{"id":16813,"date":"2024-12-31T16:25:29","date_gmt":"2024-12-31T08:25:29","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=16813"},"modified":"2025-01-05T09:31:10","modified_gmt":"2025-01-05T01:31:10","slug":"chaijiegoujianzhishica","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/en\/chaijiegoujianzhishica\/","title":{"rendered":"\u62c6\u89e3\u5229\u7528\u5927\u6a21\u578b\u6784\u5efa\u77e5\u8bc6\u56fe\u8c31\u8fc7\u7a0b\u4e2d\u5e94\u7528\u5230\u7684\u63d0\u793a\u8bcd\uff08\u793a\u4f8b\uff09"},"content":{"rendered":"

\u9274\u4e8e\u5f88\u591a\u8981\u5e38\u8bc6\u4f7f\u7528\u77e5\u8bc6\u56fe\u8c31\u63d0\u5347\u53ec\u56de\u7387\u6216\u4f5c\u4e3a\u957f\u671f\u8bb0\u5fc6\u5b58\u50a8\u7684\u670b\u53cb\u6ca1\u641e\u61c2\u5927\u6a21\u578b\u65f6\u4ee3\u7684\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u65b9\u6cd5\uff0c\u8fd9\u91cc\u7b80\u5355\u666e\u53ca\u5e38\u8bc6\uff0c\u548c\u5b9e\u9645\u6784\u5efa\u4f1a\u6709\u8f83\u591a\u5dee\u5f02\u3002<\/p>\n

\u5176\u5b9e\u77e5\u8bc6\u56fe\u8c31\u8bf4\u7b80\u5355\u4e5f\u7b80\u5355\uff0c\u5927\u6a21\u578b\u5728\u5176\u6d41\u7a0b\u4e2d\u4e3b\u8981\u8d1f\u8d23\u4e09\u5927\u5757\uff1a\u5173\u7cfb\u63d0\u53d6\u3001\u6784\u5efa\u67e5\u8be2\u3001\u56de\u7b54\u95ee\u9898\u3002<\/p>\n

\u5b9e\u9645\u4e0a\u8fd8\u6709\u5f88\u591a\u7ec6\u8282\u95ee\u9898\uff0c\u5982\u4e0b\u793a\u4f8b\uff1a\uff08\u4f18\u5316\u4e2d\u95f4\u67e5\u8be2\u8fc7\u7a0b\uff09<\/p>\n

 <\/p>\n

\u77e5\u8bc6\u56fe\u8c31\u63d0\u793a\u8bcd\u6784\u5efa<\/h2>\n
# \u6b65\u9aa41: \u95ee\u9898\u5206\u6790\u63d0\u793a\u8bcd\r\nQUESTION_ANALYSIS_PROMPT = \"\"\"\r\nYou are an expert query analyzer. Your task is to analyze the given question and extract key search terms.\r\n\r\nInput Question: {query}\r\n\r\nFollow these steps:\r\n1. Identify main entities (nouns, proper nouns)\r\n2. Extract important attributes (adjectives, descriptors)\r\n3. Identify relationship indicators (verbs, prepositions)\r\n4. Note any temporal or conditional terms\r\n\r\nFormat your output as:\r\n{\r\n\"main_entities\": [\"entity1\", \"entity2\"...],\r\n\"attributes\": [\"attr1\", \"attr2\"...],\r\n\"relationships\": [\"rel1\", \"rel2\"...],\r\n\"conditions\": [\"cond1\", \"cond2\"...]\r\n}\r\n\r\nEnsure each term is:\r\n- Specific and relevant\r\n- In its base\/root form\r\n- Without duplicates\r\n\r\nExamples:\r\nQ: \"Who created the Python programming language?\"\r\n{\r\n\"main_entities\": [\"Python\", \"programming language\"],\r\n\"attributes\": [\"created\"],\r\n\"relationships\": [\"create\", \"develop\"],\r\n\"conditions\": []\r\n}\r\n\"\"\"\r\n\r\n# \u6b65\u9aa42: \u540c\u4e49\u8bcd\u6269\u5c55\u63d0\u793a\u8bcd\r\nSYNONYM_EXPANSION_PROMPT = \"\"\"\r\nYou are an expert synonym expansion system. Your task is to expand each term with relevant alternatives.\r\n\r\nInput Terms: {terms}\r\n\r\nFor each term, provide:\r\n1. Exact synonyms\r\n2. Related terms\r\n3. Common variations\r\n4. Abbreviations\/Acronyms\r\n5. Full forms\r\n6. Common misspellings\r\n\r\nRules:\r\n- Include industry-standard terminology\r\n- Consider different naming conventions\r\n- Include both formal and informal terms\r\n- Maintain semantic equivalence\r\n\r\nFormat your output as:\r\n{\r\n\"term\": {\r\n\"synonyms\": [\"syn1\", \"syn2\"...],\r\n\"variations\": [\"var1\", \"var2\"...],\r\n\"abbreviations\": [\"abbr1\", \"abbr2\"...],\r\n\"related_terms\": [\"rel1\", \"rel2\"...]\r\n}\r\n}\r\n\r\nExample:\r\nInput: \"Python\"\r\n{\r\n\"Python\": {\r\n\"synonyms\": [\"Python programming language\", \"Python lang\"],\r\n\"variations\": [\"python\", \"Python3\", \"Python2\"],\r\n\"abbreviations\": [\"py\", \".py\"],\r\n\"related_terms\": [\"CPython\", \"Jython\", \"IronPython\"]\r\n}\r\n}\r\n\"\"\"\r\n\r\n# \u6b65\u9aa43: \u67e5\u8be2\u6784\u5efa\u63d0\u793a\u8bcd\r\nQUERY_CONSTRUCTION_PROMPT = \"\"\"\r\nYou are an expert in constructing graph database queries. Your task is to create an optimized search pattern.\r\n\r\nInput:\r\n- Primary Terms: {primary_terms}\r\n- Expanded Terms: {expanded_terms}\r\n- Relationships: {relationships}\r\n\r\nGenerate a query pattern that:\r\n1. Prioritizes exact matches\r\n2. Includes synonym matches\r\n3. Considers relationship patterns\r\n4. Handles variations in terminology\r\n\r\nRules:\r\n- Start with most specific terms\r\n- Include all relevant relationships\r\n- Consider bidirectional relationships\r\n- Limit path length appropriately\r\n\r\nFormat your output as:\r\n{\r\n\"exact_match_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"fuzzy_match_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"relationship_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"priority_order\": [\"high\", \"medium\", \"low\"]\r\n}\r\n\r\nExample:\r\n{\r\n\"exact_match_patterns\": [\"MATCH (n:Entity {name: 'Python'})\", \"MATCH (n:Language {type: 'programming'})\"],\r\n\"fuzzy_match_patterns\": [\"MATCH (n) WHERE n.name =~ '(?i).*python.*'\"],\r\n\"relationship_patterns\": [\"MATCH (creator)-[:CREATED]->(lang)\", \"MATCH (lang)-[:TYPE_OF]->(prog_lang)\"],\r\n\"priority_order\": [\"exact_name_match\", \"fuzzy_name_match\", \"relationship_match\"]\r\n}\r\n\"\"\"\r\n\r\n# \u6b65\u9aa44: \u7ed3\u679c\u6392\u5e8f\u63d0\u793a\u8bcd\r\nRESULT_RANKING_PROMPT = \"\"\"\r\nYou are an expert in ranking and ordering search results. Your task is to score and rank the retrieved matches.\r\n\r\nInput Results: {query_results}\r\nOriginal Query: {original_query}\r\n\r\nRanking Criteria:\r\n1. Relevance<\/a> to original query\r\n2. Match quality (exact vs partial)\r\n3. Relationship distance\r\n4. Information completeness\r\n5. Source reliability\r\n\r\nScore each result on:\r\n- Relevance (0-10)\r\n- Confidence (0-10)\r\n- Completeness (0-10)\r\n- Path length penalty (-1 per hop)\r\n\r\nFormat your output as:\r\n{\r\n\"ranked_results\": [\r\n{\r\n\"result\": \"result_content\",\r\n\"relevance_score\": score,\r\n\"confidence_score\": score,\r\n\"completeness_score\": score,\r\n\"final_score\": score,\r\n\"reasoning\": \"explanation\"\r\n}\r\n],\r\n\"summary\": {\r\n\"total_results\": number,\r\n\"high_confidence_count\": number,\r\n\"average_score\": number\r\n}\r\n}\r\n\"\"\"<\/pre>\n
\u6839\u636e\u4e0a\u8ff0\u63d0\u793a\u8bcd\u53ef\u4ee5\u770b\u5230\uff0c\u6784\u5efa\u77e5\u8bc6\u56fe\u8c31\u7684\u4e2d\u95f4\u8fc7\u7a0b\u53ef\u4ee5\u6709\u5f88\u591a\u7684\u7ec6\u8282\u8981\u5904\u7406\u3002\u4ee5\u4e0a\u793a\u4f8b\u611f\u5b98\u4e0d\u660e\u663e\uff0c\u7ed9\u51fa\u4ee5\u4e0b\u8f93\u51fa\u793a\u4f8b\u4f5c\u4e3a\u53c2\u8003\u3002<\/p>\n
 <\/p>\n
\u77e5\u8bc6\u56fe\u8c31\u5185\u5bb9\u8f93\u5165\u4e0e\u8f93\u51fa<\/h2>\n
\u6b65\u9aa41: \u95ee\u9898\u5206\u6790<\/h3>\n
\u8f93\u5165:<\/p>\n
{\r\n    \"query\"<\/span>: \"\u8c37\u6b4c\u5f00\u53d1\u4e86\u54ea\u4e9b\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff1f\"<\/span>\r\n}\r\n<\/code><\/pre>\n
\u8f93\u51fa:<\/p>\n
{\r\n    \"main_entities\"<\/span>: [\"\u8c37\u6b4c\"<\/span>, \"\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>],\r\n    \"attributes\"<\/span>: [\"\u5f00\u53d1\"<\/span>],\r\n    \"relationships\"<\/span>: [\"\u5f00\u53d1\"<\/span>, \"\u521b\u5efa\"<\/span>, \"\u7ef4\u62a4\"<\/span>],\r\n    \"conditions\"<\/span>: []\r\n}\r\n<\/code><\/pre>\n
\u6b65\u9aa42: \u540c\u4e49\u8bcd\u6269\u5c55<\/h3>\n
\u8f93\u5165:<\/p>\n
{\r\n    \"terms\"<\/span>: [\"\u8c37\u6b4c\"<\/span>, \"\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>]\r\n}\r\n<\/code><\/pre>\n
\u8f93\u51fa:<\/p>\n
{\r\n    \"\u8c37\u6b4c\"<\/span>: {\r\n        \"synonyms\"<\/span>: [\"Google LLC\"<\/span>, \"Google Inc.\"<\/span>, \"Google Corporation\"<\/span>],\r\n        \"variations\"<\/span>: [\"google\"<\/span>, \"GOOGLE\"<\/span>],\r\n        \"abbreviations\"<\/span>: [\"GOOG\"<\/span>],\r\n        \"related_terms\"<\/span>: [\"Alphabet Inc.\"<\/span>, \"Google AI\"<\/span>, \"Google Research\"<\/span>]\r\n    },\r\n    \"\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>: {\r\n        \"synonyms\"<\/span>: [\"\u6df1\u5ea6\u5b66\u4e60\u5e93\"<\/span>, \"\u795e\u7ecf\u7f51\u7edc\u6846\u67b6\"<\/span>],\r\n        \"variations\"<\/span>: [\"DL framework\"<\/span>, \"\u6df1\u5ea6\u5b66\u4e60\u5e73\u53f0\"<\/span>],\r\n        \"abbreviations\"<\/span>: [\"DL framework\"<\/span>, \"DNN framework\"<\/span>],\r\n        \"related_terms\"<\/span>: [\"\u673a\u5668\u5b66\u4e60\u6846\u67b6\"<\/span>, \"AI\u6846\u67b6\"<\/span>, \"\u795e\u7ecf\u7f51\u7edc\u5e93\"<\/span>]\r\n    }\r\n}\r\n<\/code><\/pre>\n
\u6b65\u9aa43: \u67e5\u8be2\u6784\u5efa<\/h3>\n
\u8f93\u5165:<\/p>\n
{\r\n    \"primary_terms\"<\/span>: [\"\u8c37\u6b4c\"<\/span>, \"\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>],\r\n    \"expanded_terms\"<\/span>: [\r\n        \"\u8c37\u6b4c\"<\/span>, \"Google LLC\"<\/span>, \"Google Inc.\"<\/span>, \"Google AI\"<\/span>,\r\n        \"\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>, \"\u6df1\u5ea6\u5b66\u4e60\u5e93\"<\/span>, \"\u795e\u7ecf\u7f51\u7edc\u6846\u67b6\"<\/span>\r\n    ],\r\n    \"relationships\"<\/span>: [\"\u5f00\u53d1\"<\/span>, \"\u521b\u5efa\"<\/span>, \"\u7ef4\u62a4\"<\/span>]\r\n}\r\n<\/code><\/pre>\n
\u8f93\u51fa:<\/p>\n
{\r\n    \"exact_match_patterns\"<\/span>: [\r\n        \"MATCH (org:Organization {name: 'Google'})\"<\/span>,\r\n        \"MATCH (framework:Framework {type: 'deep_learning'})\"<\/span>\r\n    ],\r\n    \"fuzzy_match_patterns\"<\/span>: [\r\n        \"MATCH (org) WHERE org.name =~ '(?i).*google.*'\"<\/span>,\r\n        \"MATCH (framework) WHERE framework.type =~ '(?i).*(deep learning|neural network).*'\"<\/span>\r\n    ],\r\n    \"relationship_patterns\"<\/span>: [\r\n        \"MATCH (org)-[:DEVELOPED]->(framework)\"<\/span>,\r\n        \"MATCH (org)-[:CREATED]->(framework)\"<\/span>,\r\n        \"MATCH (org)-[:MAINTAINS]->(framework)\"<\/span>,\r\n        \"MATCH (framework)-[:DEVELOPED_BY]->(org)\"<\/span>\r\n    ],\r\n    \"priority_order\"<\/span>: [\r\n        \"exact_organization_match\"<\/span>,\r\n        \"exact_framework_match\"<\/span>,\r\n        \"relationship_match\"<\/span>,\r\n        \"fuzzy_framework_match\"<\/span>\r\n    ]\r\n}\r\n<\/code><\/pre>\n
\u6b65\u9aa44: \u7ed3\u679c\u6392\u5e8f<\/h3>\n
\u8f93\u5165:<\/p>\n
{\r\n    \"query_results\"<\/span>: [\r\n        {\r\n            \"org\"<\/span>: \"\u8c37\u6b4c\"<\/span>,\r\n            \"framework\"<\/span>: \"TensorFlow\"<\/span>,\r\n            \"relationship\"<\/span>: \"\u5f00\u53d1\"<\/span>,\r\n            \"year\"<\/span>: \"2015\"<\/span>\r\n        },\r\n        {\r\n            \"org\"<\/span>: \"\u8c37\u6b4c\"<\/span>,\r\n            \"framework\"<\/span>: \"JAX\"<\/span>,\r\n            \"relationship\"<\/span>: \"\u5f00\u53d1\"<\/span>,\r\n            \"year\"<\/span>: \"2018\"<\/span>\r\n        }\r\n    ],\r\n    \"original_query\"<\/span>: \"\u8c37\u6b4c\u5f00\u53d1\u4e86\u54ea\u4e9b\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff1f\"<\/span>\r\n}\r\n<\/code><\/pre>\n
\u8f93\u51fa:<\/p>\n
{\r\n    \"ranked_results\"<\/span>: [\r\n        {\r\n            \"result\"<\/span>: \"\u8c37\u6b4c\u5f00\u53d1\u4e86TensorFlow\uff0c\u8fd9\u662f\u4e00\u6b3e\u6d41\u884c\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>,\r\n            \"relevance_score\"<\/span>: 10.0<\/span>,\r\n            \"confidence_score\"<\/span>: 9.5<\/span>,\r\n            \"completeness_score\"<\/span>: 9.0<\/span>,\r\n            \"final_score\"<\/span>: 9.5<\/span>,\r\n            \"reasoning\"<\/span>: \"\u76f4\u63a5\u7b54\u6848\uff0c\u4fe1\u5fc3\u9ad8 - TensorFlow\u662f\u8c37\u6b4c\u7684\u4e3b\u8981\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>\r\n        },\r\n        {\r\n            \"result\"<\/span>: \"\u8c37\u6b4c\u5f00\u53d1\u4e86JAX\uff0c\u4e00\u4e2a\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\"<\/span>,\r\n            \"relevance_score\"<\/span>: 9.0<\/span>,\r\n            \"confidence_score\"<\/span>: 8.5<\/span>,\r\n            \"completeness_score\"<\/span>: 8.0<\/span>,\r\n            \"final_score\"<\/span>: 8.5<\/span>,\r\n            \"reasoning\"<\/span>: \"\u76f8\u5173\u4f46\u77e5\u540d\u5ea6\u4f4e\u4e8eTensorFlow\"<\/span>\r\n        }\r\n    ],\r\n    \"summary\"<\/span>: {\r\n        \"total_results\"<\/span>: 2<\/span>,\r\n        \"high_confidence_count\"<\/span>: 2<\/span>,\r\n        \"average_score\"<\/span>: 9.0<\/span>\r\n    }\r\n}\r\n<\/code><\/pre>\n
\u6700\u7ec8\u7ec4\u5408\u67e5\u8be2 (Cypher):<\/h3>\n
\/\/ 1. \u9996\u5148\u5339\u914d\u7ec4\u7ec7\u548c\u6846\u67b6\u7684\u5173\u7cfb\r\nMATCH (org:Organization)-[r:DEVELOPED|CREATED|MAINTAINS]->(framework:Framework)\r\nWHERE org.name =~ '(?i).*google.*'\r\n  AND framework.type =~ '(?i).*(deep learning|neural network).*'\r\n\r\n\/\/ 2. \u4fdd\u5b58\u4e2d\u95f4\u7ed3\u679c\r\nWITH org, framework, r\r\n\r\n\/\/ 3. \u5bfb\u627e\u6700\u77ed\u8def\u5f84\r\nMATCH p = shortestPath((org)-[*1..2]-(framework))\r\n\r\n\/\/ 4. \u8fd4\u56de\u8def\u5f84\u5e76\u6309\u6d41\u884c\u5ea6\u6392\u5e8f\r\nRETURN p\r\nORDER BY framework.popularity DESC\r\nLIMIT 10\r\n<\/code><\/pre>\n
\u6700\u7ec8\u7b54\u6848\uff1a<\/strong>Google \u5f00\u53d1\u4e86\u591a\u4e2a\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff0c” +”\u6700\u8457\u540d\u7684\u662f TensorFlow (2015) \u548c JAX (2018)\u3002” +”TensorFlow \u662f\u4ed6\u4eec\u7684\u4e3b\u8981\u6846\u67b6\uff0c\u4e5f\u662f\u4f7f\u7528\u6700\u5e7f\u6cdb\u7684\u6846\u67b6\u3002<\/p>\n
\u4ee5\u4e0a\u6267\u884c\u8fc7\u7a0b\u4e5f\u5e76\u4e0d\u5b8c\u6574\uff0c\u5efa\u8bae\u62c6\u89e3\u4e00\u4e2a\u5b8c\u6574\u7684\u9879\u76ee\u8fdb\u884c\u5206\u6790\uff0c Llamaindex \u662f\u4e0d\u9519\u7684\u9009\u62e9\u3002\u4e0b\u9762\u6536\u96c6\u4e00\u4e9b\u63d0\u793a\u8bcd\u793a\u4f8b\u7ed9\u5927\u5bb6\u627e\u627e\u611f\u89c9\u3002<\/p>\n
 <\/p>\n
\u77e5\u8bc6\u56fe\u8c31\u63d0\u793a\u8bcd\u793a\u4f8b<\/h2>\n
\u5b9e\u4f53\u5173\u7cfb\u63d0\u53d6<\/strong><\/h3>\n
\u7cfb\u7edf\uff1a\u4f60\u662f\u4e00\u540d\u63d0\u53d6\u77e5\u8bc6\u4e09\u5143\u7ec4\u7684\u4e13\u5bb6\u3002 \r\n\u4f60\u7684\u4efb\u52a1\u662f\u4ece\u7ed9\u5b9a\u6587\u672c\u4e2d\u8bc6\u522b\u5b9e\u4f53\u53ca\u5176\u5173\u7cfb\u3002\r\n\r\n\u5bf9\u4e8e\u6bcf\u6bb5\u6587\u672c\uff0c\u751f\u6210\u683c\u5f0f\u5982\u4e0b\u7684\u77e5\u8bc6\u4e09\u5143\u7ec4\uff1a \r\n(subject, relation, object)\r\n\r\n\u89c4\u5219\uff1a \r\n1. \u4e3b\u4f53\uff08subject\uff09\u548c\u5ba2\u4f53\uff08object\uff09\u5fc5\u987b\u662f\u5177\u4f53\u7684\u5b9e\u4f53 \r\n2. \u5173\u7cfb\uff08relation\uff09\u5e94\u4e3a\u6e05\u6670\u7b80\u6d01\u7684\u8c13\u8bcd \r\n3. \u6bcf\u4e2a\u4e09\u5143\u7ec4\u5e94\u8868\u793a\u5355\u4e00\u7684\u3001\u57fa\u4e8e\u4e8b\u5b9e\u7684\u9648\u8ff0 \r\n4. \u907f\u514d\u6cdb\u5316\u6216\u6a21\u7cca\u7684\u5173\u7cfb \r\n5. \u4fdd\u6301\u5b9e\u4f53\u547d\u540d\u7684\u4e00\u81f4\u6027\r\n\r\n\u8f93\u5165\u6587\u672c\uff1a{text_chunk}\r\n\r\n\u4ece\u4e0a\u8ff0\u6587\u672c\u4e2d\u63d0\u53d6\u77e5\u8bc6\u4e09\u5143\u7ec4\u3002\u5c06\u6bcf\u4e2a\u4e09\u5143\u7ec4\u683c\u5f0f\u5316\u4e3a\uff1a \r\n(entity1) -> [relationship] -> (entity2)<\/pre>\n
\u6ce8\uff1a\u8981\u6839\u636e\u81ea\u8eab\u884c\u4e1a\u5408\u8ba1\u5b9e\u4f53\u5173\u7cfb\uff0c\u63d0\u53d6\u5b9e\u4f53\u5173\u7cfb\u8bbe\u8ba1\u5f88\u590d\u6742\uff0c\u6240\u4ee5\u8fd9\u91cc\u4ec5\u63d0\u4f9b\u57fa\u7840\u6a21\u677f\u3002<\/strong><\/p>\n
 <\/p>\n
\u95ee\u9898\u5206\u6790\u63d0\u793a\u8bcd<\/strong><\/h3>\n
\u4f60\u662f\u4e00\u4f4d\u4e13\u4e1a\u7684\u67e5\u8be2\u5206\u6790\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u5206\u6790\u7ed9\u5b9a\u7684\u95ee\u9898\uff0c\u5e76\u63d0\u53d6\u5173\u952e\u7ec4\u4ef6\u4ee5\u7528\u4e8e\u77e5\u8bc6\u56fe\u8c31\u641c\u7d22\u3002\r\n\r\n\u8f93\u5165\u95ee\u9898: {query}\r\n\r\n\u8bf7\u4ed4\u7ec6\u6309\u7167\u4ee5\u4e0b\u6b65\u9aa4\u64cd\u4f5c\uff1a\r\n\r\n1. \u63d0\u53d6\u5173\u952e\u5b9e\u4f53\uff1a\r\n- \u8bc6\u522b\u6240\u6709\u4e3b\u8981\u5b9e\u4f53\uff08\u540d\u8bcd\u3001\u4e13\u6709\u540d\u8bcd\u3001\u6280\u672f\u672f\u8bed\uff09\r\n- \u5217\u51fa\u63d0\u5230\u7684\u4efb\u4f55\u7279\u5b9a\u5c5e\u6027\r\n- \u8bb0\u5f55\u4efb\u4f55\u6570\u503c\u6216\u65e5\u671f\r\n\r\n2. \u786e\u5b9a\u5173\u7cfb\uff1a\r\n- \u627e\u51fa\u8868\u793a\u5b9e\u4f53\u95f4\u5173\u7cfb\u7684\u52a8\u8bcd\r\n- \u786e\u5b9a\u663e\u793a\u8fde\u63a5\u7684\u4ecb\u8bcd\r\n- \u6ce8\u610f\u4efb\u4f55\u9690\u542b\u7684\u5173\u7cfb\r\n\r\n3. \u68c0\u6d4b\u67e5\u8be2\u7c7b\u578b\uff1a\r\n- \u5224\u65ad\u662f\u5426\u5c5e\u4e8e\u4ee5\u4e0b\u7c7b\u578b\uff1a\r\n* \u4e8b\u5b9e\u67e5\u8be2\r\n* \u5173\u7cfb\u67e5\u8be2\r\n* \u6bd4\u8f83\u67e5\u8be2\r\n* \u5c5e\u6027\u67e5\u8be2\r\n* \u65f6\u95f4\u7ebf\u67e5\u8be2\r\n\r\n4. \u63d0\u53d6\u7ea6\u675f\u6761\u4ef6\uff1a\r\n- \u65f6\u95f4\u7ea6\u675f\r\n- \u5730\u70b9\u7ea6\u675f\r\n- \u6761\u4ef6\u7ea6\u675f\r\n- \u6570\u91cf\u7ea6\u675f\r\n\r\n\u8bf7\u6309\u4ee5\u4e0b\u683c\u5f0f\u8f93\u51fa\u4f60\u7684\u5206\u6790\u7ed3\u679c\uff1a\r\n{\r\n\"entities\": [\"entity1\", \"entity2\", ...],\r\n\"attributes\": [\"attribute1\", \"attribute2\", ...],\r\n\"relationships\": [\"relationship1\", \"relationship2\", ...],\r\n\"query_type\": \"type_of_query\",\r\n\"constraints\": {\r\n\"time\": [],\r\n\"location\": [],\r\n\"condition\": [],\r\n\"quantity\": []\r\n}\r\n}\r\n\r\n\u8bb0\u4f4f\uff1a\r\n- \u7cbe\u786e\u8bc6\u522b\u5b9e\u4f53\r\n- \u5305\u62ec\u6240\u6709\u53ef\u80fd\u7684\u672f\u8bed\u53d8\u4f53\r\n- \u4fdd\u6301\u6280\u672f\u51c6\u786e\u6027\r\n- \u4fdd\u7559\u9886\u57df\u7279\u5b9a\u672f\u8bed<\/pre>\n
 <\/p>\n
\u4f60\u662f\u4e00\u4f4d\u4e13\u4e1a\u7684\u67e5\u8be2\u5206\u6790\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u5206\u6790\u7ed9\u5b9a\u7684\u95ee\u9898\u5e76\u63d0\u53d6\u5173\u952e\u641c\u7d22\u8bcd\u3002\r\n\r\n\u8f93\u5165\u95ee\u9898: {query}\r\n\r\n\u6309\u7167\u4ee5\u4e0b\u6b65\u9aa4\u64cd\u4f5c\uff1a\r\n1. \u786e\u5b9a\u4e3b\u8981\u5b9e\u4f53\uff08\u540d\u8bcd\u3001\u4e13\u6709\u540d\u8bcd\uff09\r\n2. \u63d0\u53d6\u91cd\u8981\u5c5e\u6027\uff08\u5f62\u5bb9\u8bcd\u3001\u63cf\u8ff0\u8bcd\uff09\r\n3. \u786e\u5b9a\u5173\u7cfb\u6307\u793a\u8bcd\uff08\u52a8\u8bcd\u3001\u4ecb\u8bcd\uff09\r\n4. \u6ce8\u610f\u4efb\u4f55\u65f6\u95f4\u6216\u6761\u4ef6\u76f8\u5173\u7684\u672f\u8bed\r\n\r\n\u5c06\u8f93\u51fa\u683c\u5f0f\u5316\u4e3a\uff1a\r\n{\r\n\"main_entities\": [\"entity1\", \"entity2\"...],\r\n\"attributes\": [\"attr1\", \"attr2\"...],\r\n\"relationships\": [\"rel1\", \"rel2\"...],\r\n\"conditions\": [\"cond1\", \"cond2\"...]\r\n}\r\n\r\n\u786e\u4fdd\u6bcf\u4e2a\u672f\u8bed\uff1a\r\n- \u5177\u4f53\u4e14\u76f8\u5173\r\n- \u4e3a\u57fa\u7840\/\u8bcd\u6839\u5f62\u5f0f\r\n- \u65e0\u91cd\u590d\r\n\r\n\u793a\u4f8b\uff1a\r\nQ: \"\u8c01\u521b\u5efa\u4e86 Python \u7f16\u7a0b\u8bed\u8a00\uff1f\"\r\n{\r\n\"main_entities\": [\"Python\", \"\u7f16\u7a0b\u8bed\u8a00\"],\r\n\"attributes\": [\"\u521b\u5efa\"],\r\n\"relationships\": [\"create\", \"develop\"],\r\n\"conditions\": []\r\n}<\/pre>\n
 <\/p>\n
\u5206\u6790\u4ee5\u4e0b\u95ee\u9898\u5e76\u63d0\u53d6\u5173\u952e\u672f\u8bed\u4ee5\u7528\u4e8e\u641c\u7d22\uff1a \r\n\u95ee\u9898\uff1a{query}\r\n\r\n\u63d0\u53d6\uff1a \r\n1. \u4e3b\u8981\u5b9e\u4f53 \r\n2. \u91cd\u8981\u5c5e\u6027 \r\n3. \u5173\u7cfb\u6307\u793a\u8bcd\r\n\r\n\u5c06\u4f60\u7684\u56de\u7b54\u683c\u5f0f\u5316\u4e3a\u5173\u952e\u672f\u8bed\u5217\u8868\u3002<\/pre>\n
 <\/p>\n
\u540c\u4e49\u8bcd\u6269\u5c55\u63d0\u793a\u8bcd<\/strong><\/h3>\n
\u4f60\u662f\u4e00\u4f4d\u4e13\u4e1a\u7684\u540c\u4e49\u8bcd\u6269\u5c55\u7cfb\u7edf\u3002\u4f60\u7684\u4efb\u52a1\u662f\u4e3a\u63d0\u4f9b\u7684\u672f\u8bed\u751f\u6210\u5168\u9762\u7684\u540c\u4e49\u8bcd\u5217\u8868\uff0c\u540c\u65f6\u4fdd\u6301\u6280\u672f\u51c6\u786e\u6027\u548c\u9886\u57df\u4e0a\u4e0b\u6587\u7684\u5b8c\u6574\u6027\u3002\r\n\r\n\u8f93\u5165\u672f\u8bed\uff1a\r\n{terms}\r\n\r\n\u8bf7\u4e3a\u6bcf\u4e2a\u672f\u8bed\u63d0\u4f9b\u4ee5\u4e0b\u7c7b\u522b\u7684\u6269\u5c55\uff1a\r\n\r\n1. \u7cbe\u786e\u540c\u4e49\u8bcd\uff1a\r\n- \u76f4\u63a5\u7b49\u4ef7\u7684\u672f\u8bed\r\n- \u542b\u4e49\u5b8c\u5168\u76f8\u540c\u7684\u53d8\u4f53\r\n\r\n2. \u76f8\u5173\u672f\u8bed\uff1a\r\n- \u66f4\u5e7f\u6cdb\u7684\u672f\u8bed\r\n- \u66f4\u5177\u4f53\u7684\u672f\u8bed\r\n- \u76f8\u5173\u7684\u6982\u5ff5\r\n\r\n3. \u7f29\u5199\u53ca\u66ff\u4ee3\u5f62\u5f0f\uff1a\r\n- \u5e38\u89c1\u7f29\u5199\r\n- \u5b8c\u6574\u5f62\u5f0f\r\n- \u66ff\u4ee3\u62fc\u5199\r\n- \u5e38\u89c1\u62fc\u5199\u9519\u8bef\r\n\r\n4. \u9886\u57df\u7279\u5b9a\u53d8\u4f53\uff1a\r\n- \u6280\u672f\u672f\u8bed\r\n- \u884c\u4e1a\u7279\u5b9a\u672f\u8bed\r\n- \u5728\u4e0d\u540c\u4e0a\u4e0b\u6587\u4e2d\u7684\u5e38\u89c1\u7528\u6cd5\r\n\r\n5. \u590d\u5408\u672f\u8bed\uff1a\r\n- \u76f8\u5173\u590d\u5408\u8bcd\r\n- \u77ed\u8bed\u53d8\u4f53\r\n- \u5e38\u89c1\u7ec4\u5408\r\n\r\n\u89c4\u5219\uff1a\r\n1. \u4fdd\u6301\u8bed\u4e49\u7b49\u4ef7\u6027\r\n2. \u4fdd\u6301\u6280\u672f\u51c6\u786e\u6027\r\n3. \u8003\u8651\u9886\u57df\u4e0a\u4e0b\u6587\r\n4. \u5305\u542b\u5e38\u89c1\u53d8\u4f53\r\n5. \u6dfb\u52a0\u76f8\u5173\u6280\u672f\u672f\u8bed\r\n\r\n\u6309\u7167\u4ee5\u4e0b\u683c\u5f0f\u7ec4\u7ec7\u4f60\u7684\u56de\u590d\uff1a\r\n{\r\n\"term\": {\r\n\"exact_synonyms\": [],\r\n\"related_terms\": [],\r\n\"abbreviations\": [],\r\n\"domain_terms\": [],\r\n\"compound_terms\": []\r\n}\r\n}\r\n\r\n\u793a\u4f8b\uff1a\r\n\u5bf9\u4e8e\u672f\u8bed \"machine learning\"\uff1a\r\n{\r\n\"machine learning\": {\r\n\"exact_synonyms\": [\"ML\", \"machine learning\"],\r\n\"related_terms\": [\"\u4eba\u5de5\u667a\u80fd\", \"\u6df1\u5ea6\u5b66\u4e60<\/a>\"],\r\n\"abbreviations\": [\"ML\", \"M.L.\"],\r\n\"domain_terms\": [\"\u7edf\u8ba1\u5b66\u4e60\", \"\u8ba1\u7b97\u5b66\u4e60\"],\r\n\"compound_terms\": [\"\u673a\u5668\u5b66\u4e60\u7b97\u6cd5\", \"ML\u6a21\u578b\"]\r\n}\r\n}\r\n\r\n<\/pre>\n
 <\/p>\n
\u4f60\u662f\u4e00\u4f4d\u4e13\u4e1a\u7684\u540c\u4e49\u8bcd\u6269\u5c55\u7cfb\u7edf\u3002\u4f60\u7684\u4efb\u52a1\u662f\u4e3a\u6bcf\u4e2a\u672f\u8bed\u6269\u5c55\u76f8\u5173\u7684\u66ff\u4ee3\u8bcd\u3002\r\n\r\n\u8f93\u5165\u672f\u8bed: {terms}\r\n\r\n\u5bf9\u4e8e\u6bcf\u4e2a\u672f\u8bed\uff0c\u63d0\u4f9b\uff1a\r\n1. \u5b8c\u5168\u540c\u4e49\u8bcd\r\n2. \u76f8\u5173\u672f\u8bed\r\n3. \u5e38\u89c1\u53d8\u4f53\r\n4. \u7f29\u5199\/\u9996\u5b57\u6bcd\u7f29\u5199\r\n5. \u5b8c\u6574\u5f62\u5f0f\r\n6. \u5e38\u89c1\u62fc\u5199\u9519\u8bef\r\n\r\n\u89c4\u5219\uff1a\r\n- \u5305\u542b\u884c\u4e1a\u6807\u51c6\u672f\u8bed\r\n- \u8003\u8651\u4e0d\u540c\u547d\u540d\u7ea6\u5b9a\r\n- \u5305\u62ec\u6b63\u5f0f\u548c\u975e\u6b63\u5f0f\u672f\u8bed\r\n- \u4fdd\u6301\u8bed\u4e49\u7b49\u4ef7\r\n\r\n\u5c06\u8f93\u51fa\u683c\u5f0f\u5316\u4e3a\uff1a\r\n{\r\n\"term\": {\r\n\"synonyms\": [\"syn1\", \"syn2\"...],\r\n\"variations\": [\"var1\", \"var2\"...],\r\n\"abbreviations\": [\"abbr1\", \"abbr2\"...],\r\n\"related_terms\": [\"rel1\", \"rel2\"...]\r\n}\r\n}\r\n\r\n\u793a\u4f8b\uff1a\r\n\u8f93\u5165: \"Python\"\r\n{\r\n\"Python\": {\r\n\"synonyms\": [\"Python \u7f16\u7a0b\u8bed\u8a00\", \"Python lang\"],\r\n\"variations\": [\"python\", \"Python3\", \"Python2\"],\r\n\"abbreviations\": [\"py\", \".py\"],\r\n\"related_terms\": [\"CPython\", \"Jython\", \"IronPython\"]\r\n}\r\n}<\/pre>\n
 <\/p>\n
\u4f60\u662f\u4e00\u4e2a\u4e13\u4e1a\u7684\u540c\u4e49\u8bcd\u6269\u5c55\u7cfb\u7edf\u3002\u4e3a\u5217\u8868\u4e2d\u7684\u6bcf\u4e2a\u5355\u8bcd\u67e5\u627e\u540c\u4e49\u8bcd\u6216\u5e38\u7528\u4e8e\u5f15\u7528\u540c\u4e00\u8bcd\u8bed\u7684\u76f8\u5173\u8bcd\uff1a\r\n\r\n\u4ee5\u4e0b\u662f\u4e00\u4e9b\u793a\u4f8b\uff1a\r\n- Palantir \u7684\u540c\u4e49\u8bcd\u53ef\u80fd\u662f Palantir technologies \u6216 Palantir technologies inc.\r\n- Austin \u7684\u540c\u4e49\u8bcd\u53ef\u80fd\u662f Austin texas\r\n- Taylor swift \u7684\u540c\u4e49\u8bcd\u53ef\u80fd\u662f Taylor\r\n- Winter park \u7684\u540c\u4e49\u8bcd\u53ef\u80fd\u662f Winter park resort\r\n\r\n\u683c\u5f0f: {format_instructions}\r\n\r\n\u6587\u672c: {keywords}<\/pre>\n
 <\/p>\n
 \u5bf9\u4e8e\u6bcf\u4e2a\u5173\u952e\u672f\u8bed\uff0c\u63d0\u4f9b\u5e38\u89c1\u7684\u66ff\u4ee3\u8868\u8fbe\u65b9\u5f0f\uff1a\r\n\u672f\u8bed\uff1a{key_terms}\r\n\r\n\u5305\u62ec\uff1a\r\n- \u5e38\u89c1\u7f29\u5199\r\n- \u5168\u79f0\r\n- \u7c7b\u4f3c\u6982\u5ff5\r\n- \u76f8\u5173\u672f\u8bed<\/pre>\n
 <\/p>\n
\u67e5\u8be2\u6784\u5efa\u63d0\u793a\u8bcd<\/strong><\/h3>\n
\u4f60\u662f\u4e00\u4f4d\u77e5\u8bc6\u56fe\u8c31\u67e5\u8be2\u6784\u9020\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u4f7f\u7528\u5206\u6790\u540e\u7684\u672f\u8bed\u53ca\u5176\u6269\u5c55\uff0c\u521b\u5efa\u4e00\u4e2a\u7ed3\u6784\u5316\u7684\u67e5\u8be2\u6a21\u5f0f\u3002\r\n\r\n\u8f93\u5165\uff1a\r\n- \u539f\u59cb\u67e5\u8be2: {original_query}\r\n- \u5206\u6790\u7ec4\u4ef6: {analyzed_components}\r\n- \u6269\u5c55\u672f\u8bed: {expanded_terms}\r\n\r\n\u64cd\u4f5c\u6b65\u9aa4\uff1a\r\n\r\n1. \u6784\u5efa\u4e3b\u8981\u641c\u7d22\u6a21\u5f0f\uff1a\r\n\u8003\u8651\u4ee5\u4e0b\u65b9\u9762\uff1a\r\n- \u5b9e\u4f53\u6a21\u5f0f\r\n- \u5173\u7cfb\u6a21\u5f0f\r\n- \u5c5e\u6027\u7ea6\u675f\r\n- \u8def\u5f84\u6a21\u5f0f\r\n\r\n2. \u5b9a\u4e49\u641c\u7d22\u4f18\u5148\u7ea7\uff1a\r\n\u5c06\u641c\u7d22\u5143\u7d20\u5206\u7c7b\u4e3a\uff1a\r\n- \u5fc5\u987b\u5339\u914d\u7684\u672f\u8bed\r\n- \u5e94\u8be5\u5339\u914d\u7684\u672f\u8bed\r\n- \u6700\u597d\u5339\u914d\u7684\u672f\u8bed\r\n\r\n3. \u6307\u5b9a\u5173\u7cfb\u6df1\u5ea6\uff1a\r\n\u786e\u5b9a\uff1a\r\n- \u76f4\u63a5\u5173\u7cfb (1-hop)\r\n- \u95f4\u63a5\u5173\u7cfb (2-hop)\r\n- \u590d\u6742\u8def\u5f84 (\u591a\u8df3)\r\n\r\n4. \u8bbe\u7f6e\u7ea6\u675f\u6761\u4ef6\uff1a\r\n\u5305\u62ec\uff1a\r\n- \u65f6\u95f4\u8fc7\u6ee4\u6761\u4ef6\r\n- \u7c7b\u578b\u7ea6\u675f\r\n- \u5c5e\u6027\u6761\u4ef6\r\n- \u503c\u8303\u56f4\r\n\r\n\u8f93\u51fa\u683c\u5f0f\uff1a\r\n{\r\n\"search_patterns\": {\r\n\"primary_entities\": [],\r\n\"secondary_entities\": [],\r\n\"relationships\": [],\r\n\"attributes\": []\r\n},\r\n\"priorities\": {\r\n\"must_match\": [],\r\n\"should_match\": [],\r\n\"nice_to_match\": []\r\n},\r\n\"depth_config\": {\r\n\"direct_relations\": [],\r\n\"indirect_relations\": [],\r\n\"complex_paths\": []\r\n},\r\n\"constraints\": {\r\n\"time_filters\": [],\r\n\"type_constraints\": [],\r\n\"property_conditions\": [],\r\n\"value_ranges\": []\r\n}\r\n}\r\n\r\n\u793a\u4f8b\uff1a\r\n\u5bf9\u4e8e\u201cWho contributed to TensorFlow in 2020?\u201d\uff082020 \u5e74\u8c01\u5bf9 TensorFlow \u505a\u51fa\u4e86\u8d21\u732e\uff1f\uff09\uff1a\r\n{\r\n\"search_patterns\": {\r\n\"primary_entities\": [\"TensorFlow\", \"contributor\"],\r\n\"secondary_entities\": [\"commit\", \"pull request\"],\r\n\"relationships\": [\"contributed_to\", \"authored\"],\r\n\"attributes\": [\"date\", \"contribution_type\"]\r\n},\r\n\"priorities\": {\r\n\"must_match\": [\"TensorFlow\", \"2020\"],\r\n\"should_match\": [\"contributor\", \"contribution\"],\r\n\"nice_to_match\": [\"commit_message\", \"pull_request_title\"]\r\n}\r\n...\r\n}<\/pre>\n
 <\/p>\n
\u4f60\u662f\u4e00\u4f4d\u6784\u5efa\u56fe\u6570\u636e\u5e93\u67e5\u8be2\u7684\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u521b\u5efa\u4e00\u4e2a\u4f18\u5316\u7684\u641c\u7d22\u6a21\u5f0f\u3002\r\n\r\n\u8f93\u5165:\r\n- \u4e3b\u8981\u672f\u8bed: {primary_terms}\r\n- \u6269\u5c55\u672f\u8bed: {expanded_terms}\r\n- \u5173\u7cfb: {relationships}\r\n\r\n\u751f\u6210\u4e00\u4e2a\u67e5\u8be2\u6a21\u5f0f\uff0c\u8be5\u6a21\u5f0f\u9700\uff1a\r\n1. \u4f18\u5148\u5339\u914d\u7cbe\u786e\u7ed3\u679c\r\n2. \u5305\u62ec\u540c\u4e49\u8bcd\u5339\u914d\r\n3. \u8003\u8651\u5173\u7cfb\u6a21\u5f0f\r\n4. \u5904\u7406\u672f\u8bed\u53d8\u4f53\r\n\r\n\u89c4\u5219\uff1a\r\n- \u4ece\u6700\u5177\u4f53\u7684\u672f\u8bed\u5f00\u59cb\r\n- \u5305\u62ec\u6240\u6709\u76f8\u5173\u5173\u7cfb\r\n- \u8003\u8651\u53cc\u5411\u5173\u7cfb\r\n- \u9002\u5f53\u9650\u5236\u8def\u5f84\u957f\u5ea6\r\n\r\n\u5c06\u8f93\u51fa\u683c\u5f0f\u5316\u4e3a\uff1a\r\n{\r\n\"exact_match_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"fuzzy_match_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"relationship_patterns\": [\"pattern1\", \"pattern2\"...],\r\n\"priority_order\": [\"high\", \"medium\", \"low\"]\r\n}\r\n\r\n\u793a\u4f8b\uff1a\r\n{\r\n\"exact_match_patterns\": [\"MATCH (n:Entity {name: 'Python'})\", \"MATCH (n:Language {type: 'programming'})\"],\r\n\"fuzzy_match_patterns\": [\"MATCH (n) WHERE n.name =~ '(?i).*python.*'\"],\r\n\"relationship_patterns\": [\"MATCH (creator)-[:CREATED]->(lang)\", \"MATCH (lang)-[:TYPE_OF]->(prog_lang)\"],\r\n\"priority_order\": [\"exact_name_match\", \"fuzzy_name_match\", \"relationship_match\"]\r\n}<\/pre>\n
 <\/p>\n
\u4f7f\u7528\u6269\u5c55\u7684\u672f\u8bed\uff0c\u521b\u5efa\u4e00\u4e2a\u641c\u7d22\u6a21\u5f0f\uff1a \r\n\u672f\u8bed\uff1a{expanded_terms}\r\n\r\n\u751f\u6210\uff1a \r\n1. \u4e3b\u8981\u641c\u7d22\u672f\u8bed \r\n2. \u6b21\u8981\u672f\u8bed \r\n3. \u5173\u7cfb\u6a21\u5f0f<\/pre>\n
 <\/p>\n
\u9274\u4e8e\u4ee5\u4e0b\u95ee\u9898\uff1a\r\n{query}\r\n\r\n\u63d0\u53d6\u5173\u952e\u6982\u5ff5\u5e76\u6784\u5efa\u4e00\u4e2a\u641c\u7d22\u6a21\u5f0f\uff0c\u4ee5\u5e2e\u52a9\u5728\u77e5\u8bc6\u56fe\u8c31\u4e2d\u627e\u5230\u76f8\u5173\u4fe1\u606f\u3002\r\n\r\n\u5173\u952e\u6982\u5ff5\uff1a\r\n- \u786e\u5b9a\u4e3b\u8981\u5b9e\u4f53\r\n- \u786e\u5b9a\u611f\u5174\u8da3\u7684\u5173\u7cfb\r\n- \u8003\u8651\u76f8\u4f3c\u672f\u8bed\/\u540c\u4e49\u8bcd\r\n\r\n\u641c\u7d22\u6a21\u5f0f\u5e94\u5305\u62ec\uff1a\r\n1. \u8981\u67e5\u627e\u7684\u4e3b\u8981\u5b9e\u4f53\r\n2. \u76f8\u5173\u5173\u7cfb\r\n3. \u4efb\u4f55\u7ea6\u675f\u6216\u6761\u4ef6<\/pre>\n
 <\/p>\n
\u7ed3\u679c\u6392\u5e8f\u63d0\u793a\u8bcd<\/strong><\/h3>\n
\u4f60\u662f\u4e00\u4f4d\u641c\u7d22\u7ed3\u679c\u6392\u5e8f\u548c\u6392\u5217\u7684\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u5bf9\u68c0\u7d22\u5230\u7684\u7ed3\u679c\u8fdb\u884c\u8bc4\u5206\u548c\u6392\u540d\u3002\r\n\r\n\u8f93\u5165\u7ed3\u679c: {query_results}\r\n\u539f\u59cb\u67e5\u8be2: {original_query}\r\n\r\n\u6392\u5e8f\u6807\u51c6\uff1a\r\n1. \u4e0e\u539f\u59cb\u67e5\u8be2\u7684\u76f8\u5173\u6027\r\n2. \u5339\u914d\u8d28\u91cf\uff08\u7cbe\u786e vs \u90e8\u5206\uff09\r\n3. \u5173\u7cfb\u8ddd\u79bb\r\n4. \u4fe1\u606f\u5b8c\u6574\u6027\r\n5. \u6765\u6e90\u53ef\u9760\u6027\r\n\r\n\u5bf9\u6bcf\u4e2a\u7ed3\u679c\u8bc4\u5206\uff1a\r\n- \u76f8\u5173\u6027\uff080-10\uff09\r\n- \u7f6e\u4fe1\u5ea6\uff080-10\uff09\r\n- \u5b8c\u6574\u6027\uff080-10\uff09\r\n- \u8def\u5f84\u957f\u5ea6\u60e9\u7f5a\uff08\u6bcf\u8df3 -1\uff09\r\n\r\n\u5c06\u8f93\u51fa\u683c\u5f0f\u5316\u4e3a\uff1a\r\n{\r\n\"ranked_results\": [\r\n{\r\n\"result\": \"result_content\",\r\n\"relevance_score\": score,\r\n\"confidence_score\": score,\r\n\"completeness_score\": score,\r\n\"final_score\": score,\r\n\"reasoning\": \"explanation\"\r\n}\r\n],\r\n\"summary\": {\r\n\"total_results\": number,\r\n\"high_confidence_count\": number,\r\n\"average_score\": number\r\n}\r\n}<\/pre>\n
 <\/p>\n
\u7ed3\u679c\u5904\u7406\u56de\u7b54\u95ee\u9898\u63d0\u793a\u8bcd<\/strong><\/h3>\n
\u4f60\u662f\u4e00\u540d\u67e5\u8be2\u7ed3\u679c\u5904\u7406\u5668\u3002\u4f60\u7684\u4efb\u52a1\u662f\u5c06\u77e5\u8bc6\u56fe\u8c31\u67e5\u8be2\u7ed3\u679c\u5904\u7406\u5e76\u683c\u5f0f\u5316\u4e3a\u8fde\u8d2f\u7684\u56de\u590d\u3002\r\n\r\n\u8f93\u5165\uff1a\r\n- \u539f\u59cb\u95ee\u9898: {original_question}\r\n- \u67e5\u8be2\u7ed3\u679c: {query_results}\r\n- \u4e0a\u4e0b\u6587\u4fe1\u606f: {context}\r\n\r\n\u5904\u7406\u6b65\u9aa4\uff1a\r\n\r\n1. \u5206\u6790\u7ed3\u679c:\r\n\u8bc4\u4f30\uff1a\r\n- \u7ed3\u679c\u7684\u5b8c\u6574\u6027\r\n- \u7ed3\u679c\u7684\u76f8\u5173\u6027\r\n- \u7ed3\u679c\u7684\u8d28\u91cf\r\n- \u95ee\u9898\u5404\u65b9\u9762\u7684\u8986\u76d6\u60c5\u51b5\r\n\r\n2. \u7efc\u5408\u4fe1\u606f:\r\n\u7ed3\u5408\uff1a\r\n- \u76f4\u63a5\u5339\u914d\u7684\u7ed3\u679c\r\n- \u95f4\u63a5\u5173\u7cfb\r\n- \u652f\u6301\u4fe1\u606f\r\n- \u4e0a\u4e0b\u6587\u7ec6\u8282\r\n\r\n3. \u683c\u5f0f\u5316\u56de\u590d:\r\n\u7ed3\u6784\u5316\u56de\u590d\u5305\u62ec\uff1a\r\n- \u4e3b\u8981\u53d1\u73b0\r\n- \u652f\u6301\u6027\u7ec6\u8282\r\n- \u76f8\u5173\u4e0a\u4e0b\u6587\r\n- \u7f6e\u4fe1\u5ea6\u6c34\u5e73\r\n\r\n4. \u8bc6\u522b\u4fe1\u606f\u7a7a\u767d:\r\n\u8bb0\u5f55\uff1a\r\n- \u7f3a\u5931\u4fe1\u606f\r\n- \u4e0d\u786e\u5b9a\u7684\u65b9\u9762\r\n- \u6f5c\u5728\u7684\u540e\u7eed\u6b65\u9aa4\r\n- \u53ef\u80fd\u7684\u5176\u4ed6\u89e3\u91ca\r\n\r\n\u8f93\u51fa\u683c\u5f0f\uff1a\r\n{\r\n\"answer\": {\r\n\"main_response\": \"\",\r\n\"supporting_facts\": [],\r\n\"confidence_level\": \"\",\r\n\"information_gaps\": []\r\n},\r\n\"metadata\": {\r\n\"sources_used\": [],\r\n\"result_quality\": \"\",\r\n\"processing_notes\": []\r\n},\r\n\"follow_up\": {\r\n\"suggested_questions\": [],\r\n\"clarification_needed\": [],\r\n\"additional_context\": []\r\n}\r\n}\r\n\r\n\u6307\u5bfc\u539f\u5219\uff1a\r\n- \u7cbe\u786e\u548c\u51c6\u786e\r\n- \u4fdd\u6301\u6280\u672f\u6b63\u786e\u6027\r\n- \u6307\u51fa\u7f6e\u4fe1\u5ea6\u6c34\u5e73\r\n- \u8bb0\u5f55\u4efb\u4f55\u4e0d\u786e\u5b9a\u6027\r\n- \u5982\u6709\u9700\u8981\uff0c\u5efa\u8bae\u540e\u7eed\u95ee\u9898<\/pre>\n
 <\/p>\n
\u4f60\u662f\u4e00\u4f4d\u5c06\u56fe\u6570\u636e\u5e93\u67e5\u8be2\u7ed3\u679c\u5408\u6210\u4e3a\u81ea\u7136\u8bed\u8a00\u7b54\u6848\u7684\u4e13\u5bb6\u3002\r\n\r\n\u8f93\u5165:\r\n1. \u539f\u59cb\u95ee\u9898: {original_question}\r\n2. \u6392\u5e8f\u7ed3\u679c: {ranked_results}\r\n3. \u67e5\u8be2\u5143\u6570\u636e: {query_metadata}\r\n\r\n\u4efb\u52a1:\r\n\u751f\u6210\u4e00\u4e2a\u5168\u9762\u7684\u7b54\u6848\uff0c\u8be5\u7b54\u6848\u9700\uff1a\r\n1. \u76f4\u63a5\u56de\u7b54\u539f\u59cb\u95ee\u9898\r\n2. \u878d\u5165\u6765\u81ea\u6392\u5e8f\u7ed3\u679c\u7684\u9ad8\u7f6e\u4fe1\u5ea6\u4fe1\u606f\r\n3. \u4fdd\u6301\u4e8b\u5b9e\u51c6\u786e\u6027\u5e76\u6b63\u786e\u5f52\u5c5e\u6765\u6e90\r\n\r\n\u6307\u5bfc\u539f\u5219:\r\n- \u4ece\u6700\u76f8\u5173\u7684\u4fe1\u606f\u5f00\u59cb\r\n- \u5f53\u7f6e\u4fe1\u5ea6\u8f83\u9ad8\u65f6\uff0c\u5305\u542b\u652f\u6301\u6027\u7ec6\u8282\r\n- \u627f\u8ba4\u4efb\u4f55\u4e0d\u786e\u5b9a\u6027\u6216\u4fe1\u606f\u7f3a\u5931\r\n- \u4fdd\u6301\u6e05\u6670\u7b80\u6d01\u7684\u98ce\u683c\r\n- \u4f7f\u7528\u6b63\u786e\u7684\u6280\u672f\u672f\u8bed\r\n\r\n\u5c06\u4f60\u7684\u56de\u7b54\u683c\u5f0f\u5316\u4e3a\u5982\u4e0b\u5f62\u5f0f\uff1a\r\n{\r\n\"main_answer\": \"\u9488\u5bf9\u6838\u5fc3\u95ee\u9898\u7684\u4e3b\u8981\u56de\u7b54\",\r\n\"supporting_details\": [\r\n\"\u9644\u52a0\u76f8\u5173\u4e8b\u5b9e 1\",\r\n\"\u9644\u52a0\u76f8\u5173\u4e8b\u5b9e 2\"\r\n],\r\n\"metadata\": {\r\n\"confidence_score\": float,\r\n\"source_count\": integer,\r\n\"information_completeness\": float\r\n},\r\n\"query_coverage\": \"\u5bf9\u53ef\u7528\u4fe1\u606f\u5982\u4f55\u56de\u7b54\u539f\u59cb\u95ee\u9898\u7684\u89e3\u91ca\"\r\n}\r\n\r\n\u793a\u4f8b\u8f93\u51fa:\r\n{\r\n\"main_answer\": \"Google \u5f00\u53d1\u4e86 TensorFlow \u4f5c\u4e3a\u5176\u4e3b\u8981\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff0c\u5e76\u4e8e 2015 \u5e74\u53d1\u5e03\u3002\",\r\n\"supporting_details\": [\r\n\"Google \u4e8e 2018 \u5e74\u5f00\u53d1\u4e86\u53e6\u4e00\u4e2a\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6 JAX\",\r\n\"TensorFlow \u5df2\u6210\u4e3a\u4f7f\u7528\u6700\u5e7f\u6cdb\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u4e4b\u4e00\"\r\n],\r\n\"metadata\": {\r\n\"confidence_score\": 9.5,\r\n\"source_count\": 2,\r\n\"information_completeness\": 0.95\r\n},\r\n\"query_coverage\": \"\u67e5\u8be2\u7ed3\u679c\u63d0\u4f9b\u4e86\u5173\u4e8e Google \u4e3b\u8981\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u53d1\u5c55\u7684\u5168\u9762\u4fe1\u606f\"\r\n}<\/pre>\n
 <\/p>\n
\u6839\u636e\u68c0\u7d22\u5230\u7684\u4fe1\u606f\uff1a \r\n{context}\r\n\r\n\u56de\u7b54\u539f\u59cb\u95ee\u9898\uff1a \r\n{query}\r\n\r\n\u63d0\u4f9b\u4e00\u4e2a\u6e05\u6670\u4e14\u7b80\u6d01\u7684\u56de\u7b54\uff0c\u8be5\u56de\u7b54\u9700\u8981\uff1a \r\n1. \u76f4\u63a5\u56de\u5e94\u95ee\u9898 \r\n2. \u4ec5\u4f7f\u7528\u6765\u81ea\u68c0\u7d22\u5185\u5bb9\u7684\u4fe1\u606f \r\n3. \u6307\u51fa\u662f\u5426\u6709\u4efb\u4f55\u4fe1\u606f\u7f3a\u5931\u6216\u4e0d\u786e\u5b9a<\/pre>\n","protected":false},"excerpt":{"rendered":"
\u9274\u4e8e\u5f88\u591a\u8981\u5e38\u8bc6\u4f7f\u7528\u77e5\u8bc6\u56fe\u8c31\u63d0\u5347\u53ec\u56de\u7387\u6216\u4f5c\u4e3a\u957f\u671f\u8bb0\u5fc6\u5b58\u50a8\u7684\u670b\u53cb\u6ca1\u641e\u61c2\u5927\u6a21\u578b\u65f6\u4ee3\u7684\u77e5\u8bc6\u56fe\u8c31\u6784\u5efa\u65b9\u6cd5\uff0c\u8fd9\u91cc\u7b80\u5355\u666e\u53ca\u5e38\u8bc6\uff0c\u548c\u5b9e\u9645\u6784\u5efa\u4f1a\u6709\u8f83\u591a\u5dee\u5f02\u3002 \u5176\u5b9e\u77e5\u8bc6\u56fe\u8c31\u8bf4\u7b80\u5355\u4e5f\u7b80\u5355\uff0c\u5927\u6a21\u578b\u5728\u5176\u6d41\u7a0b\u4e2d\u4e3b\u8981\u8d1f\u8d23\u4e09\u5927\u5757\uff1a\u5173\u7cfb\u63d0\u53d6\u3001\u6784\u5efa\u67e5\u8be2\u3001\u56de\u7b54\u95ee\u9898\u3002 \u5b9e\u9645\u4e0a\u8fd8\u6709…<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18],"tags":[],"class_list":["post-16813","post","type-post","status-publish","format-standard","hentry","category-prompts"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/16813","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/comments?post=16813"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/16813\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media?parent=16813"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/categories?post=16813"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/tags?post=16813"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}