#!/usr/bin/env python3 """给「阿里云合同条款」接上知识库。现网已经执行过,知识库 id 固定。 重复跑「创建知识库」会再索引一份文档。只在知识库不存在时使用创建请求。 """ DATASET_ID = "71042108-5f9a-4856-a0db-4359e8e7fa98" FILE_ID = "c74afc20-f95b-4be1-b548-aad73abb5325" AGENT_ID = "01a1236c-4cc3-7fb9-bfce-c730a318bdf9" PROVIDER = "langgenius/xinference/xinference" init_body = { "name": "阿里云合同条款", "indexing_technique": "high_quality", "doc_form": "text_model", "doc_language": "Chinese", "embedding_model": "bge-m3", "embedding_model_provider": PROVIDER, "data_source": { "info_list": { "data_source_type": "upload_file", "file_info_list": {"file_ids": [FILE_ID]}, } }, "process_rule": {"mode": "automatic"}, "retrieval_model": { "search_method": "hybrid_search", "reranking_enable": True, "reranking_mode": "reranking_model", "reranking_model": { "reranking_provider_name": PROVIDER, "reranking_model_name": "bge-reranker-v2-m3", }, "top_k": 4, "score_threshold_enabled": False, "weights": { "weight_type": "customized", "vector_setting": { "vector_weight": 0.7, "embedding_provider_name": PROVIDER, "embedding_model_name": "bge-m3", }, "keyword_setting": {"keyword_weight": 0.3}, }, }, } knowledge_set = { "id": "retrieval-1", "name": "Retrieval 1", "description": "阿里云框架合同原文及附件", "datasets": [{"id": DATASET_ID, "name": "阿里云合同条款"}], "query": {"mode": "generated_query"}, "retrieval": { "mode": "multiple", "top_k": 4, "reranking_mode": "reranking_model", "reranking_enable": True, "reranking_model": {"provider": PROVIDER, "model": "bge-reranker-v2-m3"}, }, "metadata_filtering": {"mode": "disabled"}, } print("POST /console/api/datasets/init") print("PUT /console/api/agent/%s/composer" % AGENT_ID) print("save_strategy = save_to_current_version") print("POST /console/api/agent/%s/publish" % AGENT_ID) print("knowledge set", knowledge_set["name"], "dataset", DATASET_ID) print("init file", FILE_ID, "chunks expected", 75)