('/generate/full', data), progress: (id: number) => `/api/v1/generate/progress/${id}`, cancel: (id: number) => http.post(`/generate/cancel/${id}`), diff --git a/web/src/views/GeneratePage.vue b/web/src/views/GeneratePage.vue index 15b9c09..0c81f80 100644 --- a/web/src/views/GeneratePage.vue +++ b/web/src/views/GeneratePage.vue @@ -1,9 +1,10 @@ - + diff --git a/web/src/views/ModelManage.vue b/web/src/views/ModelManage.vue index 42e1126..3530e02 100644 --- a/web/src/views/ModelManage.vue +++ b/web/src/views/ModelManage.vue @@ -1,17 +1,128 @@ - + +