From bfcf84001343e20bf272a47f76c3820106c79b60 Mon Sep 17 00:00:00 2001 From: along <1015042407@qq.com> Date: Wed, 1 Apr 2026 15:55:53 +0800 Subject: [PATCH] =?UTF-8?q?=E4=BF=AE=E6=94=B9GraphRAG=E5=90=8E=E7=AB=AF?= =?UTF-8?q?=EF=BC=8C=E5=8F=A6=E5=A4=96=E4=B8=A4=E6=9C=AC=E5=B0=8F=E8=AF=B4?= =?UTF-8?q?=E5=85=A5=E5=BA=93?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- backend/app.py | 11 +-- backend/graph_builder.py | 160 ++++++++++++++++++++++++++++++--------- backend/prompts.py | 61 ++++++++------- backend/run_import.py | 8 +- 4 files changed, 170 insertions(+), 70 deletions(-) diff --git a/backend/app.py b/backend/app.py index c764c2e..5003825 100644 --- a/backend/app.py +++ b/backend/app.py @@ -1,5 +1,5 @@ """ -大唐双龙传 GraphRAG — FastAPI 后端 +武侠三部曲 GraphRAG — FastAPI 后端 端点: GET /api/health — 健康检查（含 Neo4j 连通性） @@ -13,11 +13,11 @@ from fastapi.middleware.cors import CORSMiddleware from pydantic import BaseModel from graph_query import get_driver, get_graph_stats -from graph_builder import build_graph +from graph_builder import build_all_graphs from llm_router import answer_question import uvicorn -app = FastAPI(title="大唐双龙传 GraphRAG API", version="1.0.0") +app = FastAPI(title="武侠三部曲 GraphRAG API", version="1.1.0") app.add_middleware( CORSMiddleware, @@ -37,6 +37,7 @@ class ChatRequest(BaseModel): class ImportRequest(BaseModel): clear: bool = False # True = 先清空图谱再重新导入 + novels: list[str] | None = None # 默认导入 dtslz/ldj/tlbb # ── Endpoints ───────────────────────────────────────────── @@ -62,10 +63,10 @@ def stats(): @app.post("/api/import") def import_data(req: ImportRequest = ImportRequest()): - """导入所有卷数据到 Neo4j（耗时约 1-3 分钟，请勿重复调用）""" + """导入小说数据到 Neo4j（耗时约 1-3 分钟，请勿重复调用）""" driver = get_driver() try: - build_graph(driver, clear=req.clear) + build_all_graphs(driver, novels=req.novels, clear=req.clear) stats = get_graph_stats() return {"status": "ok", "stats": stats} except Exception as e: diff --git a/backend/graph_builder.py b/backend/graph_builder.py index 8e24e8f..df6c8f9 100644 --- a/backend/graph_builder.py +++ b/backend/graph_builder.py @@ -10,7 +10,8 @@ import json from pathlib import Path from neo4j import Driver -DATA_DIR = Path(__file__).parent.parent / "data" +DATA_DIR_BASE = Path(__file__).parent.parent / "fiction" +SUPPORTED_NOVELS = ("dtslz", "ldj", "tlbb") # ── 工具函数 ────────────────────────────────────────────── @@ -26,32 +27,77 @@ def _split_leaders(leader: str) -> list[str]: return [p for p in parts if p not in ("未提及", "")] +def _node_id(novel: str, raw_id: str) -> str: + return f"{novel}:{raw_id}" + + +def _get_data_dir(novel: str) -> Path: + data_dir = DATA_DIR_BASE / novel / "data" + if not data_dir.exists(): + raise ValueError(f"小说数据目录不存在: {data_dir}") + return data_dir + + +def _iter_volume_files(data_dir: Path): + for filepath in sorted(data_dir.glob("vol*.json")): + stem = filepath.stem # vol01 + if len(stem) >= 5 and stem[:3] == "vol" and stem[3:].isdigit(): + yield int(stem[3:]), filepath + + # ── Schema 初始化 ───────────────────────────────────────── +def _drop_legacy_constraints(session): + """兼容旧版本：移除 Character(name) 唯一约束，改为 (novel, name) 复合唯一约束。""" + rows = session.run( + """ + SHOW CONSTRAINTS + YIELD name, labelsOrTypes, properties + RETURN name, labelsOrTypes, properties + """ + ) + for row in rows: + labels = row.get("labelsOrTypes") or [] + properties = row.get("properties") or [] + if labels == ["Character"] and properties == ["name"]: + constraint_name = row["name"].replace("`", "") + session.run(f"DROP CONSTRAINT `{constraint_name}` IF EXISTS") + + def setup_schema(driver: Driver): with driver.session() as s: - s.run("CREATE CONSTRAINT IF NOT EXISTS FOR (n:Character) REQUIRE n.name IS UNIQUE") + _drop_legacy_constraints(s) + s.run("CREATE CONSTRAINT IF NOT EXISTS FOR (n:Character) REQUIRE (n.novel, n.name) IS UNIQUE") s.run("CREATE CONSTRAINT IF NOT EXISTS FOR (n:Location) REQUIRE n.id IS UNIQUE") s.run("CREATE CONSTRAINT IF NOT EXISTS FOR (n:Faction) REQUIRE n.id IS UNIQUE") s.run("CREATE CONSTRAINT IF NOT EXISTS FOR (n:Event) REQUIRE n.id IS UNIQUE") s.run("CREATE INDEX IF NOT EXISTS FOR (e:Event) ON (e.vol)") + s.run("CREATE INDEX IF NOT EXISTS FOR (e:Event) ON (e.novel)") + s.run("CREATE INDEX IF NOT EXISTS FOR (c:Character) ON (c.novel)") + s.run("CREATE INDEX IF NOT EXISTS FOR (l:Location) ON (l.novel)") + s.run("CREATE INDEX IF NOT EXISTS FOR (f:Faction) ON (f.novel)") s.run("CREATE INDEX IF NOT EXISTS FOR ()-[r:VISITED]-() ON (r.vol)") s.run("CREATE INDEX IF NOT EXISTS FOR ()-[r:CONTROLS]-() ON (r.vol)") # ── 各类型导入 ──────────────────────────────────────────── -def _import_locations(session, locations: list[dict]): +def _import_locations(session, novel: str, locations: list[dict]): for loc in locations: + raw_id = loc["id"] session.run( """ MERGE (l:Location {id: $id}) SET l.name = $name, + l.source_id = $source_id, + l.novel = $novel, l.type = $type, l.lat = $lat, l.lng = $lng """, - id=loc["id"], + id=_node_id(novel, raw_id), + source_id=raw_id, + novel=novel, name=loc["name"], type=loc.get("type", ""), lat=loc.get("lat"), @@ -59,14 +105,22 @@ def _import_locations(session, locations: list[dict]): ) -def _import_factions(session, factions: list[dict], vol: int): +def _import_factions(session, novel: str, factions: list[dict], vol: int): for f in factions: + raw_id = f["id"] session.run( """ MERGE (n:Faction {id: $id}) - SET n.name = $name, n.type = $type, n.color = $color + SET n.name = $name, + n.source_id = $source_id, + n.novel = $novel, + n.type = $type, + n.color = $color """, - id=f["id"], name=f["name"], + id=_node_id(novel, raw_id), + source_id=raw_id, + novel=novel, + name=f["name"], type=f.get("type", ""), color=f.get("color", ""), ) @@ -76,9 +130,12 @@ def _import_factions(session, factions: list[dict], vol: int): """ MATCH (fac:Faction {id: $fid}) MATCH (loc:Location {id: $lid}) - MERGE (fac)-[:CONTROLS {vol: $vol}]->(loc) + MERGE (fac)-[:CONTROLS {novel: $novel, vol: $vol}]->(loc) """, - fid=f["id"], lid=loc_id, vol=vol, + fid=_node_id(novel, raw_id), + lid=_node_id(novel, loc_id), + novel=novel, + vol=vol, ) # Faction → HAS_MEMBER → Character @@ -87,36 +144,44 @@ def _import_factions(session, factions: list[dict], vol: int): continue session.run( """ - MERGE (c:Character {name: $name}) + MERGE (c:Character {novel: $novel, name: $name}) WITH c MATCH (fac:Faction {id: $fid}) - MERGE (fac)-[:HAS_MEMBER {vol: $vol}]->(c) + MERGE (fac)-[:HAS_MEMBER {novel: $novel, vol: $vol}]->(c) """, - name=figure, fid=f["id"], vol=vol, + novel=novel, + name=figure, + fid=_node_id(novel, raw_id), + vol=vol, ) # Character → LEADS → Faction for leader_name in _split_leaders(f.get("leader", "")): session.run( """ - MERGE (c:Character {name: $name}) + MERGE (c:Character {novel: $novel, name: $name}) WITH c MATCH (fac:Faction {id: $fid}) - MERGE (c)-[:LEADS {vol: $vol}]->(fac) + MERGE (c)-[:LEADS {novel: $novel, vol: $vol}]->(fac) """, - name=leader_name, fid=f["id"], vol=vol, + novel=novel, + name=leader_name, + fid=_node_id(novel, raw_id), + vol=vol, ) -def _import_routes(session, routes: list[dict], vol: int): +def _import_routes(session, novel: str, routes: list[dict], vol: int): for route in routes: char_color = route.get("color", "") char_names = _split_characters(route["character"]) for char_name in char_names: session.run( - "MERGE (c:Character {name: $name}) SET c.color = $color", - name=char_name, color=char_color, + "MERGE (c:Character {novel: $novel, name: $name}) SET c.color = $color", + novel=novel, + name=char_name, + color=char_color, ) for wp in route.get("route", []): @@ -128,28 +193,37 @@ def _import_routes(session, routes: list[dict], vol: int): session.run( """ - MATCH (c:Character {name: $char}) + MATCH (c:Character {novel: $novel, name: $char}) MATCH (l:Location {id: $lid}) - MERGE (c)-[v:VISITED {vol: $vol, chapter: $chapter}]->(l) + MERGE (c)-[v:VISITED {novel: $novel, vol: $vol, chapter: $chapter}]->(l) SET v.event = $event """, - char=char_name, lid=loc_id, + novel=novel, + char=char_name, + lid=_node_id(novel, loc_id), vol=vol, chapter=chapter, event=event, ) -def _import_events(session, events: list[dict], vol: int): +def _import_events(session, novel: str, events: list[dict], vol: int): for i, evt in enumerate(events): - event_id = f"v{vol:02d}_e{i:03d}" + event_id = _node_id(novel, f"v{vol:02d}_e{i:03d}") chapter = evt.get("chapter", 0) description = evt.get("event", "") session.run( """ MERGE (e:Event {id: $id}) - SET e.vol = $vol, e.chapter = $chapter, e.description = $description + SET e.novel = $novel, + e.vol = $vol, + e.chapter = $chapter, + e.description = $description """, - id=event_id, vol=vol, chapter=chapter, description=description, + id=event_id, + novel=novel, + vol=vol, + chapter=chapter, + description=description, ) # 只在有命名地点 id 时建立关系（lat/lng 条目跳过） @@ -161,13 +235,17 @@ def _import_events(session, events: list[dict], vol: int): MATCH (l:Location {id: $lid}) MERGE (e)-[:OCCURRED_AT]->(l) """, - eid=event_id, lid=loc_ref, + eid=event_id, + lid=_node_id(novel, loc_ref), ) # ── 主入口 ──────────────────────────────────────────────── -def build_graph(driver: Driver, clear: bool = False): +def build_graph(driver: Driver, novel: str, clear: bool = False): + if novel not in SUPPORTED_NOVELS: + raise ValueError(f"不支持的小说标识: {novel}，可选: {', '.join(SUPPORTED_NOVELS)}") + if clear: print("Clearing existing graph data...") with driver.session() as s: @@ -176,22 +254,30 @@ def build_graph(driver: Driver, clear: bool = False): print("Setting up schema constraints and indexes...") setup_schema(driver) + data_dir = _get_data_dir(novel) imported = 0 - for vol_num in range(1, 64): - filepath = DATA_DIR / f"vol{vol_num:02d}.json" - if not filepath.exists(): - continue + print(f"Importing novel: {novel} ({data_dir})") + for vol_num, filepath in _iter_volume_files(data_dir): with open(filepath, encoding="utf-8") as f: data = json.load(f) with driver.session() as session: - _import_locations(session, data.get("locations", [])) - _import_factions(session, data.get("factions", []), vol_num) - _import_routes(session, data.get("character_routes", []), vol_num) - _import_events(session, data.get("key_events", []), vol_num) + _import_locations(session, novel, data.get("locations", [])) + _import_factions(session, novel, data.get("factions", []), vol_num) + _import_routes(session, novel, data.get("character_routes", []), vol_num) + _import_events(session, novel, data.get("key_events", []), vol_num) imported += 1 - print(f" [✓] vol{vol_num:02d} imported") + print(f" [✓] {novel}/vol{vol_num:02d} imported") - print(f"\nDone. Imported {imported} volumes.") + print(f"Done. Imported {imported} volumes for {novel}.\n") + + +def build_all_graphs(driver: Driver, novels: list[str] | None = None, clear: bool = False): + selected = novels or list(SUPPORTED_NOVELS) + if not selected: + raise ValueError("novels 不能为空") + + for i, novel in enumerate(selected): + build_graph(driver, novel=novel, clear=(clear and i == 0)) diff --git a/backend/prompts.py b/backend/prompts.py index 6c8bf20..9b14992 100644 --- a/backend/prompts.py +++ b/backend/prompts.py @@ -1,44 +1,49 @@ SCHEMA_DESCRIPTION = """ -大唐双龙传知识图谱 Schema（Neo4j）： +武侠三部曲知识图谱 Schema（Neo4j）： + +小说标识（novel）： +- dtslz = 大唐双龙传 +- ldj = 鹿鼎记 +- tlbb = 天龙八部 节点类型： -- Character {name, color} - 主要人物：寇仲、徐子陵、宇文化及、傅君婥、宋师道、李靖、石青璇、李密、李子通、 - 杜伏威、跋锋寒、李世民、李渊、宋缺、寇仲、毕玄、阴后 +- Character {novel, name, color} + 同名人物在不同小说中会按 novel 隔离 -- Location {id, name, type, lat, lng} +- Location {id, novel, source_id, name, type, lat, lng} type 取值：city / town / waterway / landmark / grassland / forest / region - 主要城市：扬州(yangzhou)、洛阳(luoyang)、长安/大兴(daxing)、丹阳(danyang)、 - 梁都、历阳(liyang)、江陵 + id 为全局唯一键，格式：{novel}:{source_id} -- Faction {id, name, type, color} +- Faction {id, novel, source_id, name, type, color} type 取值：朝廷 / 门阀 / 义军 / 游牧政权 / 江湖势力 / 地方军阀 / 帮会 / 外族 - 主要势力：隋朝(sui)、李阀(li_clan)、宋阀(song_clan)、宇文阀(yuwen)、 - 瓦岗军(wagang_army)、突厥(turks)、慈航静斋、阴癸派 + id 为全局唯一键，格式：{novel}:{source_id} -- Event {id, vol, chapter, description} - vol 是卷号（整数 1-63），chapter 是章节号 +- Event {id, novel, vol, chapter, description} + id 为全局唯一键，格式：{novel}:vXX_eYYY + vol 是该小说内部卷号（整数），chapter 是章节号 关系类型： -- (Character)-[:VISITED {vol, chapter, event}]->(Location) +- (Character)-[:VISITED {novel, vol, chapter, event}]->(Location) 人物在某卷某章到访某地 -- (Faction)-[:CONTROLS {vol}]->(Location) +- (Faction)-[:CONTROLS {novel, vol}]->(Location) 势力在某卷控制某地 -- (Faction)-[:HAS_MEMBER {vol}]->(Character) +- (Faction)-[:HAS_MEMBER {novel, vol}]->(Character) 势力在某卷拥有某成员 -- (Character)-[:LEADS {vol}]->(Faction) +- (Character)-[:LEADS {novel, vol}]->(Faction) 人物在某卷领导某势力 - (Event)-[:OCCURRED_AT]->(Location) 事件发生于某地 -注意：vol 属性用整数表示（如 vol=1 代表第一卷，vol=20 代表第二十卷） +查询建议： +- 用户明确提到小说名时，务必加 novel 过滤 +- 未指定小说时，可跨小说查询 """ -CYPHER_SYSTEM_PROMPT = f"""你是大唐双龙传知识图谱的 Cypher 查询专家。 +CYPHER_SYSTEM_PROMPT = f"""你是武侠三部曲知识图谱的 Cypher 查询专家。 {SCHEMA_DESCRIPTION} @@ -49,26 +54,30 @@ CYPHER_SYSTEM_PROMPT = f"""你是大唐双龙传知识图谱的 Cypher 查询专 4. 默认加 LIMIT 30，除非用户指定数量 5. 使用 DISTINCT 去重 6. 属性名用 n.name、r.vol 格式，不要用整个节点 -7. 如果问题完全无法用图谱回答，只输出单词：UNSUPPORTED +7. 若问题指定小说，优先使用 novel 过滤： + - 大唐双龙传 => novel = "dtslz" + - 鹿鼎记 => novel = "ldj" + - 天龙八部 => novel = "tlbb" +8. 如果问题完全无法用图谱回答，只输出单词：UNSUPPORTED 示例： Q: 寇仲去过哪些地方？ -A: MATCH (c:Character {{name: "寇仲"}})-[v:VISITED]->(l:Location) RETURN DISTINCT l.name, l.type, min(v.vol) AS first_vol ORDER BY first_vol LIMIT 30 +A: MATCH (c:Character {{novel: "dtslz", name: "寇仲"}})-[v:VISITED]->(l:Location) RETURN DISTINCT l.name, l.type, min(v.vol) AS first_vol ORDER BY first_vol LIMIT 30 Q: 第30卷时宇文阀控制哪些城市？ -A: MATCH (f:Faction {{name: "宇文阀"}})-[r:CONTROLS]->(l:Location) WHERE r.vol <= 30 AND l.type = "city" RETURN DISTINCT l.name, r.vol ORDER BY r.vol LIMIT 30 +A: MATCH (f:Faction {{novel: "dtslz", name: "宇文阀"}})-[r:CONTROLS]->(l:Location) WHERE r.vol <= 30 AND l.type = "city" RETURN DISTINCT l.name, r.vol ORDER BY r.vol LIMIT 30 Q: 扬州发生过哪些重要事件？ -A: MATCH (e:Event)-[:OCCURRED_AT]->(l:Location {{name: "扬州"}}) RETURN e.description, e.vol, e.chapter ORDER BY e.vol, e.chapter LIMIT 30 +A: MATCH (e:Event {{novel: "dtslz"}})-[:OCCURRED_AT]->(l:Location {{novel: "dtslz", name: "扬州"}}) RETURN e.description, e.vol, e.chapter ORDER BY e.vol, e.chapter LIMIT 30 Q: 谁领导过瓦岗军？ -A: MATCH (c:Character)-[r:LEADS]->(f:Faction {{name: "瓦岗军"}}) RETURN DISTINCT c.name, r.vol ORDER BY r.vol LIMIT 30 +A: MATCH (c:Character)-[r:LEADS]->(f:Faction {{novel: "dtslz", name: "瓦岗军"}}) RETURN DISTINCT c.name, r.vol ORDER BY r.vol LIMIT 30 -Q: 寇仲和哪些势力有过关联？ -A: MATCH (c:Character {{name: "寇仲"}})-[:VISITED]->(l:Location)<-[:CONTROLS]-(f:Faction) RETURN DISTINCT f.name, f.type LIMIT 30 +Q: 韦小宝加入过哪些势力？ +A: MATCH (c:Character {{novel: "ldj", name: "韦小宝"}})<-[:HAS_MEMBER]-(f:Faction) RETURN DISTINCT f.name, f.type LIMIT 30 """ -ANSWER_SYSTEM_PROMPT = """你是大唐双龙传的知识问答助手，熟悉小说中的人物、势力、地点和事件。 +ANSWER_SYSTEM_PROMPT = """你是武侠三部曲（大唐双龙传、鹿鼎记、天龙八部）的知识问答助手，熟悉小说中的人物、势力、地点和事件。 请根据知识图谱的查询结果，用中文给出准确、自然的回答： - 直接回答问题，语言简洁流畅 diff --git a/backend/run_import.py b/backend/run_import.py index 4d62237..76a2819 100644 --- a/backend/run_import.py +++ b/backend/run_import.py @@ -5,25 +5,29 @@ 用法: python run_import.py # 增量导入（MERGE，不删除现有数据） python run_import.py --clear # 清空图谱后全量重新导入 + python run_import.py ldj # 仅导入鹿鼎记 + python run_import.py dtslz tlbb --clear # 清空后导入指定小说 """ import sys from dotenv import load_dotenv from graph_query import get_driver -from graph_builder import build_graph +from graph_builder import build_all_graphs, SUPPORTED_NOVELS load_dotenv() def main(): clear = "--clear" in sys.argv + novels = [arg for arg in sys.argv[1:] if not arg.startswith("--")] + selected = novels or list(SUPPORTED_NOVELS) print("Connecting to Neo4j...") driver = get_driver() driver.verify_connectivity() print("Connected.\n") - build_graph(driver, clear=clear) + build_all_graphs(driver, novels=selected, clear=clear) print("\nGraph stats:") from graph_query import get_graph_stats