lingyue-zhibao/python-services/ner-service/app/routers/ner.py

"""
NER 路由
"""
import json
import asyncio
from fastapi import APIRouter, HTTPException
from fastapi.responses import StreamingResponse
from loguru import logger
import time

from ..models import NerRequest, NerResponse, EntityInfo
from ..services.ner_service import ner_service

router = APIRouter()


async def sse_event(event: str, data: dict):
    """生成 SSE 事件格式"""
    return f"event: {event}\ndata: {json.dumps(data, ensure_ascii=False)}\n\n"


@router.post("/extract", response_model=NerResponse)
async def extract_entities(request: NerRequest):
    """
    从文本中提取命名实体
    """
    start_time = time.time()
    
    try:
        logger.info(f"开始提取实体: document_id={request.document_id}, text_length={len(request.text)}")
        
        # 验证文本长度
        if len(request.text) > 50000:
            raise HTTPException(status_code=400, detail="文本长度超过限制（最大50000字符）")
        
        # 调用 NER 服务
        entities = await ner_service.extract_entities(
            text=request.text,
            entity_types=request.entity_types
        )
        
        # 如果需要提取关系
        relations = []
        if request.extract_relations and len(entities) > 1:
            from ..services.relation_service import relation_service
            relations = await relation_service.extract_relations(
                text=request.text,
                entities=entities
            )
        
        processing_time = int((time.time() - start_time) * 1000)
        
        logger.info(f"实体提取完成: document_id={request.document_id}, "
                   f"entity_count={len(entities)}, relation_count={len(relations)}, "
                   f"processing_time={processing_time}ms")
        
        # 输出完整的实体列表
        logger.info(f"========== 实体列表 ({len(entities)} 个) ==========")
        for i, entity in enumerate(entities, 1):
            logger.info(f"  [{i}] {entity.type}: {entity.name}")
        
        # 输出完整的关系列表
        if relations:
            logger.info(f"========== 关系列表 ({len(relations)} 个) ==========")
            for i, rel in enumerate(relations, 1):
                logger.info(f"  [{i}] {rel.source_entity} --[{rel.relation_type}]--> {rel.target_entity}")
        
        return NerResponse.success_response(
            document_id=request.document_id,
            entities=entities,
            relations=relations,
            processing_time=processing_time
        )
        
    except HTTPException:
        raise
    except Exception as e:
        logger.error(f"实体提取失败: document_id={request.document_id}, error={str(e)}")
        return NerResponse.error_response(
            document_id=request.document_id,
            error_message=str(e)
        )


@router.post("/extract/stream")
async def extract_entities_stream(request: NerRequest):
    """
    从文本中提取命名实体（SSE 流式响应）
    
    实时推送进度事件：
    - progress: 处理进度（分块处理时）
    - entity: 发现新实体
    - complete: 处理完成
    - error: 处理出错
    """
    async def generate():
        start_time = time.time()
        
        try:
            # 发送开始事件
            yield await sse_event("start", {
                "document_id": request.document_id,
                "text_length": len(request.text),
                "message": "开始 NER 提取"
            })
            
            # 验证文本长度
            if len(request.text) > 50000:
                yield await sse_event("error", {
                    "document_id": request.document_id,
                    "error": "文本长度超过限制（最大50000字符）"
                })
                return
            
            # 使用带进度回调的提取方法
            all_entities = []
            all_relations = []
            
            # 调用带进度的 NER 服务
            from ..services.ner_service import ner_service
            
            # 检查是否支持流式提取
            if hasattr(ner_service, 'extract_entities_with_progress'):
                async for event_str in ner_service.extract_entities_with_progress(
                    text=request.text,
                    entity_types=request.entity_types
                ):
                    # 转发进度事件
                    yield event_str
                    
                    # 解析事件获取实体数据
                    if "entities_data" in event_str:
                        try:
                            # 从 SSE 格式中提取 JSON 数据
                            lines = event_str.strip().split('\n')
                            for line in lines:
                                if line.startswith('data:'):
                                    data = json.loads(line[5:].strip())
                                    if 'entities' in data:
                                        # 将字典转换回 EntityInfo 对象
                                        all_entities = [
                                            EntityInfo(**e) for e in data['entities']
                                        ]
                        except Exception as parse_err:
                            logger.warning(f"解析实体数据事件失败: {parse_err}")
            else:
                # 回退到普通提取
                all_entities = await ner_service.extract_entities(
                    text=request.text,
                    entity_types=request.entity_types
                )
                yield await sse_event("progress", {
                    "document_id": request.document_id,
                    "chunk_index": 1,
                    "total_chunks": 1,
                    "total_entities": len(all_entities),
                    "progress_percent": 100
                })
            
            # 提取关系
            if request.extract_relations and len(all_entities) > 1:
                yield await sse_event("progress", {
                    "document_id": request.document_id,
                    "message": "正在提取实体关系...",
                    "stage": "relations"
                })
                
                from ..services.relation_service import relation_service
                all_relations = await relation_service.extract_relations(
                    text=request.text,
                    entities=all_entities
                )
            
            processing_time = int((time.time() - start_time) * 1000)
            
            # 发送完成事件（包含完整结果）
            response = NerResponse.success_response(
                document_id=request.document_id,
                entities=all_entities,
                relations=all_relations,
                processing_time=processing_time
            )
            
            yield await sse_event("complete", response.dict())
            
            logger.info(f"SSE 提取完成: document_id={request.document_id}, "
                       f"entity_count={len(all_entities)}, relation_count={len(all_relations)}, "
                       f"processing_time={processing_time}ms")
            
        except Exception as e:
            logger.error(f"SSE 提取失败: document_id={request.document_id}, error={str(e)}")
            yield await sse_event("error", {
                "document_id": request.document_id,
                "error": str(e)
            })
    
    return StreamingResponse(
        generate(),
        media_type="text/event-stream",
        headers={
            "Cache-Control": "no-cache",
            "Connection": "keep-alive",
            "X-Accel-Buffering": "no"
        }
    )