You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

kb_app.py 14KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360
  1. #
  2. # Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
  3. #
  4. # Licensed under the Apache License, Version 2.0 (the "License");
  5. # you may not use this file except in compliance with the License.
  6. # You may obtain a copy of the License at
  7. #
  8. # http://www.apache.org/licenses/LICENSE-2.0
  9. #
  10. # Unless required by applicable law or agreed to in writing, software
  11. # distributed under the License is distributed on an "AS IS" BASIS,
  12. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13. # See the License for the specific language governing permissions and
  14. # limitations under the License.
  15. #
  16. import json
  17. from flask import request
  18. from flask_login import login_required, current_user
  19. from api.db.services import duplicate_name
  20. from api.db.services.document_service import DocumentService
  21. from api.db.services.file2document_service import File2DocumentService
  22. from api.db.services.file_service import FileService
  23. from api.db.services.user_service import TenantService, UserTenantService
  24. from api.utils.api_utils import server_error_response, get_data_error_result, validate_request, not_allowed_parameters
  25. from api.utils import get_uuid
  26. from api.db import StatusEnum, FileSource
  27. from api.db.services.knowledgebase_service import KnowledgebaseService
  28. from api.db.db_models import File
  29. from api.utils.api_utils import get_json_result
  30. from api import settings
  31. from rag.nlp import search
  32. from api.constants import DATASET_NAME_LIMIT
  33. from rag.settings import PAGERANK_FLD
  34. from rag.utils.storage_factory import STORAGE_IMPL
  35. @manager.route('/create', methods=['post']) # noqa: F821
  36. @login_required
  37. @validate_request("name")
  38. def create():
  39. req = request.json
  40. dataset_name = req["name"]
  41. if not isinstance(dataset_name, str):
  42. return get_data_error_result(message="Dataset name must be string.")
  43. if dataset_name.strip() == "":
  44. return get_data_error_result(message="Dataset name can't be empty.")
  45. if len(dataset_name.encode("utf-8")) > DATASET_NAME_LIMIT:
  46. return get_data_error_result(
  47. message=f"Dataset name length is {len(dataset_name)} which is larger than {DATASET_NAME_LIMIT}")
  48. dataset_name = dataset_name.strip()
  49. dataset_name = duplicate_name(
  50. KnowledgebaseService.query,
  51. name=dataset_name,
  52. tenant_id=current_user.id,
  53. status=StatusEnum.VALID.value)
  54. try:
  55. req["id"] = get_uuid()
  56. req["name"] = dataset_name
  57. req["tenant_id"] = current_user.id
  58. req["created_by"] = current_user.id
  59. e, t = TenantService.get_by_id(current_user.id)
  60. if not e:
  61. return get_data_error_result(message="Tenant not found.")
  62. req["embd_id"] = t.embd_id
  63. if not KnowledgebaseService.save(**req):
  64. return get_data_error_result()
  65. return get_json_result(data={"kb_id": req["id"]})
  66. except Exception as e:
  67. return server_error_response(e)
  68. @manager.route('/update', methods=['post']) # noqa: F821
  69. @login_required
  70. @validate_request("kb_id", "name", "description", "parser_id")
  71. @not_allowed_parameters("id", "tenant_id", "created_by", "create_time", "update_time", "create_date", "update_date", "created_by")
  72. def update():
  73. req = request.json
  74. if not isinstance(req["name"], str):
  75. return get_data_error_result(message="Dataset name must be string.")
  76. if req["name"].strip() == "":
  77. return get_data_error_result(message="Dataset name can't be empty.")
  78. if len(req["name"].encode("utf-8")) > DATASET_NAME_LIMIT:
  79. return get_data_error_result(
  80. message=f"Dataset name length is {len(req['name'])} which is large than {DATASET_NAME_LIMIT}")
  81. req["name"] = req["name"].strip()
  82. if not KnowledgebaseService.accessible4deletion(req["kb_id"], current_user.id):
  83. return get_json_result(
  84. data=False,
  85. message='No authorization.',
  86. code=settings.RetCode.AUTHENTICATION_ERROR
  87. )
  88. try:
  89. if not KnowledgebaseService.query(
  90. created_by=current_user.id, id=req["kb_id"]):
  91. return get_json_result(
  92. data=False, message='Only owner of knowledgebase authorized for this operation.',
  93. code=settings.RetCode.OPERATING_ERROR)
  94. e, kb = KnowledgebaseService.get_by_id(req["kb_id"])
  95. if not e:
  96. return get_data_error_result(
  97. message="Can't find this knowledgebase!")
  98. if req["name"].lower() != kb.name.lower() \
  99. and len(
  100. KnowledgebaseService.query(name=req["name"], tenant_id=current_user.id, status=StatusEnum.VALID.value)) >= 1:
  101. return get_data_error_result(
  102. message="Duplicated knowledgebase name.")
  103. del req["kb_id"]
  104. if not KnowledgebaseService.update_by_id(kb.id, req):
  105. return get_data_error_result()
  106. if kb.pagerank != req.get("pagerank", 0):
  107. if req.get("pagerank", 0) > 0:
  108. settings.docStoreConn.update({"kb_id": kb.id}, {PAGERANK_FLD: req["pagerank"]},
  109. search.index_name(kb.tenant_id), kb.id)
  110. else:
  111. # Elasticsearch requires PAGERANK_FLD be non-zero!
  112. settings.docStoreConn.update({"exists": PAGERANK_FLD}, {"remove": PAGERANK_FLD},
  113. search.index_name(kb.tenant_id), kb.id)
  114. e, kb = KnowledgebaseService.get_by_id(kb.id)
  115. if not e:
  116. return get_data_error_result(
  117. message="Database error (Knowledgebase rename)!")
  118. kb = kb.to_dict()
  119. kb.update(req)
  120. return get_json_result(data=kb)
  121. except Exception as e:
  122. return server_error_response(e)
  123. @manager.route('/detail', methods=['GET']) # noqa: F821
  124. @login_required
  125. def detail():
  126. kb_id = request.args["kb_id"]
  127. try:
  128. tenants = UserTenantService.query(user_id=current_user.id)
  129. for tenant in tenants:
  130. if KnowledgebaseService.query(
  131. tenant_id=tenant.tenant_id, id=kb_id):
  132. break
  133. else:
  134. return get_json_result(
  135. data=False, message='Only owner of knowledgebase authorized for this operation.',
  136. code=settings.RetCode.OPERATING_ERROR)
  137. kb = KnowledgebaseService.get_detail(kb_id)
  138. if not kb:
  139. return get_data_error_result(
  140. message="Can't find this knowledgebase!")
  141. kb["size"] = DocumentService.get_total_size_by_kb_id(kb_id=kb["id"],keywords="", run_status=[], types=[])
  142. return get_json_result(data=kb)
  143. except Exception as e:
  144. return server_error_response(e)
  145. @manager.route('/list', methods=['POST']) # noqa: F821
  146. @login_required
  147. def list_kbs():
  148. keywords = request.args.get("keywords", "")
  149. page_number = int(request.args.get("page", 0))
  150. items_per_page = int(request.args.get("page_size", 0))
  151. parser_id = request.args.get("parser_id")
  152. orderby = request.args.get("orderby", "create_time")
  153. if request.args.get("desc", "true").lower() == "false":
  154. desc = False
  155. else:
  156. desc = True
  157. req = request.get_json()
  158. owner_ids = req.get("owner_ids", [])
  159. try:
  160. if not owner_ids:
  161. tenants = TenantService.get_joined_tenants_by_user_id(current_user.id)
  162. tenants = [m["tenant_id"] for m in tenants]
  163. kbs, total = KnowledgebaseService.get_by_tenant_ids(
  164. tenants, current_user.id, page_number,
  165. items_per_page, orderby, desc, keywords, parser_id)
  166. else:
  167. tenants = owner_ids
  168. kbs, total = KnowledgebaseService.get_by_tenant_ids(
  169. tenants, current_user.id, 0,
  170. 0, orderby, desc, keywords, parser_id)
  171. kbs = [kb for kb in kbs if kb["tenant_id"] in tenants]
  172. total = len(kbs)
  173. if page_number and items_per_page:
  174. kbs = kbs[(page_number-1)*items_per_page:page_number*items_per_page]
  175. return get_json_result(data={"kbs": kbs, "total": total})
  176. except Exception as e:
  177. return server_error_response(e)
  178. @manager.route('/rm', methods=['post']) # noqa: F821
  179. @login_required
  180. @validate_request("kb_id")
  181. def rm():
  182. req = request.json
  183. if not KnowledgebaseService.accessible4deletion(req["kb_id"], current_user.id):
  184. return get_json_result(
  185. data=False,
  186. message='No authorization.',
  187. code=settings.RetCode.AUTHENTICATION_ERROR
  188. )
  189. try:
  190. kbs = KnowledgebaseService.query(
  191. created_by=current_user.id, id=req["kb_id"])
  192. if not kbs:
  193. return get_json_result(
  194. data=False, message='Only owner of knowledgebase authorized for this operation.',
  195. code=settings.RetCode.OPERATING_ERROR)
  196. for doc in DocumentService.query(kb_id=req["kb_id"]):
  197. if not DocumentService.remove_document(doc, kbs[0].tenant_id):
  198. return get_data_error_result(
  199. message="Database error (Document removal)!")
  200. f2d = File2DocumentService.get_by_document_id(doc.id)
  201. if f2d:
  202. FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
  203. File2DocumentService.delete_by_document_id(doc.id)
  204. FileService.filter_delete(
  205. [File.source_type == FileSource.KNOWLEDGEBASE, File.type == "folder", File.name == kbs[0].name])
  206. if not KnowledgebaseService.delete_by_id(req["kb_id"]):
  207. return get_data_error_result(
  208. message="Database error (Knowledgebase removal)!")
  209. for kb in kbs:
  210. settings.docStoreConn.delete({"kb_id": kb.id}, search.index_name(kb.tenant_id), kb.id)
  211. settings.docStoreConn.deleteIdx(search.index_name(kb.tenant_id), kb.id)
  212. if hasattr(STORAGE_IMPL, 'remove_bucket'):
  213. STORAGE_IMPL.remove_bucket(kb.id)
  214. return get_json_result(data=True)
  215. except Exception as e:
  216. return server_error_response(e)
  217. @manager.route('/<kb_id>/tags', methods=['GET']) # noqa: F821
  218. @login_required
  219. def list_tags(kb_id):
  220. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  221. return get_json_result(
  222. data=False,
  223. message='No authorization.',
  224. code=settings.RetCode.AUTHENTICATION_ERROR
  225. )
  226. tags = settings.retrievaler.all_tags(current_user.id, [kb_id])
  227. return get_json_result(data=tags)
  228. @manager.route('/tags', methods=['GET']) # noqa: F821
  229. @login_required
  230. def list_tags_from_kbs():
  231. kb_ids = request.args.get("kb_ids", "").split(",")
  232. for kb_id in kb_ids:
  233. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  234. return get_json_result(
  235. data=False,
  236. message='No authorization.',
  237. code=settings.RetCode.AUTHENTICATION_ERROR
  238. )
  239. tags = settings.retrievaler.all_tags(current_user.id, kb_ids)
  240. return get_json_result(data=tags)
  241. @manager.route('/<kb_id>/rm_tags', methods=['POST']) # noqa: F821
  242. @login_required
  243. def rm_tags(kb_id):
  244. req = request.json
  245. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  246. return get_json_result(
  247. data=False,
  248. message='No authorization.',
  249. code=settings.RetCode.AUTHENTICATION_ERROR
  250. )
  251. e, kb = KnowledgebaseService.get_by_id(kb_id)
  252. for t in req["tags"]:
  253. settings.docStoreConn.update({"tag_kwd": t, "kb_id": [kb_id]},
  254. {"remove": {"tag_kwd": t}},
  255. search.index_name(kb.tenant_id),
  256. kb_id)
  257. return get_json_result(data=True)
  258. @manager.route('/<kb_id>/rename_tag', methods=['POST']) # noqa: F821
  259. @login_required
  260. def rename_tags(kb_id):
  261. req = request.json
  262. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  263. return get_json_result(
  264. data=False,
  265. message='No authorization.',
  266. code=settings.RetCode.AUTHENTICATION_ERROR
  267. )
  268. e, kb = KnowledgebaseService.get_by_id(kb_id)
  269. settings.docStoreConn.update({"tag_kwd": req["from_tag"], "kb_id": [kb_id]},
  270. {"remove": {"tag_kwd": req["from_tag"].strip()}, "add": {"tag_kwd": req["to_tag"]}},
  271. search.index_name(kb.tenant_id),
  272. kb_id)
  273. return get_json_result(data=True)
  274. @manager.route('/<kb_id>/knowledge_graph', methods=['GET']) # noqa: F821
  275. @login_required
  276. def knowledge_graph(kb_id):
  277. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  278. return get_json_result(
  279. data=False,
  280. message='No authorization.',
  281. code=settings.RetCode.AUTHENTICATION_ERROR
  282. )
  283. _, kb = KnowledgebaseService.get_by_id(kb_id)
  284. req = {
  285. "kb_id": [kb_id],
  286. "knowledge_graph_kwd": ["graph"]
  287. }
  288. obj = {"graph": {}, "mind_map": {}}
  289. if not settings.docStoreConn.indexExist(search.index_name(kb.tenant_id), kb_id):
  290. return get_json_result(data=obj)
  291. sres = settings.retrievaler.search(req, search.index_name(kb.tenant_id), [kb_id])
  292. if not len(sres.ids):
  293. return get_json_result(data=obj)
  294. for id in sres.ids[:1]:
  295. ty = sres.field[id]["knowledge_graph_kwd"]
  296. try:
  297. content_json = json.loads(sres.field[id]["content_with_weight"])
  298. except Exception:
  299. continue
  300. obj[ty] = content_json
  301. if "nodes" in obj["graph"]:
  302. obj["graph"]["nodes"] = sorted(obj["graph"]["nodes"], key=lambda x: x.get("pagerank", 0), reverse=True)[:256]
  303. if "edges" in obj["graph"]:
  304. node_id_set = { o["id"] for o in obj["graph"]["nodes"] }
  305. filtered_edges = [o for o in obj["graph"]["edges"] if o["source"] != o["target"] and o["source"] in node_id_set and o["target"] in node_id_set]
  306. obj["graph"]["edges"] = sorted(filtered_edges, key=lambda x: x.get("weight", 0), reverse=True)[:128]
  307. return get_json_result(data=obj)
  308. @manager.route('/<kb_id>/knowledge_graph', methods=['DELETE']) # noqa: F821
  309. @login_required
  310. def delete_knowledge_graph(kb_id):
  311. if not KnowledgebaseService.accessible(kb_id, current_user.id):
  312. return get_json_result(
  313. data=False,
  314. message='No authorization.',
  315. code=settings.RetCode.AUTHENTICATION_ERROR
  316. )
  317. _, kb = KnowledgebaseService.get_by_id(kb_id)
  318. settings.docStoreConn.delete({"knowledge_graph_kwd": ["graph", "subgraph", "entity", "relation"]}, search.index_name(kb.tenant_id), kb_id)
  319. return get_json_result(data=True)