feat: 提升截图基金名称识别准确率

app/hooks/useFundFuzzyMatcher.js

@@ -0,0 +1,204 @@
+import { useCallback, useRef } from 'react';
+import { cachedRequest, clearCachedRequest } from '../lib/cacheRequest';
+
+const FUND_CODE_SEARCH_URL = 'https://fund.eastmoney.com/js/fundcode_search.js';
+const FUND_LIST_CACHE_KEY = 'eastmoney_fundcode_search_list';
+const FUND_LIST_CACHE_TIME = 24 * 60 * 60 * 1000;
+
+const formatEastMoneyFundList = (rawList) => {
+  if (!Array.isArray(rawList)) return [];
+
+  return rawList
+    .map((item) => {
+      if (!Array.isArray(item)) return null;
+      const code = String(item[0] ?? '').trim();
+      const name = String(item[2] ?? '').trim();
+      if (!code || !name) return null;
+      return { code, name };
+    })
+    .filter(Boolean);
+};
+
+export const useFundFuzzyMatcher = () => {
+  const allFundFuseRef = useRef(null);
+  const allFundLoadPromiseRef = useRef(null);
+
+  const getAllFundFuse = useCallback(async () => {
+    if (allFundFuseRef.current) return allFundFuseRef.current;
+    if (allFundLoadPromiseRef.current) return allFundLoadPromiseRef.current;
+
+    allFundLoadPromiseRef.current = (async () => {
+      const [fuseModule, allFundList] = await Promise.all([
+        import('fuse.js'),
+        cachedRequest(
+          () =>
+            new Promise((resolve, reject) => {
+              if (typeof window === 'undefined' || typeof document === 'undefined' || !document.body) {
+                reject(new Error('NO_BROWSER_ENV'));
+                return;
+              }
+
+              const prevR = window.r;
+              const script = document.createElement('script');
+              script.src = `${FUND_CODE_SEARCH_URL}?_=${Date.now()}`;
+              script.async = true;
+
+              const cleanup = () => {
+                if (document.body.contains(script)) {
+                  document.body.removeChild(script);
+                }
+                if (prevR === undefined) {
+                  try {
+                    delete window.r;
+                  } catch (e) {
+                    window.r = undefined;
+                  }
+                } else {
+                  window.r = prevR;
+                }
+              };
+
+              script.onload = () => {
+                const snapshot = Array.isArray(window.r) ? JSON.parse(JSON.stringify(window.r)) : [];
+                cleanup();
+                const parsed = formatEastMoneyFundList(snapshot);
+                if (!parsed.length) {
+                  reject(new Error('PARSE_ALL_FUND_FAILED'));
+                  return;
+                }
+                resolve(parsed);
+              };
+
+              script.onerror = () => {
+                cleanup();
+                reject(new Error('LOAD_ALL_FUND_FAILED'));
+              };
+
+              document.body.appendChild(script);
+            }),
+          FUND_LIST_CACHE_KEY,
+          { cacheTime: FUND_LIST_CACHE_TIME }
+        ),
+      ]);
+      const Fuse = fuseModule.default;
+      const fuse = new Fuse(Array.isArray(allFundList) ? allFundList : [], {
+        keys: ['name', 'code'],
+        includeScore: true,
+        threshold: 0.5,
+        ignoreLocation: true,
+        minMatchCharLength: 2,
+      });
+
+      allFundFuseRef.current = fuse;
+      return fuse;
+    })();
+
+    try {
+      return await allFundLoadPromiseRef.current;
+    } catch (e) {
+      allFundLoadPromiseRef.current = null;
+      clearCachedRequest(FUND_LIST_CACHE_KEY);
+      throw e;
+    }
+  }, []);
+
+  const normalizeFundText = useCallback((value) => {
+    if (typeof value !== 'string') return '';
+    return value
+      .toUpperCase()
+      .replace(/[（(]/g, '(')
+      .replace(/[）)]/g, ')')
+      .replace(/[·•]/g, '')
+      .replace(/\s+/g, '')
+      .replace(/[^\u4e00-\u9fa5A-Z0-9()]/g, '');
+  }, []);
+
+  const parseFundQuerySignals = useCallback((rawName) => {
+    const normalized = normalizeFundText(rawName);
+    const hasETF = normalized.includes('ETF');
+    const hasLOF = normalized.includes('LOF');
+    const hasLink = normalized.includes('联接');
+    const shareMatch = normalized.match(/([A-Z])(?:类)?$/i);
+    const shareClass = shareMatch ? shareMatch[1].toUpperCase() : null;
+
+    const core = normalized
+      .replace(/基金/g, '')
+      .replace(/ETF联接/g, '')
+      .replace(/联接[A-Z]?/g, '')
+      .replace(/ETF/g, '')
+      .replace(/LOF/g, '')
+      .replace(/[A-Z](?:类)?$/g, '');
+
+    return {
+      normalized,
+      core,
+      hasETF,
+      hasLOF,
+      hasLink,
+      shareClass,
+    };
+  }, [normalizeFundText]);
+
+  const resolveFundCodeByFuzzy = useCallback(async (name) => {
+    const querySignals = parseFundQuerySignals(name);
+    if (!querySignals.normalized) return null;
+
+    const len = querySignals.normalized.length;
+    const strictThreshold = len <= 4 ? 0.16 : len <= 8 ? 0.22 : 0.28;
+    const relaxedThreshold = Math.min(0.45, strictThreshold + 0.16);
+    const scoreGapThreshold = len <= 5 ? 0.08 : 0.06;
+
+    const fuse = await getAllFundFuse();
+    const recalled = fuse.search(name, { limit: 50 });
+    if (!recalled.length) return null;
+
+    const stage1 = recalled.filter((item) => (item.score ?? 1) <= relaxedThreshold);
+    if (!stage1.length) return null;
+
+    const ranked = stage1
+      .map((item) => {
+        const candidateSignals = parseFundQuerySignals(item?.item?.name || '');
+        let finalScore = item.score ?? 1;
+
+        if (querySignals.hasETF) {
+          finalScore += candidateSignals.hasETF ? -0.04 : 0.2;
+        }
+        if (querySignals.hasLOF) {
+          finalScore += candidateSignals.hasLOF ? -0.04 : 0.2;
+        }
+        if (querySignals.hasLink) {
+          finalScore += candidateSignals.hasLink ? -0.03 : 0.18;
+        }
+        if (querySignals.shareClass) {
+          finalScore += candidateSignals.shareClass === querySignals.shareClass ? -0.03 : 0.18;
+        }
+
+        if (querySignals.core && candidateSignals.core) {
+          if (candidateSignals.core.includes(querySignals.core)) {
+            finalScore -= 0.06;
+          } else if (!querySignals.core.includes(candidateSignals.core)) {
+            finalScore += 0.06;
+          }
+        }
+
+        return { ...item, finalScore };
+      })
+      .sort((a, b) => a.finalScore - b.finalScore);
+
+    const top1 = ranked[0];
+    if (!top1 || top1.finalScore > strictThreshold) return null;
+
+    const top2 = ranked[1];
+    if (top2 && (top2.finalScore - top1.finalScore) < scoreGapThreshold) {
+      return null;
+    }
+
+    return top1?.item?.code || null;
+  }, [getAllFundFuse, parseFundQuerySignals]);
+
+  return {
+    resolveFundCodeByFuzzy,
+  };
+};
+
+export default useFundFuzzyMatcher;

app/page.jsx

@@ -49,6 +49,7 @@ import { fetchFundData, fetchLatestRelease, fetchShanghaiIndexDate, fetchSmartFu
 import packageJson from '../package.json';
 import PcFundTable from './components/PcFundTable';
 import MobileFundTable from './components/MobileFundTable';
+import { useFundFuzzyMatcher } from './hooks/useFundFuzzyMatcher';
 
 dayjs.extend(utc);
 dayjs.extend(timezone);
@@ -1173,6 +1174,7 @@ export default function HomePage() {
   const abortScanRef = useRef(false); // 终止扫描标记
   const fileInputRef = useRef(null);
   const ocrWorkerRef = useRef(null);
+  const { resolveFundCodeByFuzzy } = useFundFuzzyMatcher();
 
   const handleScanClick = () => {
     setScanModalOpen(true);
@@ -1327,6 +1329,15 @@ export default function HomePage() {
               if (found && found.CODE) {
                 allCodes.add(found.CODE);
               }
+            } else {
+              // 使用 fuse.js 读取 Public 中的 allFunds 数据进行模糊匹配，补充搜索接口的不足
+              try {
+                const fuzzyCode = await resolveFundCodeByFuzzy(name);
+                if (fuzzyCode) {
+                  allCodes.add(fuzzyCode);
+                }
+              } catch (e) {
+              }
             }
           } catch (e) {
           }

package-lock.json

@@ -19,6 +19,7 @@
         "chart.js": "^4.5.1",
         "dayjs": "^1.11.19",
         "framer-motion": "^12.29.2",
+        "fuse.js": "^7.1.0",
         "lodash": "^4.17.23",
         "next": "^16.1.5",
         "react": "18.3.1",
@@ -4149,6 +4150,15 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
+    "node_modules/fuse.js": {
+      "version": "7.1.0",
+      "resolved": "https://registry.npmmirror.com/fuse.js/-/fuse.js-7.1.0.tgz",
+      "integrity": "sha512-trLf4SzuuUxfusZADLINj+dE8clK1frKdmqiJNb1Es75fmI5oY6X2mxLVUciLLjxqw/xr72Dhy+lER6dGd02FQ==",
+      "license": "Apache-2.0",
+      "engines": {
+        "node": ">=10"
+      }
+    },
     "node_modules/generator-function": {
       "version": "2.0.1",
       "resolved": "https://registry.npmjs.org/generator-function/-/generator-function-2.0.1.tgz",

package.json

@@ -22,6 +22,7 @@
     "chart.js": "^4.5.1",
     "dayjs": "^1.11.19",
     "framer-motion": "^12.29.2",
+    "fuse.js": "^7.1.0",
     "lodash": "^4.17.23",
     "next": "^16.1.5",
     "react": "18.3.1",