Blame view

api/routes/search.py 14.9 KB
be52af70   tangwang   first commit
1
2
3
4
  """
  Search API routes.
  """
  
16c42787   tangwang   feat: implement r...
5
  from fastapi import APIRouter, HTTPException, Query, Request
be52af70   tangwang   first commit
6
  from typing import Optional
16c42787   tangwang   feat: implement r...
7
  import uuid
28e57bb1   tangwang   日志体系优化
8
9
10
  import hashlib
  import json
  import logging
be52af70   tangwang   first commit
11
12
13
14
15
  
  from ..models import (
      SearchRequest,
      ImageSearchRequest,
      SearchResponse,
6aa246be   tangwang   问题:Pydantic 应该能自动...
16
      SearchSuggestResponse,
be52af70   tangwang   first commit
17
18
19
      DocumentResponse,
      ErrorResponse
  )
16c42787   tangwang   feat: implement r...
20
  from context.request_context import create_request_context, set_current_request_context, clear_current_request_context
be52af70   tangwang   first commit
21
22
  
  router = APIRouter(prefix="/search", tags=["search"])
28e57bb1   tangwang   日志体系优化
23
24
25
26
27
28
29
30
31
  backend_verbose_logger = logging.getLogger("backend.verbose")
  
  
  def _log_backend_verbose(payload: dict) -> None:
      if not backend_verbose_logger.handlers:
          return
      backend_verbose_logger.info(
          json.dumps(payload, ensure_ascii=False, separators=(",", ":"))
      )
be52af70   tangwang   first commit
32
33
  
  
16c42787   tangwang   feat: implement r...
34
35
36
37
38
  def extract_request_info(request: Request) -> tuple[str, str]:
      """Extract request ID and user ID from HTTP request"""
      # Try to get request ID from headers
      reqid = request.headers.get('X-Request-ID') or str(uuid.uuid4())[:8]
  
99bea633   tangwang   add logs
39
40
      # Try to get user ID from headers; if not found, use "-1" for correlation
      uid = request.headers.get('X-User-ID') or request.headers.get('User-ID') or "-1"
16c42787   tangwang   feat: implement r...
41
42
43
44
  
      return reqid, uid
  
  
be52af70   tangwang   first commit
45
  @router.post("/", response_model=SearchResponse)
16c42787   tangwang   feat: implement r...
46
  async def search(request: SearchRequest, http_request: Request):
be52af70   tangwang   first commit
47
      """
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
48
      Execute text search query (外部友好格式).
be52af70   tangwang   first commit
49
50
51
  
      Supports:
      - Multi-language query processing
bd96cead   tangwang   1. 动态多语言字段与统一策略配置
52
      - Unified text retrieval strategy (no boolean AST parsing)
be52af70   tangwang   first commit
53
54
      - Semantic search with embeddings
      - Custom ranking functions
6aa246be   tangwang   问题:Pydantic 应该能自动...
55
56
      - Exact match filters and range filters
      - Faceted search
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
57
58
      
      Requires tenant_id in header (X-Tenant-ID) or query parameter (tenant_id).
be52af70   tangwang   first commit
59
      """
16c42787   tangwang   feat: implement r...
60
61
      reqid, uid = extract_request_info(http_request)
  
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
      # Extract tenant_id (required)
      tenant_id = http_request.headers.get('X-Tenant-ID')
      if not tenant_id:
          # Try to get from query string
          from urllib.parse import parse_qs
          query_string = http_request.url.query
          if query_string:
              params = parse_qs(query_string)
              tenant_id = params.get('tenant_id', [None])[0]
      
      if not tenant_id:
          raise HTTPException(
              status_code=400,
              detail="tenant_id is required. Provide it via header 'X-Tenant-ID' or query parameter 'tenant_id'"
          )
  
16c42787   tangwang   feat: implement r...
78
79
80
81
82
      # Create request context
      context = create_request_context(reqid=reqid, uid=uid)
  
      # Set context in thread-local storage
      set_current_request_context(context)
be52af70   tangwang   first commit
83
84
  
      try:
99bea633   tangwang   add logs
85
86
87
          # Log request start (English logs, with key search parameters)
          client_ip = http_request.client.host if http_request.client else "unknown"
          user_agent = http_request.headers.get("User-Agent", "unknown")[:200]
16c42787   tangwang   feat: implement r...
88
          context.logger.info(
99bea633   tangwang   add logs
89
90
91
92
93
94
95
96
97
98
              "Received search request | "
              f"Tenant: {tenant_id} | "
              f"Query: {request.query} | "
              f"IP: {client_ip} | "
              f"User agent: {user_agent} | "
              f"size: {request.size} | from: {request.from_} | "
              f"sort_by: {request.sort_by} | sort_order: {request.sort_order} | "
              f"min_score: {request.min_score} | "
              f"language: {request.language} | "
              f"debug: {request.debug} | "
ff32d894   tangwang   rerank
99
100
101
              f"enable_rerank: {request.enable_rerank} | "
              f"rerank_query_template: {request.rerank_query_template} | "
              f"rerank_doc_template: {request.rerank_doc_template} | "
99bea633   tangwang   add logs
102
103
104
105
              f"sku_filter_dimension: {request.sku_filter_dimension} | "
              f"filters: {request.filters} | "
              f"range_filters: {request.range_filters} | "
              f"facets: {request.facets}",
16c42787   tangwang   feat: implement r...
106
107
108
              extra={'reqid': context.reqid, 'uid': context.uid}
          )
  
be52af70   tangwang   first commit
109
          # Get searcher from app state
bb3c5ef8   tangwang   灌入数据流程跑通
110
          from api.app import get_searcher
be52af70   tangwang   first commit
111
112
          searcher = get_searcher()
  
16c42787   tangwang   feat: implement r...
113
          # Execute search with context (using backend defaults from config)
be52af70   tangwang   first commit
114
115
          result = searcher.search(
              query=request.query,
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
116
              tenant_id=tenant_id,
be52af70   tangwang   first commit
117
118
119
              size=request.size,
              from_=request.from_,
              filters=request.filters,
6aa246be   tangwang   问题:Pydantic 应该能自动...
120
121
              range_filters=request.range_filters,
              facets=request.facets,
16c42787   tangwang   feat: implement r...
122
              min_score=request.min_score,
c86c8237   tangwang   支持聚合。过滤项补充了逻辑,但是有问题
123
              context=context,
c86c8237   tangwang   支持聚合。过滤项补充了逻辑,但是有问题
124
              sort_by=request.sort_by,
1f071951   tangwang   补充调试信息,记录包括各个阶段的 ...
125
              sort_order=request.sort_order,
577ec972   tangwang   返回给前端的字段、格式适配。主要包...
126
127
              debug=request.debug,
              language=request.language,
ca91352a   tangwang   更新文档
128
              sku_filter_dimension=request.sku_filter_dimension,
ff32d894   tangwang   rerank
129
130
131
              enable_rerank=request.enable_rerank,
              rerank_query_template=request.rerank_query_template,
              rerank_doc_template=request.rerank_doc_template,
be52af70   tangwang   first commit
132
133
          )
  
16c42787   tangwang   feat: implement r...
134
135
136
          # Include performance summary in response
          performance_summary = context.get_summary() if context else None
  
be52af70   tangwang   first commit
137
          # Convert to response model
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
138
          response = SearchResponse(
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
139
              results=result.results,
be52af70   tangwang   first commit
140
141
142
              total=result.total,
              max_score=result.max_score,
              took_ms=result.took_ms,
6aa246be   tangwang   问题:Pydantic 应该能自动...
143
              facets=result.facets,
16c42787   tangwang   feat: implement r...
144
              query_info=result.query_info,
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
145
146
              suggestions=result.suggestions,
              related_searches=result.related_searches,
1f071951   tangwang   补充调试信息,记录包括各个阶段的 ...
147
148
              performance_info=performance_summary,
              debug_info=result.debug_info
be52af70   tangwang   first commit
149
150
          )
  
28e57bb1   tangwang   日志体系优化
151
152
153
154
155
          response_payload = response.model_dump(mode="json")
          response_json = json.dumps(response_payload, ensure_ascii=False, separators=(",", ":"))
          response_digest = hashlib.sha256(response_json.encode("utf-8")).hexdigest()[:16]
          max_score = float(response.max_score or 0.0)
  
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
156
          context.logger.info(
28e57bb1   tangwang   日志体系优化
157
158
159
160
161
162
              "Search response | Total results: %s | Max score: %.4f | Time: %sms | payload_size: %s chars | digest: %s",
              response.total,
              max_score,
              response.took_ms,
              len(response_json),
              response_digest,
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
163
164
              extra={'reqid': context.reqid, 'uid': context.uid}
          )
28e57bb1   tangwang   日志体系优化
165
166
167
168
169
170
171
172
173
174
175
176
          _log_backend_verbose({
              "event": "search_response",
              "reqid": context.reqid,
              "uid": context.uid,
              "tenant_id": tenant_id,
              "total_results": response.total,
              "max_score": max_score,
              "took_ms": response.took_ms,
              "payload_size_chars": len(response_json),
              "sha256_16": response_digest,
              "response": response_payload,
          })
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
177
178
179
  
          return response
  
be52af70   tangwang   first commit
180
      except Exception as e:
16c42787   tangwang   feat: implement r...
181
182
183
184
          # Log error in context
          if context:
              context.set_error(e)
              context.logger.error(
99bea633   tangwang   add logs
185
                  f"Search request failed | error: {str(e)}",
16c42787   tangwang   feat: implement r...
186
187
                  extra={'reqid': context.reqid, 'uid': context.uid}
              )
be52af70   tangwang   first commit
188
          raise HTTPException(status_code=500, detail=str(e))
16c42787   tangwang   feat: implement r...
189
190
191
      finally:
          # Clear thread-local context
          clear_current_request_context()
be52af70   tangwang   first commit
192
193
194
  
  
  @router.post("/image", response_model=SearchResponse)
16c42787   tangwang   feat: implement r...
195
  async def search_by_image(request: ImageSearchRequest, http_request: Request):
be52af70   tangwang   first commit
196
      """
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
197
      Search by image similarity (外部友好格式).
be52af70   tangwang   first commit
198
199
  
      Uses image embeddings to find visually similar products.
6aa246be   tangwang   问题:Pydantic 应该能自动...
200
      Supports exact match filters and range filters.
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
201
202
      
      Requires tenant_id in header (X-Tenant-ID) or query parameter (tenant_id).
be52af70   tangwang   first commit
203
      """
16c42787   tangwang   feat: implement r...
204
205
      reqid, uid = extract_request_info(http_request)
  
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
      # Extract tenant_id (required)
      tenant_id = http_request.headers.get('X-Tenant-ID')
      if not tenant_id:
          from urllib.parse import parse_qs
          query_string = http_request.url.query
          if query_string:
              params = parse_qs(query_string)
              tenant_id = params.get('tenant_id', [None])[0]
      
      if not tenant_id:
          raise HTTPException(
              status_code=400,
              detail="tenant_id is required. Provide it via header 'X-Tenant-ID' or query parameter 'tenant_id'"
          )
  
16c42787   tangwang   feat: implement r...
221
222
223
224
225
226
      # Create request context
      context = create_request_context(reqid=reqid, uid=uid)
  
      # Set context in thread-local storage
      set_current_request_context(context)
  
be52af70   tangwang   first commit
227
      try:
99bea633   tangwang   add logs
228
229
          # Log request start for image search (English)
          client_ip = http_request.client.host if http_request.client else "unknown"
16c42787   tangwang   feat: implement r...
230
          context.logger.info(
99bea633   tangwang   add logs
231
232
233
234
              "Received image search request | "
              f"Tenant: {tenant_id} | "
              f"Image URL: {request.image_url} | "
              f"IP: {client_ip}",
16c42787   tangwang   feat: implement r...
235
236
237
              extra={'reqid': context.reqid, 'uid': context.uid}
          )
  
bb3c5ef8   tangwang   灌入数据流程跑通
238
          from api.app import get_searcher
be52af70   tangwang   first commit
239
240
241
242
243
          searcher = get_searcher()
  
          # Execute image search
          result = searcher.search_by_image(
              image_url=request.image_url,
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
244
              tenant_id=tenant_id,
be52af70   tangwang   first commit
245
              size=request.size,
6aa246be   tangwang   问题:Pydantic 应该能自动...
246
247
              filters=request.filters,
              range_filters=request.range_filters
be52af70   tangwang   first commit
248
249
          )
  
16c42787   tangwang   feat: implement r...
250
251
252
          # Include performance summary in response
          performance_summary = context.get_summary() if context else None
  
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
253
          response = SearchResponse(
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
254
              results=result.results,
be52af70   tangwang   first commit
255
256
257
              total=result.total,
              max_score=result.max_score,
              took_ms=result.took_ms,
6aa246be   tangwang   问题:Pydantic 应该能自动...
258
              facets=result.facets,
16c42787   tangwang   feat: implement r...
259
              query_info=result.query_info,
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
260
261
              suggestions=result.suggestions,
              related_searches=result.related_searches,
16c42787   tangwang   feat: implement r...
262
              performance_info=performance_summary
be52af70   tangwang   first commit
263
264
          )
  
28e57bb1   tangwang   日志体系优化
265
266
267
268
269
          response_payload = response.model_dump(mode="json")
          response_json = json.dumps(response_payload, ensure_ascii=False, separators=(",", ":"))
          response_digest = hashlib.sha256(response_json.encode("utf-8")).hexdigest()[:16]
          max_score = float(response.max_score or 0.0)
  
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
270
          context.logger.info(
28e57bb1   tangwang   日志体系优化
271
272
273
274
275
276
              "Image search response | Total results: %s | Max score: %.4f | Time: %sms | payload_size: %s chars | digest: %s",
              response.total,
              max_score,
              response.took_ms,
              len(response_json),
              response_digest,
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
277
278
              extra={'reqid': context.reqid, 'uid': context.uid}
          )
28e57bb1   tangwang   日志体系优化
279
280
281
282
283
284
285
286
287
288
289
290
          _log_backend_verbose({
              "event": "image_search_response",
              "reqid": context.reqid,
              "uid": context.uid,
              "tenant_id": tenant_id,
              "total_results": response.total,
              "max_score": max_score,
              "took_ms": response.took_ms,
              "payload_size_chars": len(response_json),
              "sha256_16": response_digest,
              "response": response_payload,
          })
3cd09b3b   tangwang   翻译接口改为调用qwen-mt-f...
291
292
293
  
          return response
  
be52af70   tangwang   first commit
294
      except ValueError as e:
16c42787   tangwang   feat: implement r...
295
296
297
          if context:
              context.set_error(e)
              context.logger.error(
99bea633   tangwang   add logs
298
                  f"Image search request parameter error | error: {str(e)}",
16c42787   tangwang   feat: implement r...
299
300
                  extra={'reqid': context.reqid, 'uid': context.uid}
              )
be52af70   tangwang   first commit
301
302
          raise HTTPException(status_code=400, detail=str(e))
      except Exception as e:
16c42787   tangwang   feat: implement r...
303
304
305
          if context:
              context.set_error(e)
              context.logger.error(
99bea633   tangwang   add logs
306
                  f"Image search request failed | error: {str(e)}",
16c42787   tangwang   feat: implement r...
307
308
                  extra={'reqid': context.reqid, 'uid': context.uid}
              )
be52af70   tangwang   first commit
309
          raise HTTPException(status_code=500, detail=str(e))
16c42787   tangwang   feat: implement r...
310
311
312
      finally:
          # Clear thread-local context
          clear_current_request_context()
be52af70   tangwang   first commit
313
314
  
  
6aa246be   tangwang   问题:Pydantic 应该能自动...
315
316
317
  @router.get("/suggestions", response_model=SearchSuggestResponse)
  async def search_suggestions(
      q: str = Query(..., min_length=1, description="搜索查询"),
ff9efda0   tangwang   suggest
318
      size: int = Query(10, ge=1, le=50, description="建议数量(1-50)"),
ded6f29e   tangwang   补充suggestion模块
319
      language: str = Query("en", description="请求语言,如 zh/en/ar/ru"),
ded6f29e   tangwang   补充suggestion模块
320
321
      debug: bool = Query(False, description="是否返回调试信息"),
      http_request: Request = None,
6aa246be   tangwang   问题:Pydantic 应该能自动...
322
323
324
325
  ):
      """
      获取搜索建议(自动补全)。
      
ff9efda0   tangwang   suggest
326
      获取搜索建议(自动补全,支持多语言)。
6aa246be   tangwang   问题:Pydantic 应该能自动...
327
      """
ded6f29e   tangwang   补充suggestion模块
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
      # Extract tenant_id (required)
      tenant_id = http_request.headers.get("X-Tenant-ID") if http_request else None
      if not tenant_id and http_request:
          from urllib.parse import parse_qs
          query_string = http_request.url.query
          if query_string:
              params = parse_qs(query_string)
              tenant_id = params.get("tenant_id", [None])[0]
  
      if not tenant_id:
          raise HTTPException(
              status_code=400,
              detail="tenant_id is required. Provide it via header 'X-Tenant-ID' or query parameter 'tenant_id'",
          )
  
      try:
          from api.app import get_suggestion_service
  
          service = get_suggestion_service()
          result = service.search(
              tenant_id=tenant_id,
              query=q,
              language=language,
              size=size,
ded6f29e   tangwang   补充suggestion模块
352
353
354
355
356
357
358
359
360
361
362
363
364
365
          )
          response = SearchSuggestResponse(
              query=result["query"],
              language=result.get("language"),
              resolved_language=result.get("resolved_language"),
              suggestions=result["suggestions"],
              took_ms=result["took_ms"],
          )
          if debug:
              # keep response_model stable; debug info stays inside suggestions payload for now
              return response
          return response
      except Exception as e:
          raise HTTPException(status_code=500, detail=str(e))
6aa246be   tangwang   问题:Pydantic 应该能自动...
366
367
368
369
370
  
  
  @router.get("/instant", response_model=SearchResponse)
  async def instant_search(
      q: str = Query(..., min_length=2, description="搜索查询"),
1f6d15fa   tangwang   重构:SPU级别索引、统一索引架构...
371
      size: int = Query(5, ge=1, le=20, description="结果数量"),
6aa246be   tangwang   问题:Pydantic 应该能自动...
372
373
374
375
376
377
378
379
  ):
      """
      即时搜索(Instant Search)。
      
      功能说明:
      - 边输入边搜索,无需点击搜索按钮
      - 返回简化的搜索结果
      
26b910bd   tangwang   refactor service ...
380
      注意:此功能暂未开放,当前明确返回 501
6aa246be   tangwang   问题:Pydantic 应该能自动...
381
      """
26b910bd   tangwang   refactor service ...
382
383
384
385
386
387
388
      # 明确暴露当前接口尚未完成实现,避免调用不完整逻辑导致隐式运行时错误。
      raise HTTPException(
          status_code=501,
          detail=(
              "/search/instant is not implemented yet. "
              "Use POST /search/ for production traffic."
          ),
6aa246be   tangwang   问题:Pydantic 应该能自动...
389
390
391
      )
  
  
be52af70   tangwang   first commit
392
  @router.get("/{doc_id}", response_model=DocumentResponse)
e4a39cc8   tangwang   索引隔离。 不同的tenant_i...
393
  async def get_document(doc_id: str, http_request: Request):
be52af70   tangwang   first commit
394
395
      """
      Get a single document by ID.
e4a39cc8   tangwang   索引隔离。 不同的tenant_i...
396
397
      
      Requires tenant_id in header (X-Tenant-ID) or query parameter (tenant_id).
be52af70   tangwang   first commit
398
399
      """
      try:
e4a39cc8   tangwang   索引隔离。 不同的tenant_i...
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
          # Extract tenant_id (required)
          tenant_id = http_request.headers.get('X-Tenant-ID')
          if not tenant_id:
              # Try to get from query string
              from urllib.parse import parse_qs
              query_string = http_request.url.query
              if query_string:
                  params = parse_qs(query_string)
                  tenant_id = params.get('tenant_id', [None])[0]
          
          if not tenant_id:
              raise HTTPException(
                  status_code=400,
                  detail="tenant_id is required. Provide it via header 'X-Tenant-ID' or query parameter 'tenant_id'"
              )
          
bb3c5ef8   tangwang   灌入数据流程跑通
416
          from api.app import get_searcher
be52af70   tangwang   first commit
417
418
          searcher = get_searcher()
  
e4a39cc8   tangwang   索引隔离。 不同的tenant_i...
419
          doc = searcher.get_document(tenant_id=tenant_id, doc_id=doc_id)
be52af70   tangwang   first commit
420
421
  
          if doc is None:
e4a39cc8   tangwang   索引隔离。 不同的tenant_i...
422
              raise HTTPException(status_code=404, detail=f"Document {doc_id} not found for tenant {tenant_id}")
be52af70   tangwang   first commit
423
424
425
426
427
428
429
  
          return DocumentResponse(id=doc_id, source=doc)
  
      except HTTPException:
          raise
      except Exception as e:
          raise HTTPException(status_code=500, detail=str(e))