@natlibfi/melinda-record-matching 5.0.1-alpha.1 → 5.0.2-alpha.1

package/dist/match-detection/features/bib/host-component.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../../src/match-detection/features/bib/host-component.js"],
-  "sourcesContent": ["\nexport default () => ({\n  name: 'Host/Component record',\n  extract: ({record}) => record.get(/^773$/u).length > 0 ? ['component'] : ['host'],\n  compare: (a, b) => a[0] === b[0] ? 0.0 : -1.0\n});\n"],
-  "mappings": "AACA,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC,EAAC,OAAM,MAAM,OAAO,IAAI,QAAQ,EAAE,SAAS,IAAI,CAAC,WAAW,IAAI,CAAC,MAAM;AAAA,EAChF,SAAS,CAAC,GAAG,MAAM,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,IAAM;AAC3C;",
+  "sourcesContent": ["\nexport default () => ({\n  name: 'Host/Component record',\n  extract: ({record}) => record.get(/^[79]73$/u).length > 0 ? ['component'] : ['host'],\n  compare: (a, b) => a[0] === b[0] ? 0.0 : -1.0\n});\n"],
+  "mappings": "AACA,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC,EAAC,OAAM,MAAM,OAAO,IAAI,WAAW,EAAE,SAAS,IAAI,CAAC,WAAW,IAAI,CAAC,MAAM;AAAA,EACnF,SAAS,CAAC,GAAG,MAAM,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,IAAM;AAC3C;",
   "names": []
 }

package/dist/match-detection/features/bib/isbn.js

@@ -1,21 +1,128 @@
 import createDebugLogger from "debug";
 import { parse as isbnParse } from "isbn3";
-import createInterface from "./standard-identifier-factory.js";
+import { uniqArray } from "./issn.js";
+import { isComponentRecord } from "@natlibfi/melinda-commons";
 const debug = createDebugLogger(`@natlibfi/melinda-record-matching:match-detection:features:standard-identifiers:ISBN`);
 const debugData = debug.extend("data");
-export default () => {
-  const IDENTIFIER_NAME = "ISBN";
-  function validatorAndNormalizer(string) {
-    const isbnParseResult = isbnParse(string);
-    debugData(`isbnParseResult: ${JSON.stringify(isbnParseResult)}`);
-    if (isbnParseResult === null) {
-      debug(`Not parseable ISBN, just removing hyphens`);
-      return { valid: false, value: string.replace(/-/ug, "") };
-    }
-    debug(`Parseable ISBN, normalizing to ISBN-13`);
-    return { valid: true, value: isbnParseResult.isbn13 };
+const MAX_SCORE = 0.75;
+export default () => ({
+  name: "ISBN",
+  extract: ({
+    record
+    /*, recordExternal*/
+  }) => {
+    const host = !isComponentRecord(record, false, []);
+    if (host) {
+      return record.get("020").filter((f) => f.subfields?.some((sf) => ["a", "z"].includes(sf.code) && sf.value));
+    }
+    return record.get("773").filter((f) => f.subfields?.some((sf) => ["z"].includes(sf.code) && sf.value));
+  },
+  // eslint-disable-next-line max-statements
+  compare: (aa, bb) => {
+    debugData(`Comparing ISBN sets ${JSON.stringify(aa)} and ${JSON.stringify(bb)}`);
+    if (aa.length === 0 || bb.length === 0) {
+      return 0;
+    }
+    const [subfieldCodeForGoodValues, subfieldCodeForBadValues] = getSubfieldCodes(aa[0].tag);
+    const [goodValuesA, badValuesA] = getValues(aa);
+    if (goodValuesA.length) {
+      debug(`GOOD VALUES (A): ${goodValuesA.join(", ")}`);
+    }
+    if (badValuesA.length) {
+      debug(`BAD VALUES (A): ${badValuesA.join(", ")}`);
+    }
+    const [goodValuesB, badValuesB] = getValues(bb);
+    if (goodValuesB.length) {
+      debug(`GOOD VALUES (B): ${goodValuesB.join(", ")}`);
+    }
+    if (badValuesB.length) {
+      debug(`BAD VALUES (B): ${badValuesB.join(", ")}`);
+    }
+    const [sharedGoodValues, goodValuesAOnly, goodValuesBOnly] = getUnionData(goodValuesA, goodValuesB);
+    const hitScore = scoreHit();
+    function scoreHit() {
+      if (sharedGoodValues.length > 0) {
+        return MAX_SCORE;
+      }
+      if (goodValuesA.some((valA) => badValuesB.includes(valA)) || goodValuesB.some((valB) => badValuesA.includes(valB))) {
+        return MAX_SCORE;
+      }
+      if (badValuesA.some((valA) => looksGood(valA) && badValuesB.includes(valA)) || badValuesB.some((valB) => looksGood(valB) && badValuesA.includes(valB))) {
+        return MAX_SCORE * 2 / 3;
+      }
+      return 0;
+    }
+    if (sharedGoodValues.length > 0) {
+      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length);
+    }
+    if (hitScore === MAX_SCORE) {
+      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length - 1);
+    }
+    if (hitScore > 0) {
+      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length);
+    }
+    if (goodValuesA.length === 0 || goodValuesB === 0) {
+      return 0;
+    }
+    return -0.75;
+    function getSubfieldCodes(tag) {
+      if (tag === "773") {
+        return ["z", void 0];
+      }
+      return ["a", "z"];
+    }
+    function looksGood(val) {
+      if (/^([0-9]-?){9}[0-9X]$/u.test(val)) {
+        return true;
+      }
+      if (/^([0-9]-?){12}[0-9]$/u.test(val)) {
+        return true;
+      }
+      return false;
+    }
+    function getValues(fields) {
+      const goodValues = fields.flatMap((f) => f.subfields.filter((sf) => sf.code === subfieldCodeForGoodValues)).map((sf) => validatorAndNormalizer(sf.value));
+      const trueGoodValues = goodValues.filter((val) => val.valid).map((val) => val.value);
+      const wannabeGoodValues = goodValues.filter((val) => !val.valid).map((val) => val.value);
+      if (!subfieldCodeForBadValues) {
+        return [trueGoodValues, wannabeGoodValues];
+      }
+      const badValues = fields.flatMap((f) => f.subfields.filter((sf) => sf.code === subfieldCodeForBadValues)).map((sf) => sf.value);
+      return [uniqArray(trueGoodValues), uniqArray([...wannabeGoodValues, ...badValues])];
+    }
+    function validatorAndNormalizer(string) {
+      const string2 = string.replace(/\. -$/u, "");
+      const string3 = string2.replace(/ .*$/u, "");
+      if (string2 !== string3) {
+        const altResult = validatorAndNormalizer(string3);
+        if (altResult.valid) {
+          return altResult;
+        }
+      }
+      const isbnParseResult = isbnParse(string2, "") || "";
+      debugData(`isbnParseResult: ${JSON.stringify(isbnParseResult)}`);
+      if (!isbnParseResult.isValid) {
+        debug(`Not parseable ISBN '${string2}', just removing hyphens`);
+        return { valid: false, value: string2.replace(/-/ug, "") };
+      }
+      debug(`Parseable ISBN '${string2}', normalizing to ISBN-13 '${isbnParseResult.isbn13}'`);
+      return { valid: true, value: isbnParseResult.isbn13 };
+    }
+  }
+});
+function getUnionData(set1, set2) {
+  const shared = set1.filter((val) => set2.includes(val));
+  const onlyInSet1 = set1.filter((val) => !shared.includes(val));
+  const onlyInSet2 = set2.filter((val) => !shared.includes(val));
+  return [shared, onlyInSet1, onlyInSet2];
+}
+function scoreData(score, factor, n) {
+  return innerScoreData(score, n);
+  function innerScoreData(currScore, remaining) {
+    if (remaining > 0) {
+      return innerScoreData(currScore * factor, remaining - 1);
+    }
+    return Math.round(currScore * 100) / 100;
   }
-  const { extract, compare } = createInterface({ identifier: IDENTIFIER_NAME, pattern: /^020$/u, subfieldCodes: ["a", "z"], validIdentifierSubfieldCodes: ["a"], invalidIdentifierSubfieldCodes: ["z"], validatorAndNormalizer });
-  return { extract, compare, name: IDENTIFIER_NAME };
-};
+}
 //# sourceMappingURL=isbn.js.map

package/dist/match-detection/features/bib/isbn.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../../src/match-detection/features/bib/isbn.js"],
-  "sourcesContent": ["\nimport createDebugLogger from 'debug';\nimport {parse as isbnParse} from 'isbn3';\nimport createInterface from './standard-identifier-factory.js';\n\nconst debug = createDebugLogger(`@natlibfi/melinda-record-matching:match-detection:features:standard-identifiers:ISBN`);\nconst debugData = debug.extend('data');\n\nexport default () => {\n  const IDENTIFIER_NAME = 'ISBN';\n\n  function validatorAndNormalizer(string) {\n    const isbnParseResult = isbnParse(string);\n    debugData(`isbnParseResult: ${JSON.stringify(isbnParseResult)}`);\n    if (isbnParseResult === null) {\n      debug(`Not parseable ISBN, just removing hyphens`);\n      return {valid: false, value: string.replace(/-/ug, '')};\n    }\n    debug(`Parseable ISBN, normalizing to ISBN-13`);\n    return {valid: true, value: isbnParseResult.isbn13};\n  }\n\n  const {extract, compare} = createInterface({identifier: IDENTIFIER_NAME, pattern: /^020$/u, subfieldCodes: ['a', 'z'], validIdentifierSubfieldCodes: ['a'], invalidIdentifierSubfieldCodes: ['z'], validatorAndNormalizer});\n  return {extract, compare, name: IDENTIFIER_NAME};\n};\n\n"],
-  "mappings": "AACA,OAAO,uBAAuB;AAC9B,SAAQ,SAAS,iBAAgB;AACjC,OAAO,qBAAqB;AAE5B,MAAM,QAAQ,kBAAkB,sFAAsF;AACtH,MAAM,YAAY,MAAM,OAAO,MAAM;AAErC,eAAe,MAAM;AACnB,QAAM,kBAAkB;AAExB,WAAS,uBAAuB,QAAQ;AACtC,UAAM,kBAAkB,UAAU,MAAM;AACxC,cAAU,oBAAoB,KAAK,UAAU,eAAe,CAAC,EAAE;AAC/D,QAAI,oBAAoB,MAAM;AAC5B,YAAM,2CAA2C;AACjD,aAAO,EAAC,OAAO,OAAO,OAAO,OAAO,QAAQ,OAAO,EAAE,EAAC;AAAA,IACxD;AACA,UAAM,wCAAwC;AAC9C,WAAO,EAAC,OAAO,MAAM,OAAO,gBAAgB,OAAM;AAAA,EACpD;AAEA,QAAM,EAAC,SAAS,QAAO,IAAI,gBAAgB,EAAC,YAAY,iBAAiB,SAAS,UAAU,eAAe,CAAC,KAAK,GAAG,GAAG,8BAA8B,CAAC,GAAG,GAAG,gCAAgC,CAAC,GAAG,GAAG,uBAAsB,CAAC;AAC1N,SAAO,EAAC,SAAS,SAAS,MAAM,gBAAe;AACjD;",
+  "sourcesContent": ["\nimport createDebugLogger from 'debug';\nimport {parse as isbnParse} from 'isbn3';\n\nimport {uniqArray} from './issn.js';\nimport {isComponentRecord} from '@natlibfi/melinda-commons';\n\n\nconst debug = createDebugLogger(`@natlibfi/melinda-record-matching:match-detection:features:standard-identifiers:ISBN`);\nconst debugData = debug.extend('data');\n\nconst MAX_SCORE = 0.75;\n\nexport default () => ({\n  name: 'ISBN',\n  extract: ({record/*, recordExternal*/}) => {\n    //const label = recordExternal && recordExternal.label ? recordExternal.label : 'record';\n\n    const host = !isComponentRecord(record, false, []);\n\n    if (host) {\n      return record.get('020').filter(f => f.subfields?.some(sf => ['a', 'z'].includes(sf.code) && sf.value));\n    }\n    return record.get('773').filter(f => f.subfields?.some(sf => ['z'].includes(sf.code) && sf.value));\n  },\n  // eslint-disable-next-line max-statements\n  compare: (aa, bb) => {\n    debugData(`Comparing ISBN sets ${JSON.stringify(aa)} and ${JSON.stringify(bb)}`);\n    if (aa.length === 0 || bb.length === 0) {\n      // No data for decision\n      return 0;\n    }\n\n    const [subfieldCodeForGoodValues, subfieldCodeForBadValues] = getSubfieldCodes(aa[0].tag);\n\n    const [goodValuesA, badValuesA] = getValues(aa);\n    if (goodValuesA.length) {\n      debug(`GOOD VALUES (A): ${goodValuesA.join(', ')}`);\n    }\n    if (badValuesA.length) {\n      debug(`BAD VALUES (A): ${badValuesA.join(', ')}`);\n    }\n\n    const [goodValuesB, badValuesB] = getValues(bb);\n    if (goodValuesB.length) {\n      debug(`GOOD VALUES (B): ${goodValuesB.join(', ')}`);\n    }\n    if (badValuesB.length) {\n      debug(`BAD VALUES (B): ${badValuesB.join(', ')}`);\n    }\n\n    const [sharedGoodValues, goodValuesAOnly, goodValuesBOnly] = getUnionData(goodValuesA, goodValuesB);\n\n    //debug(`GOOD\\tBOTH: ${sharedGoodValues.length}, A only: ${goodValuesAOnly.length}, B only: ${goodValuesBOnly.length}`);\n\n    const hitScore = scoreHit();\n\n    function scoreHit() {\n      if (sharedGoodValues.length > 0) {\n        return MAX_SCORE;\n      }\n      // One record consider ISBN good and the other record considered it's canceled:\n      if (goodValuesA.some(valA => badValuesB.includes(valA)) || goodValuesB.some(valB => badValuesA.includes(valB))) {\n        return MAX_SCORE;\n      }\n\n      // Value is bad, but looks isbn-ish to a human eye (not validating the isbn again, note than invalid isbns in 020$a are considered bad):\n      // Could happen for two canceled ISBNs for example. I'll give this two thirds of the full score\n      if (badValuesA.some(valA => looksGood(valA) && badValuesB.includes(valA)) || badValuesB.some(valB => looksGood(valB) && badValuesA.includes(valB))) {\n        return MAX_SCORE * 2 / 3;\n      }\n\n      return 0;\n    }\n\n    if (sharedGoodValues.length > 0) {\n      // Third argument (aka 'N times') is >= 0\n      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length);\n    }\n\n    if (hitScore === MAX_SCORE) {\n      // -1 is needed to make the third argument >= 0 (otherwise min val would be 0)\n      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length - 1);\n    }\n\n    if (hitScore > 0) { // Canceled/invalid ISBNs match\n      // Note that this is not (currently) penalized for non-matching canceled/invalid ISBNs. Maybe it should be.\n      return scoreData(hitScore, 0.8, goodValuesAOnly.length + goodValuesBOnly.length);\n    }\n\n    // No match:\n\n    if (goodValuesA.length === 0 || goodValuesB === 0) { // At least one record did not have any good ISBNs, so not penalizing here! (Invalid 020$as are counted a bad.)\n      return 0.0;\n    }\n\n    return -0.75; // Has good ISBNs on both records, but they did not match\n\n\n    function getSubfieldCodes(tag) {\n      if (tag === '773') {\n        return ['z', undefined];\n      }\n      return ['a', 'z'];\n    }\n\n    function looksGood(val) {\n      // isbn10 can end in X:\n      if (/^([0-9]-?){9}[0-9X]$/u.test(val)) {\n        return true;\n      }\n      // isbn13 can not:\n      if (/^([0-9]-?){12}[0-9]$/u.test(val)) {\n        return true;\n      }\n      return false;\n    }\n\n    function getValues(fields) {\n      const goodValues = fields.flatMap(f => f.subfields.filter(sf => sf.code === subfieldCodeForGoodValues)).map(sf => validatorAndNormalizer(sf.value));\n      const trueGoodValues = goodValues.filter(val => val.valid).map(val => val.value);\n      const wannabeGoodValues = goodValues.filter(val => !val.valid).map(val => val.value);\n      if (!subfieldCodeForBadValues) { // 773\n        return [trueGoodValues, wannabeGoodValues];\n      }\n      const badValues = fields.flatMap(f => f.subfields.filter(sf => sf.code === subfieldCodeForBadValues)).map(sf => sf.value);\n      return [uniqArray(trueGoodValues), uniqArray([...wannabeGoodValues, ...badValues])];\n    }\n\n    function validatorAndNormalizer(string) {\n      const string2 = string.replace(/\\. -$/u, ''); // Remove punctuation (773$z)\n\n      // Hack: Historically we 020$a \"1234567890 sidottu\" etc. Try the LHS alone:\n      const string3 = string2.replace(/ .*$/u, '');\n      if (string2 !== string3) {\n        const altResult = validatorAndNormalizer(string3);\n        if (altResult.valid) {\n          return altResult;\n        }\n      }\n\n      const isbnParseResult = isbnParse(string2, '') || '';\n      debugData(`isbnParseResult: ${JSON.stringify(isbnParseResult)}`);\n      if (!isbnParseResult.isValid) {\n        debug(`Not parseable ISBN '${string2}', just removing hyphens`);\n        return {valid: false, value: string2.replace(/-/ug, '')};\n      }\n\n      debug(`Parseable ISBN '${string2}', normalizing to ISBN-13 '${isbnParseResult.isbn13}'`);\n      return {valid: true, value: isbnParseResult.isbn13};\n\n\n    }\n  }\n});\n\n// These are outside the default function as I'll probably want to export these later on\n\nfunction getUnionData(set1, set2) {\n  const shared = set1.filter(val => set2.includes(val));\n  const onlyInSet1 = set1.filter(val => !shared.includes(val));\n  const onlyInSet2 = set2.filter(val => !shared.includes(val));\n  return [shared, onlyInSet1, onlyInSet2];\n}\n\nfunction scoreData(score, factor, n) {\n  return innerScoreData(score, n);\n  function innerScoreData(currScore, remaining) {\n    if (remaining > 0) {\n      return innerScoreData(currScore * factor, remaining-1);\n    }\n    return Math.round(currScore * 100)/100; // 0.600000000001 => 0.6\n  }\n}"],
+  "mappings": "AACA,OAAO,uBAAuB;AAC9B,SAAQ,SAAS,iBAAgB;AAEjC,SAAQ,iBAAgB;AACxB,SAAQ,yBAAwB;AAGhC,MAAM,QAAQ,kBAAkB,sFAAsF;AACtH,MAAM,YAAY,MAAM,OAAO,MAAM;AAErC,MAAM,YAAY;AAElB,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC;AAAA,IAAC;AAAA;AAAA,EAA0B,MAAM;AAGzC,UAAM,OAAO,CAAC,kBAAkB,QAAQ,OAAO,CAAC,CAAC;AAEjD,QAAI,MAAM;AACR,aAAO,OAAO,IAAI,KAAK,EAAE,OAAO,OAAK,EAAE,WAAW,KAAK,QAAM,CAAC,KAAK,GAAG,EAAE,SAAS,GAAG,IAAI,KAAK,GAAG,KAAK,CAAC;AAAA,IACxG;AACA,WAAO,OAAO,IAAI,KAAK,EAAE,OAAO,OAAK,EAAE,WAAW,KAAK,QAAM,CAAC,GAAG,EAAE,SAAS,GAAG,IAAI,KAAK,GAAG,KAAK,CAAC;AAAA,EACnG;AAAA;AAAA,EAEA,SAAS,CAAC,IAAI,OAAO;AACnB,cAAU,uBAAuB,KAAK,UAAU,EAAE,CAAC,QAAQ,KAAK,UAAU,EAAE,CAAC,EAAE;AAC/E,QAAI,GAAG,WAAW,KAAK,GAAG,WAAW,GAAG;AAEtC,aAAO;AAAA,IACT;AAEA,UAAM,CAAC,2BAA2B,wBAAwB,IAAI,iBAAiB,GAAG,CAAC,EAAE,GAAG;AAExF,UAAM,CAAC,aAAa,UAAU,IAAI,UAAU,EAAE;AAC9C,QAAI,YAAY,QAAQ;AACtB,YAAM,oBAAoB,YAAY,KAAK,IAAI,CAAC,EAAE;AAAA,IACpD;AACA,QAAI,WAAW,QAAQ;AACrB,YAAM,mBAAmB,WAAW,KAAK,IAAI,CAAC,EAAE;AAAA,IAClD;AAEA,UAAM,CAAC,aAAa,UAAU,IAAI,UAAU,EAAE;AAC9C,QAAI,YAAY,QAAQ;AACtB,YAAM,oBAAoB,YAAY,KAAK,IAAI,CAAC,EAAE;AAAA,IACpD;AACA,QAAI,WAAW,QAAQ;AACrB,YAAM,mBAAmB,WAAW,KAAK,IAAI,CAAC,EAAE;AAAA,IAClD;AAEA,UAAM,CAAC,kBAAkB,iBAAiB,eAAe,IAAI,aAAa,aAAa,WAAW;AAIlG,UAAM,WAAW,SAAS;AAE1B,aAAS,WAAW;AAClB,UAAI,iBAAiB,SAAS,GAAG;AAC/B,eAAO;AAAA,MACT;AAEA,UAAI,YAAY,KAAK,UAAQ,WAAW,SAAS,IAAI,CAAC,KAAK,YAAY,KAAK,UAAQ,WAAW,SAAS,IAAI,CAAC,GAAG;AAC9G,eAAO;AAAA,MACT;AAIA,UAAI,WAAW,KAAK,UAAQ,UAAU,IAAI,KAAK,WAAW,SAAS,IAAI,CAAC,KAAK,WAAW,KAAK,UAAQ,UAAU,IAAI,KAAK,WAAW,SAAS,IAAI,CAAC,GAAG;AAClJ,eAAO,YAAY,IAAI;AAAA,MACzB;AAEA,aAAO;AAAA,IACT;AAEA,QAAI,iBAAiB,SAAS,GAAG;AAE/B,aAAO,UAAU,UAAU,KAAK,gBAAgB,SAAS,gBAAgB,MAAM;AAAA,IACjF;AAEA,QAAI,aAAa,WAAW;AAE1B,aAAO,UAAU,UAAU,KAAK,gBAAgB,SAAS,gBAAgB,SAAS,CAAC;AAAA,IACrF;AAEA,QAAI,WAAW,GAAG;AAEhB,aAAO,UAAU,UAAU,KAAK,gBAAgB,SAAS,gBAAgB,MAAM;AAAA,IACjF;AAIA,QAAI,YAAY,WAAW,KAAK,gBAAgB,GAAG;AACjD,aAAO;AAAA,IACT;AAEA,WAAO;AAGP,aAAS,iBAAiB,KAAK;AAC7B,UAAI,QAAQ,OAAO;AACjB,eAAO,CAAC,KAAK,MAAS;AAAA,MACxB;AACA,aAAO,CAAC,KAAK,GAAG;AAAA,IAClB;AAEA,aAAS,UAAU,KAAK;AAEtB,UAAI,wBAAwB,KAAK,GAAG,GAAG;AACrC,eAAO;AAAA,MACT;AAEA,UAAI,wBAAwB,KAAK,GAAG,GAAG;AACrC,eAAO;AAAA,MACT;AACA,aAAO;AAAA,IACT;AAEA,aAAS,UAAU,QAAQ;AACzB,YAAM,aAAa,OAAO,QAAQ,OAAK,EAAE,UAAU,OAAO,QAAM,GAAG,SAAS,yBAAyB,CAAC,EAAE,IAAI,QAAM,uBAAuB,GAAG,KAAK,CAAC;AAClJ,YAAM,iBAAiB,WAAW,OAAO,SAAO,IAAI,KAAK,EAAE,IAAI,SAAO,IAAI,KAAK;AAC/E,YAAM,oBAAoB,WAAW,OAAO,SAAO,CAAC,IAAI,KAAK,EAAE,IAAI,SAAO,IAAI,KAAK;AACnF,UAAI,CAAC,0BAA0B;AAC7B,eAAO,CAAC,gBAAgB,iBAAiB;AAAA,MAC3C;AACA,YAAM,YAAY,OAAO,QAAQ,OAAK,EAAE,UAAU,OAAO,QAAM,GAAG,SAAS,wBAAwB,CAAC,EAAE,IAAI,QAAM,GAAG,KAAK;AACxH,aAAO,CAAC,UAAU,cAAc,GAAG,UAAU,CAAC,GAAG,mBAAmB,GAAG,SAAS,CAAC,CAAC;AAAA,IACpF;AAEA,aAAS,uBAAuB,QAAQ;AACtC,YAAM,UAAU,OAAO,QAAQ,UAAU,EAAE;AAG3C,YAAM,UAAU,QAAQ,QAAQ,SAAS,EAAE;AAC3C,UAAI,YAAY,SAAS;AACvB,cAAM,YAAY,uBAAuB,OAAO;AAChD,YAAI,UAAU,OAAO;AACnB,iBAAO;AAAA,QACT;AAAA,MACF;AAEA,YAAM,kBAAkB,UAAU,SAAS,EAAE,KAAK;AAClD,gBAAU,oBAAoB,KAAK,UAAU,eAAe,CAAC,EAAE;AAC/D,UAAI,CAAC,gBAAgB,SAAS;AAC5B,cAAM,uBAAuB,OAAO,0BAA0B;AAC9D,eAAO,EAAC,OAAO,OAAO,OAAO,QAAQ,QAAQ,OAAO,EAAE,EAAC;AAAA,MACzD;AAEA,YAAM,mBAAmB,OAAO,8BAA8B,gBAAgB,MAAM,GAAG;AACvF,aAAO,EAAC,OAAO,MAAM,OAAO,gBAAgB,OAAM;AAAA,IAGpD;AAAA,EACF;AACF;AAIA,SAAS,aAAa,MAAM,MAAM;AAChC,QAAM,SAAS,KAAK,OAAO,SAAO,KAAK,SAAS,GAAG,CAAC;AACpD,QAAM,aAAa,KAAK,OAAO,SAAO,CAAC,OAAO,SAAS,GAAG,CAAC;AAC3D,QAAM,aAAa,KAAK,OAAO,SAAO,CAAC,OAAO,SAAS,GAAG,CAAC;AAC3D,SAAO,CAAC,QAAQ,YAAY,UAAU;AACxC;AAEA,SAAS,UAAU,OAAO,QAAQ,GAAG;AACnC,SAAO,eAAe,OAAO,CAAC;AAC9B,WAAS,eAAe,WAAW,WAAW;AAC5C,QAAI,YAAY,GAAG;AACjB,aAAO,eAAe,YAAY,QAAQ,YAAU,CAAC;AAAA,IACvD;AACA,WAAO,KAAK,MAAM,YAAY,GAAG,IAAE;AAAA,EACrC;AACF;",
   "names": []
 }

package/dist/match-detection/features/bib/issn.js

@@ -1,6 +1,63 @@
-import createInterface from "./standard-identifier-factory.js";
-export default () => {
-  const { extract, compare } = createInterface({ pattern: /^022$/u, subfieldCodes: ["a", "z", "y"] });
-  return { extract, compare, name: "ISSN" };
-};
+import createDebugLogger from "debug";
+import { isComponentRecord } from "@natlibfi/melinda-commons";
+const debug = createDebugLogger("@natlibfi/melinda-record-matching:match-detection:features:issn");
+const debugData = debug.extend("data");
+export default () => ({
+  name: "ISSN",
+  extract: ({
+    record
+    /*, recordExternal*/
+  }) => {
+    const isHost = !isComponentRecord(record, false, []);
+    return getIssns();
+    function getIssns() {
+      const fields = getRelevantFields();
+      if (fields.length === 0) {
+        return [];
+      }
+      debug(`	 ${fields.length} potential ISSN fields (${fields[0].tag})`);
+      const subfieldCodes = getRelevantSubfieldCodes();
+      const subfieldValues = fields.flatMap((f) => f.subfields.filter((sf) => subfieldCodes.includes(sf.code)).map((sf) => sf.value));
+      const validSubfieldValues = subfieldValues?.map((val) => normalizeSubfieldValue(val)).filter((val) => isValidIssn(val));
+      if (!validSubfieldValues) {
+        return [];
+      }
+      return uniqArray(validSubfieldValues);
+    }
+    function getRelevantFields() {
+      if (isHost) {
+        return record.get(/^022$/u);
+      }
+      return record.get(/^[79]73$/u);
+    }
+    function normalizeSubfieldValue(val) {
+      return val.replace(/[., :;].*$/u, "");
+    }
+    function isValidIssn(val) {
+      return /^[0-9][0-9][0-9][0-9]-[0-9][0-9][0-9][0-9X]$/u.test(val);
+    }
+    function getRelevantSubfieldCodes() {
+      if (isHost) {
+        return ["a", "y", "z"];
+      }
+      return ["x"];
+    }
+  },
+  // eslint-disable-next-line max-statements
+  compare: (aa, bb) => {
+    debugData(`Comparing ISSN sets ${JSON.stringify(aa)} and ${JSON.stringify(bb)}`);
+    if (aa.length === 0 || bb.length === 0) {
+      return 0;
+    }
+    const firstSharedIssn = aa.find((val) => bb.includes(val));
+    if (firstSharedIssn) {
+      debug(`	 Shared ISSN found: '${firstSharedIssn}'`);
+      return 0.2;
+    }
+    return -0.2;
+  }
+});
+export function uniqArray(arr) {
+  return arr.filter((val, i) => arr.indexOf(val) === i);
+}
 //# sourceMappingURL=issn.js.map

package/dist/match-detection/features/bib/issn.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../../src/match-detection/features/bib/issn.js"],
-  "sourcesContent": ["\nimport createInterface from './standard-identifier-factory.js';\n\nexport default () => {\n  const {extract, compare} = createInterface({pattern: /^022$/u, subfieldCodes: ['a', 'z', 'y']});\n  return {extract, compare, name: 'ISSN'};\n};\n"],
-  "mappings": "AACA,OAAO,qBAAqB;AAE5B,eAAe,MAAM;AACnB,QAAM,EAAC,SAAS,QAAO,IAAI,gBAAgB,EAAC,SAAS,UAAU,eAAe,CAAC,KAAK,KAAK,GAAG,EAAC,CAAC;AAC9F,SAAO,EAAC,SAAS,SAAS,MAAM,OAAM;AACxC;",
+  "sourcesContent": ["/*\n//import createInterface from './standard-identifier-factory.js';\n\n\nexport default () => {\n  const {extract, compare} = createInterface({pattern: /^022$/u, subfieldCodes: ['a', 'z', 'y']});\n  return {extract, compare, name: 'ISSN'};\n};\n*/\n\n\nimport createDebugLogger from 'debug';\n\nimport {isComponentRecord} from '@natlibfi/melinda-commons';\n\nconst debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features:issn');\nconst debugData = debug.extend('data');\n\nexport default () => ({\n  name: 'ISSN',\n  extract: ({record/*, recordExternal*/}) => {\n    //const label = recordExternal && recordExternal.label ? recordExternal.label : 'record';\n\n    const isHost = !isComponentRecord(record, false, []);\n\n    // debug(`\\t Is HOST: '${isHost}'`);\n    return getIssns();\n\n    function getIssns() {\n      const fields = getRelevantFields();\n      if (fields.length === 0) {\n        return [];\n      }\n      debug(`\\t ${fields.length} potential ISSN fields (${fields[0].tag})`);\n      const subfieldCodes = getRelevantSubfieldCodes();\n      //debug(`\\t subfield codes: '${subfieldCodes.join(\"', '\")}'`);\n      const subfieldValues = fields.flatMap(f => f.subfields.filter(sf => subfieldCodes.includes(sf.code)).map(sf => sf.value));\n      //debug(`\\t cand values: '${subfieldValues.join(\"', '\")}'`);\n      // Stripping punctuaction with substring here is pretty quick and dirty approach...\n      const validSubfieldValues = subfieldValues?.map(val => normalizeSubfieldValue(val)).filter(val => isValidIssn(val));\n      if (!validSubfieldValues) {\n        return [];\n      }\n      return uniqArray(validSubfieldValues);\n    }\n\n    function getRelevantFields() {\n      if (isHost) {\n        return record.get(/^022$/u);\n      }\n      return record.get(/^[79]73$/u);\n    }\n\n    function normalizeSubfieldValue(val) {\n      return val.replace(/[., :;].*$/u, '');\n    }\n\n    function isValidIssn(val) {\n      return /^[0-9][0-9][0-9][0-9]-[0-9][0-9][0-9][0-9X]$/u.test(val);\n    }\n\n    function getRelevantSubfieldCodes() {\n      if (isHost) {\n        return ['a', 'y', 'z'];\n      }\n      return ['x'];\n    }\n  },\n  // eslint-disable-next-line max-statements\n  compare: (aa, bb) => {\n    debugData(`Comparing ISSN sets ${JSON.stringify(aa)} and ${JSON.stringify(bb)}`);\n    if (aa.length === 0 || bb.length === 0) {\n      // No data for decision\n      return 0;\n    }\n    const firstSharedIssn = aa.find(val => bb.includes(val));\n    if (firstSharedIssn) {\n      debug(`\\t Shared ISSN found: '${firstSharedIssn}'`);\n      // Maybe less for comps?\n      return 0.2;\n    }\n    return -0.2;\n\n  }\n});\n\nexport function uniqArray(arr) {\n  return arr.filter((val, i) => arr.indexOf(val) === i);\n}"],
+  "mappings": "AAWA,OAAO,uBAAuB;AAE9B,SAAQ,yBAAwB;AAEhC,MAAM,QAAQ,kBAAkB,iEAAiE;AACjG,MAAM,YAAY,MAAM,OAAO,MAAM;AAErC,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC;AAAA,IAAC;AAAA;AAAA,EAA0B,MAAM;AAGzC,UAAM,SAAS,CAAC,kBAAkB,QAAQ,OAAO,CAAC,CAAC;AAGnD,WAAO,SAAS;AAEhB,aAAS,WAAW;AAClB,YAAM,SAAS,kBAAkB;AACjC,UAAI,OAAO,WAAW,GAAG;AACvB,eAAO,CAAC;AAAA,MACV;AACA,YAAM,KAAM,OAAO,MAAM,2BAA2B,OAAO,CAAC,EAAE,GAAG,GAAG;AACpE,YAAM,gBAAgB,yBAAyB;AAE/C,YAAM,iBAAiB,OAAO,QAAQ,OAAK,EAAE,UAAU,OAAO,QAAM,cAAc,SAAS,GAAG,IAAI,CAAC,EAAE,IAAI,QAAM,GAAG,KAAK,CAAC;AAGxH,YAAM,sBAAsB,gBAAgB,IAAI,SAAO,uBAAuB,GAAG,CAAC,EAAE,OAAO,SAAO,YAAY,GAAG,CAAC;AAClH,UAAI,CAAC,qBAAqB;AACxB,eAAO,CAAC;AAAA,MACV;AACA,aAAO,UAAU,mBAAmB;AAAA,IACtC;AAEA,aAAS,oBAAoB;AAC3B,UAAI,QAAQ;AACV,eAAO,OAAO,IAAI,QAAQ;AAAA,MAC5B;AACA,aAAO,OAAO,IAAI,WAAW;AAAA,IAC/B;AAEA,aAAS,uBAAuB,KAAK;AACnC,aAAO,IAAI,QAAQ,eAAe,EAAE;AAAA,IACtC;AAEA,aAAS,YAAY,KAAK;AACxB,aAAO,gDAAgD,KAAK,GAAG;AAAA,IACjE;AAEA,aAAS,2BAA2B;AAClC,UAAI,QAAQ;AACV,eAAO,CAAC,KAAK,KAAK,GAAG;AAAA,MACvB;AACA,aAAO,CAAC,GAAG;AAAA,IACb;AAAA,EACF;AAAA;AAAA,EAEA,SAAS,CAAC,IAAI,OAAO;AACnB,cAAU,uBAAuB,KAAK,UAAU,EAAE,CAAC,QAAQ,KAAK,UAAU,EAAE,CAAC,EAAE;AAC/E,QAAI,GAAG,WAAW,KAAK,GAAG,WAAW,GAAG;AAEtC,aAAO;AAAA,IACT;AACA,UAAM,kBAAkB,GAAG,KAAK,SAAO,GAAG,SAAS,GAAG,CAAC;AACvD,QAAI,iBAAiB;AACnB,YAAM,yBAA0B,eAAe,GAAG;AAElD,aAAO;AAAA,IACT;AACA,WAAO;AAAA,EAET;AACF;AAEO,gBAAS,UAAU,KAAK;AAC7B,SAAO,IAAI,OAAO,CAAC,KAAK,MAAM,IAAI,QAAQ,GAAG,MAAM,CAAC;AACtD;",
   "names": []
 }

package/dist/match-detection/features/bib/language.js

@@ -1,4 +1,6 @@
 import createDebugLogger from "debug";
+import { MarcRecord } from "@natlibfi/marc-record";
+import { FixSami041, Remove041zxx } from "@natlibfi/marc-record-validators-melinda";
 import { getMatchCounts } from "../../../matching-utils.js";
 const debug = createDebugLogger("@natlibfi/melinda-record-matching:match-detection:features:language");
 const debugData = debug.extend("data");
@@ -6,75 +8,190 @@ export default () => ({
   name: "Language",
   extract: ({ record, recordExternal }) => {
     const label = recordExternal && recordExternal.label ? recordExternal.label : "record";
-    const value008 = get008Value();
-    const values041 = get041Values();
-    debugData(`${label}: 008: ${JSON.stringify(value008)}, 041: ${JSON.stringify(values041)}`);
-    if (!value008 && values041.length < 1) {
-      debugData(`${label}: No actual values found`);
-      return [];
+    const clonedRecord = new MarcRecord(record, record._validationOptions);
+    FixSami041().fix(clonedRecord);
+    Remove041zxx().fix(clonedRecord);
+    return [{ leader: clonedRecord.leader, fields: clonedRecord.fields }, label];
+  },
+  // eslint-disable-next-line max-statements
+  compare: (aa, bb) => {
+    const [a, aLabel] = aa;
+    const [b, bLabel] = bb;
+    debugData(`Comparing language ${JSON.stringify(a)} and ${JSON.stringify(b)}`);
+    const score008 = compare008();
+    const score041 = compare041();
+    return applyLimits(score008 + score041);
+    function applyLimits(score) {
+      if (score > 0.1) {
+        return 0.1;
+      }
+      if (score < -1) {
+        return -1;
+      }
+      return score;
     }
-    const allValues = value008 === void 0 ? values041 : values041.concat(value008);
-    const uniqueSortedValues = [...new Set(allValues)].sort();
-    return uniqueSortedValues;
-    function get008Value() {
-      const value = record.get(/^008$/u)?.[0]?.value || void 0;
-      debugData(`${label}: 008 value: ${value}`);
-      if (!value) {
-        return void 0;
-      }
-      const code = value.slice(35, 38);
-      debugData(`${label}: 008 code: ${code}`);
-      return isLangCodeForALanguage(code) ? code : void 0;
+    function compare008() {
+      const a008 = get008Value(a, aLabel);
+      const b008 = get008Value(b, bLabel);
+      if (a008 === void 0 || b008 === void 0) {
+        return 0;
+      }
+      if (containsNoData(a008) || containsNoData(b008)) {
+        return 0;
+      }
+      if (a008 === b008) {
+        return 0.05;
+      }
+      if (a008 === "mul" || b008 === "mul") {
+        return 0;
+      }
+      return -0.2;
     }
-    function get041Values() {
-      return record.get(/^041$/u).filter(({ ind2 }) => ind2 === " ").map(({ subfields }) => subfields).flat().filter(({ code }) => code === "a" || code === "d").filter(({ value }) => value && isLangCodeForALanguage(value)).map(({ value }) => value);
+    function compare041() {
+      const a041s = getFields041(a);
+      const b041s = getFields041(b);
+      if (a041s.length === 0 || b041s.length === 0) {
+        return 0;
+      }
+      if (a041s.length === 1 && b041s.length === 1) {
+        const sourcePenalty = getSourceOfLanguageCode(a041s[0]) === getSourceOfLanguageCode(b041s[0]) ? 0 : -0.05;
+        const mainPenalty = compareLanguageCodeLists(getFieldsLanguageCodes(a041s[0]), getFieldsLanguageCodes(b041s[0]));
+        const scoreInd1 = mainPenalty === 0 ? 0 : indicator1Penalty(a041s[0], b041s[0]);
+        return mainPenalty + sourcePenalty + scoreInd1;
+      }
+      return compareLanguageCodeLists(getLanguageCodesFrom041Fields(a), getLanguageCodesFrom041Fields(b));
+      function getFields041(record) {
+        if (!record.fields) {
+          return [];
+        }
+        return record.fields.filter((f) => f.tag === "041");
+      }
     }
-    function isLangCodeForALanguage(code) {
-      if (code.length !== 3) {
-        debugData(`Code ${code} is not correct length (3) for a language code.`);
-        return false;
+    function getSourceOfLanguageCode(field) {
+      const sf2 = field.subfields.find((sf) => sf.code === "2");
+      if (!sf2) {
+        return "MARC";
       }
-      if (code === "|||" || code === "   " || code === "^^^" || code === "mul" || code === "zxx") {
-        debugData(`Code ${code} is not code for a spesific language.`);
-        return false;
+      if (sf2.value.match(/639.2/u)) {
+        return "ISO 639-2";
       }
-      const langCodePattern = /^[a-z][a-z][a-z]$/ui;
-      if (!langCodePattern.test(code)) {
-        debugData(`Code ${code} is not valid as a language code`);
-        return false;
+      if (sf2.value.match(/639.3/u)) {
+        return "ISO 639-3";
       }
-      return true;
+      return sf2.value;
     }
-  },
-  // eslint-disable-next-line max-statements
-  compare: (a, b) => {
-    debugData(`Comparing ${JSON.stringify(a)} and ${JSON.stringify(b)}`);
-    if (a.length === 0 || b.length === 0) {
-      debugData(`No language to compare`);
-      return 0;
+    function indicator1Penalty(aField, bField) {
+      if (aField.ind1 === " " || bField.ind1 === " " || aField.ind1 === bField.ind1) {
+        return 0;
+      }
+      debug(`	 Indicator penalty: '${aField.ind1}' vs '${bField.ind1}'`);
+      return -0.1;
     }
-    if (a.length === b.length && a.every((element, index) => element === b[index])) {
-      debugData(`All languages match`);
-      return 0.1;
+    function containsNoData(languageCode) {
+      return ["   ", "|||", "und"].includes(languageCode);
     }
-    const { matchingValues, possibleMatchValues, maxValues } = getMatchCounts(a, b);
-    if (matchingValues < 1) {
-      debug(`Both have languages, but none of these match.`);
-      return -1;
+    function compareLanguageCodeLists(a2, b2) {
+      if (a2.length === 0 || b2.length === 0) {
+        debugData(`No language to compare`);
+        return 0;
+      }
+      const samiLanguageCodes = ["sma", "sme", "smj", "smn", "sms"];
+      if (a2.includes("smi") && b2.includes("smi")) {
+        if (a2.some((code) => samiLanguageCodes.includes(code)) && !b2.some((code) => samiLanguageCodes.includes(code))) {
+          return compareLanguageCodeLists(a2.filter((val) => !samiLanguageCodes.includes(val)), b2);
+        }
+        if (b2.some((code) => samiLanguageCodes.includes(code)) && !a2.some((code) => samiLanguageCodes.includes(code))) {
+          return compareLanguageCodeLists(a2, b2.filter((val) => !samiLanguageCodes.includes(val)));
+        }
+      }
+      if (a2.length === b2.length && a2.every((element, index) => element === b2[index])) {
+        debugData(`All languages match`);
+        return 0.1;
+      }
+      if (a2.length === 1 && a2[0] === "mul" && b2.length > 1) {
+        return 0;
+      }
+      if (b2.length === 1 && b2[0] === "mul" && a2.length > 1) {
+        return 0;
+      }
+      const sharedValues = getSharedValues(a2, b2);
+      const aOnly = a2.filter((val) => !sharedValues.includes(val));
+      const bOnly = b2.filter((val) => !sharedValues.includes(val));
+      const hasUnd = [...aOnly, ...bOnly].includes("und");
+      if (sharedValues.length < 1) {
+        console.info(`NV: ${sharedValues.join(", ")}`);
+        if (aOnly.length === 1 && bOnly.length === 1 && hasUnd) {
+          debug(`Both have languages, but none of these match. However, the benefit of doubt is given: '${aOnly[0]}' and '${bOnly[0]}' might mean the same}`);
+          return 0;
+        }
+        debug(`Both have languages, but none of these match.`);
+        return -1;
+      }
+      const { matchingValues, possibleMatchValues, maxValues } = getMatchCounts(a2, b2);
+      debug(`Both have languages, ${matchingValues}/${possibleMatchValues} valid languages match.`);
+      debug(`Possible matches: ${possibleMatchValues}/${maxValues}`);
+      const missingCount = maxValues - possibleMatchValues;
+      const misMatchCount = possibleMatchValues - matchingValues;
+      debug(`	 missing: ${missingCount}`);
+      debug(`	 mismatches: ${misMatchCount}`);
+      const penaltyForMissing = 0.02 * (maxValues - possibleMatchValues);
+      const penaltyForMisMatch = 0.05 * (possibleMatchValues - matchingValues);
+      debug(`	 points: penaltyForMissing: ${penaltyForMissing}`);
+      debug(`	 points: penaltyForMisMatch: ${penaltyForMisMatch}`);
+      const points = Number(Number(0.1 - penaltyForMisMatch - penaltyForMissing).toFixed(2));
+      debug(`Total points: ${points}`);
+      return points;
     }
-    debug(`Both have languages, ${matchingValues}/${possibleMatchValues} valid languages match.`);
-    debug(`Possible matches: ${possibleMatchValues}/${maxValues}`);
-    const missingCount = maxValues - possibleMatchValues;
-    const misMatchCount = possibleMatchValues - matchingValues;
-    debug(`	 missing: ${missingCount}`);
-    debug(`	 mismatches: ${misMatchCount}`);
-    const penaltyForMissing = 0.02 * (maxValues - possibleMatchValues);
-    const penaltyForMisMatch = 0.05 * (possibleMatchValues - matchingValues);
-    debug(`	 points: penaltyForMissing: ${penaltyForMissing}`);
-    debug(`	 points: penaltyForMisMatch: ${penaltyForMisMatch}`);
-    const points = Number(Number(0.1 - penaltyForMisMatch - penaltyForMissing).toFixed(2));
-    debug(`Total points: ${points}`);
-    return points;
   }
 });
+function get008Value(record, label = "record") {
+  if (!record || !record.fields) {
+    return;
+  }
+  const f008 = record.fields.find((f) => f.tag === "008");
+  if (!f008) {
+    return void 0;
+  }
+  const value = f008.value || defaultValue;
+  if (!value) {
+    debugData(`${label}: Failed to extact 008/35-37 value from '${value}'`);
+    return defaultValue;
+  }
+  const code = value.slice(35, 38);
+  debugData(`${label}: 008 code: '${code}'`);
+  return code;
+}
+function isLangCodeForALanguage(code, label = "record", encoding = void 0) {
+  if (!code) {
+    return false;
+  }
+  if ([void 0, "ISO 639-2", "ISO-639-3"].includes(encoding) && code.length !== 3) {
+    debugData(`${label}: Code ${code} is not correct length (3) for a language code.`);
+    return false;
+  }
+  if (!encoding) {
+    if (code === "|||" || code === "   ") {
+      debugData(`${label}: Code ${code} is not code for a spesific language.`);
+      return false;
+    }
+  }
+  const langCodePattern = /^[a-z][a-z][a-z]$/ui;
+  if (!langCodePattern.test(code)) {
+    debugData(`${label}: Code ${code} is not valid as a language code`);
+    return false;
+  }
+  return true;
+}
+function getLanguageCodesFrom041Fields(record) {
+  const values = record.fields.filter((f) => f.tag === "041").flatMap((f) => getFieldsLanguageCodes(f));
+  return [...new Set(values)].sort();
+}
+function getFieldsLanguageCodes(field) {
+  const relevantSubfields = field.subfields.filter((sf) => sf.code === "a" || sf.code === "d").filter((sf) => isLangCodeForALanguage(sf.value, "field"));
+  const values = relevantSubfields.map((sf) => sf.value).flat();
+  return [...new Set(values)].sort();
+}
+function getSharedValues(list1, list2) {
+  return list1.filter((val) => list2.includes(val));
+}
 //# sourceMappingURL=language.js.map

package/dist/match-detection/features/bib/language.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../../src/match-detection/features/bib/language.js"],
-  "sourcesContent": ["\nimport createDebugLogger from 'debug';\nimport {getMatchCounts} from '../../../matching-utils.js';\n\nconst debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features:language');\nconst debugData = debug.extend('data');\n\nexport default () => ({\n  name: 'Language',\n  extract: ({record, recordExternal}) => {\n    const label = recordExternal && recordExternal.label ? recordExternal.label : 'record';\n\n    const value008 = get008Value();\n    const values041 = get041Values();\n    debugData(`${label}: 008: ${JSON.stringify(value008)}, 041: ${JSON.stringify(values041)}`);\n\n    if (!value008 && values041.length < 1) {\n      debugData(`${label}: No actual values found`);\n      return [];\n    }\n\n    const allValues = value008 === undefined ? values041 : values041.concat(value008);\n    const uniqueSortedValues = [...new Set(allValues)].sort();\n\n    return uniqueSortedValues;\n\n    function get008Value() {\n      const value = record.get(/^008$/u)?.[0]?.value || undefined;\n      debugData(`${label}: 008 value: ${value}`);\n\n      if (!value) {\n        return undefined;\n      }\n\n      const code = value.slice(35, 38);\n      debugData(`${label}: 008 code: ${code}`);\n      return isLangCodeForALanguage(code) ? code : undefined;\n    }\n\n    // Uses only f041s that have 2nd ind ' ', which means that the codes used are MARC 21 language codes\n\n    function get041Values() {\n      return record.get(/^041$/u)\n        .filter(({ind2}) => ind2 === ' ')\n        .map(({subfields}) => subfields)\n        .flat()\n        .filter(({code}) => code === 'a' || code === 'd')\n        .filter(({value}) => value && isLangCodeForALanguage(value))\n        .map(({value}) => value);\n    }\n\n    // Check if a string is a possible, validly formed language code for a single language\n    // Currently accept also codes in capitals\n    function isLangCodeForALanguage(code) {\n      if (code.length !== 3) {\n        debugData(`Code ${code} is not correct length (3) for a language code.`);\n        return false;\n      }\n      if (code === '|||' || code === '   ' || code === '^^^' || code === 'mul' || code === 'zxx') {\n        debugData(`Code ${code} is not code for a spesific language.`);\n        return false;\n      }\n      const langCodePattern = /^[a-z][a-z][a-z]$/ui;\n      if (!langCodePattern.test(code)) {\n        debugData(`Code ${code} is not valid as a language code`);\n        return false;\n      }\n      return true;\n    }\n\n  },\n  // eslint-disable-next-line max-statements\n  compare: (a, b) => {\n    debugData(`Comparing ${JSON.stringify(a)} and ${JSON.stringify(b)}`);\n\n    if (a.length === 0 || b.length === 0) {\n      debugData(`No language to compare`);\n      return 0;\n    }\n\n    if (a.length === b.length && a.every((element, index) => element === b[index])) {\n      debugData(`All languages match`);\n      return 0.1;\n    }\n\n    const {matchingValues, possibleMatchValues, maxValues} = getMatchCounts(a, b);\n\n    if (matchingValues < 1) {\n      debug(`Both have languages, but none of these match.`);\n      return -1.0;\n    }\n    debug(`Both have languages, ${matchingValues}/${possibleMatchValues} valid languages match.`);\n    // ignore non-matches if there is mismatching amount of values\n    debug(`Possible matches: ${possibleMatchValues}/${maxValues}`);\n    //we give some kind of penalty for mismatching amount of values instead of simple divide?\n    const missingCount = maxValues - possibleMatchValues;\n    const misMatchCount = possibleMatchValues - matchingValues;\n    debug(`\\t missing: ${missingCount}`);\n    debug(`\\t mismatches: ${misMatchCount}`);\n\n    const penaltyForMissing = 0.02 * (maxValues - possibleMatchValues);\n    const penaltyForMisMatch = 0.05 * (possibleMatchValues - matchingValues);\n    debug(`\\t points: penaltyForMissing: ${penaltyForMissing}`);\n    debug(`\\t points: penaltyForMisMatch: ${penaltyForMisMatch}`);\n\n    const points = Number(Number(0.1 - penaltyForMisMatch - penaltyForMissing).toFixed(2));\n    debug(`Total points: ${points}`);\n\n    return points;\n  }\n});\n"],
-  "mappings": "AACA,OAAO,uBAAuB;AAC9B,SAAQ,sBAAqB;AAE7B,MAAM,QAAQ,kBAAkB,qEAAqE;AACrG,MAAM,YAAY,MAAM,OAAO,MAAM;AAErC,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC,EAAC,QAAQ,eAAc,MAAM;AACrC,UAAM,QAAQ,kBAAkB,eAAe,QAAQ,eAAe,QAAQ;AAE9E,UAAM,WAAW,YAAY;AAC7B,UAAM,YAAY,aAAa;AAC/B,cAAU,GAAG,KAAK,UAAU,KAAK,UAAU,QAAQ,CAAC,UAAU,KAAK,UAAU,SAAS,CAAC,EAAE;AAEzF,QAAI,CAAC,YAAY,UAAU,SAAS,GAAG;AACrC,gBAAU,GAAG,KAAK,0BAA0B;AAC5C,aAAO,CAAC;AAAA,IACV;AAEA,UAAM,YAAY,aAAa,SAAY,YAAY,UAAU,OAAO,QAAQ;AAChF,UAAM,qBAAqB,CAAC,GAAG,IAAI,IAAI,SAAS,CAAC,EAAE,KAAK;AAExD,WAAO;AAEP,aAAS,cAAc;AACrB,YAAM,QAAQ,OAAO,IAAI,QAAQ,IAAI,CAAC,GAAG,SAAS;AAClD,gBAAU,GAAG,KAAK,gBAAgB,KAAK,EAAE;AAEzC,UAAI,CAAC,OAAO;AACV,eAAO;AAAA,MACT;AAEA,YAAM,OAAO,MAAM,MAAM,IAAI,EAAE;AAC/B,gBAAU,GAAG,KAAK,eAAe,IAAI,EAAE;AACvC,aAAO,uBAAuB,IAAI,IAAI,OAAO;AAAA,IAC/C;AAIA,aAAS,eAAe;AACtB,aAAO,OAAO,IAAI,QAAQ,EACvB,OAAO,CAAC,EAAC,KAAI,MAAM,SAAS,GAAG,EAC/B,IAAI,CAAC,EAAC,UAAS,MAAM,SAAS,EAC9B,KAAK,EACL,OAAO,CAAC,EAAC,KAAI,MAAM,SAAS,OAAO,SAAS,GAAG,EAC/C,OAAO,CAAC,EAAC,MAAK,MAAM,SAAS,uBAAuB,KAAK,CAAC,EAC1D,IAAI,CAAC,EAAC,MAAK,MAAM,KAAK;AAAA,IAC3B;AAIA,aAAS,uBAAuB,MAAM;AACpC,UAAI,KAAK,WAAW,GAAG;AACrB,kBAAU,QAAQ,IAAI,iDAAiD;AACvE,eAAO;AAAA,MACT;AACA,UAAI,SAAS,SAAS,SAAS,SAAS,SAAS,SAAS,SAAS,SAAS,SAAS,OAAO;AAC1F,kBAAU,QAAQ,IAAI,uCAAuC;AAC7D,eAAO;AAAA,MACT;AACA,YAAM,kBAAkB;AACxB,UAAI,CAAC,gBAAgB,KAAK,IAAI,GAAG;AAC/B,kBAAU,QAAQ,IAAI,kCAAkC;AACxD,eAAO;AAAA,MACT;AACA,aAAO;AAAA,IACT;AAAA,EAEF;AAAA;AAAA,EAEA,SAAS,CAAC,GAAG,MAAM;AACjB,cAAU,aAAa,KAAK,UAAU,CAAC,CAAC,QAAQ,KAAK,UAAU,CAAC,CAAC,EAAE;AAEnE,QAAI,EAAE,WAAW,KAAK,EAAE,WAAW,GAAG;AACpC,gBAAU,wBAAwB;AAClC,aAAO;AAAA,IACT;AAEA,QAAI,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,CAAC,SAAS,UAAU,YAAY,EAAE,KAAK,CAAC,GAAG;AAC9E,gBAAU,qBAAqB;AAC/B,aAAO;AAAA,IACT;AAEA,UAAM,EAAC,gBAAgB,qBAAqB,UAAS,IAAI,eAAe,GAAG,CAAC;AAE5E,QAAI,iBAAiB,GAAG;AACtB,YAAM,+CAA+C;AACrD,aAAO;AAAA,IACT;AACA,UAAM,wBAAwB,cAAc,IAAI,mBAAmB,yBAAyB;AAE5F,UAAM,qBAAqB,mBAAmB,IAAI,SAAS,EAAE;AAE7D,UAAM,eAAe,YAAY;AACjC,UAAM,gBAAgB,sBAAsB;AAC5C,UAAM,cAAe,YAAY,EAAE;AACnC,UAAM,iBAAkB,aAAa,EAAE;AAEvC,UAAM,oBAAoB,QAAQ,YAAY;AAC9C,UAAM,qBAAqB,QAAQ,sBAAsB;AACzD,UAAM,gCAAiC,iBAAiB,EAAE;AAC1D,UAAM,iCAAkC,kBAAkB,EAAE;AAE5D,UAAM,SAAS,OAAO,OAAO,MAAM,qBAAqB,iBAAiB,EAAE,QAAQ,CAAC,CAAC;AACrF,UAAM,iBAAiB,MAAM,EAAE;AAE/B,WAAO;AAAA,EACT;AACF;",
-  "names": []
+  "sourcesContent": ["\nimport createDebugLogger from 'debug';\n\nimport {MarcRecord} from '@natlibfi/marc-record';\n\nimport {FixSami041, Remove041zxx} from '@natlibfi/marc-record-validators-melinda';\n\nimport {getMatchCounts} from '../../../matching-utils.js';\n\nconst debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features:language');\nconst debugData = debug.extend('data');\n\n// NB! 2025-12-17: I changed the logic drastically. However, I tried to keep the scores as same as possible.\n// 'extract' no longer extracts anything. Instead it clones the record, and does some preprocessing (using validators) instead.\n// 'compare' had multiple changes\n// - f041 ind1 differences ('0' vs '1') now result in a small penalty\n// - Two sami languages related changes:\n// -- validator adds a 'smi' subfield before a corresponding sma/sme/...subfield, if needed (national)\n// -- 'smi' only vs 'smi'+'sma' does not cause penalty\n// - 'mul' vs 'fin'+'swe' does not cause a penalty. However, 'mul' vs 'fin' alone triggers penalty.\n// - 008/35-37 and f041$a/$d are calculated separately\n// - a threshold is applied: return value is always between -1.0 and +0.1 (as it was before)\n// - we try to handle 041 $2 ISO 639-2 and ISO 639-3 as well.\n// - 'und' vs one other language does not cause a penalty in certain contexts. In other contexts it's treated as a normal \"language code\". (Tune/alleviate penalties later on.)\n\nexport default () => ({\n  name: 'Language',\n  extract: ({record, recordExternal}) => {\n    const label = recordExternal && recordExternal.label ? recordExternal.label : 'record';\n    const clonedRecord = new MarcRecord(record, record._validationOptions); // clone(record); // NB! This loses record.get()...\n\n    FixSami041().fix(clonedRecord); // Handle 'smi' adding if needed\n    Remove041zxx().fix(clonedRecord); // Remove 'zxx' from f041s\n    // Add other language code normalizations?\n\n    // Should we return all the fields, or just the relevant fields?\n    return [{leader: clonedRecord.leader, fields: clonedRecord.fields}, label];\n  },\n  // eslint-disable-next-line max-statements\n  compare: (aa, bb) => {\n    const [a, aLabel] = aa;\n    const [b, bLabel] = bb;\n    debugData(`Comparing language ${JSON.stringify(a)} and ${JSON.stringify(b)}`);\n\n    const score008 = compare008();\n    const score041 = compare041();\n    return applyLimits(score008 + score041);\n\n    function applyLimits(score) {\n      if (score > 0.1) { // Keeps the original max, we might have 0.10 (041) + 0.05 (008/35-37) = 0.15 now\n        return 0.1;\n      }\n      if (score < -1) {\n        return -1.0;\n      }\n      return score;\n    }\n\n    function compare008() {\n      const a008 = get008Value(a, aLabel);\n      const b008 = get008Value(b, bLabel);\n      // Something seriously wrong with 008:\n      if (a008 === undefined || b008 === undefined) {\n        return 0.0; // Punish for generic badness?\n      }\n\n      if (containsNoData(a008) || containsNoData(b008)){\n        // Nothing to compare\n        return 0.0;\n      }\n\n      if (a008 === b008) {\n        return 0.05;\n      }\n\n      if (a008 === 'mul' || b008 === 'mul') {\n        return 0.0;\n      }\n\n      return -0.2;\n    }\n\n\n\n\n\n    function compare041() {\n      const a041s = getFields041(a);\n      const b041s = getFields041(b);\n      if (a041s.length === 0 || b041s.length === 0) {\n        return 0.0; // Should we punish these for badness?\n      }\n      if (a041s.length === 1 && b041s.length === 1) {\n        // The language codes are typically same between different sources. Incur a small penalty for differences though...\n        const sourcePenalty = getSourceOfLanguageCode(a041s[0]) === getSourceOfLanguageCode(b041s[0]) ? 0.0 : -0.05;\n        const mainPenalty = compareLanguageCodeLists(getFieldsLanguageCodes(a041s[0]), getFieldsLanguageCodes(b041s[0]));\n        const scoreInd1 = mainPenalty === 0.0 ? 0.0 : indicator1Penalty(a041s[0], b041s[0]);\n        return mainPenalty + sourcePenalty + scoreInd1;\n      }\n      // Things are already complicated (likely to pe penalized, so no nedd to worrty about language code sources, I daresay.\n      return compareLanguageCodeLists(getLanguageCodesFrom041Fields(a), getLanguageCodesFrom041Fields(b));\n\n      function getFields041(record) {\n        if (!record.fields) {\n          return [];\n        }\n        return record.fields.filter(f => f.tag === '041');\n      }\n\n    }\n\n    function getSourceOfLanguageCode(field) {\n      const sf2 = field.subfields.find(sf => sf.code === '2');\n      if (!sf2) {\n        return 'MARC';\n      }\n      // Normalize the two relevant language code names:\n      if (sf2.value.match(/639.2/u)) {\n        return 'ISO 639-2';\n      }\n      if (sf2.value.match(/639.3/u)) {\n        return 'ISO 639-3';\n      }\n      return sf2.value; // We don't actually have anything else in Melinda though\n    }\n\n    function indicator1Penalty(aField, bField) {\n      if (aField.ind1 === ' ' || bField.ind1 === ' '|| aField.ind1 === bField.ind1 ) {\n        return 0.0;\n      }\n      debug(`\\t Indicator penalty: '${aField.ind1}' vs '${bField.ind1}'`);\n      return -0.1;\n    }\n\n\n\n    function containsNoData(languageCode) {\n      return ['   ', '|||', 'und'].includes(languageCode);\n    }\n\n    function compareLanguageCodeLists(a, b) {\n      if (a.length === 0 || b.length === 0) {\n        debugData(`No language to compare`);\n        return 0;\n      }\n\n      // If either one of the sets has only the generic 'smi', remove the specific sami languages codes from the other as well:\n      const samiLanguageCodes = ['sma', 'sme', 'smj', 'smn', 'sms']; // NB! Don't put generic 'smi' here!\n\n      if (a.includes('smi') && b.includes('smi')) {\n        if (a.some(code => samiLanguageCodes.includes(code)) && !b.some(code => samiLanguageCodes.includes(code))) {\n          return compareLanguageCodeLists(a.filter(val => !samiLanguageCodes.includes(val)), b); // recurse\n        }\n        if (b.some(code => samiLanguageCodes.includes(code)) && !a.some(code => samiLanguageCodes.includes(code))) {\n          return compareLanguageCodeLists(a, b.filter(val => !samiLanguageCodes.includes(val))); // recurse\n        }\n      }\n\n\n\n\n      if (a.length === b.length && a.every((element, index) => element === b[index])) {\n        debugData(`All languages match`);\n        return 0.1;\n      }\n\n      // Handle 'mul'. (Should we check whether the other array contains mul as well, probably not...)\n      // (If 'mul' does not appear alone, then it is very iffy... )\n      if (a.length === 1 && a[0] === 'mul' && b.length > 1) {\n        return 0;\n      }\n      if (b.length === 1 && b[0] === 'mul' && a.length > 1) {\n        return 0;\n      }\n\n\n\n      // Damage control:\n\n      // Not using the generic solution here as eg. 'und' needs a special treatment\n      const sharedValues = getSharedValues(a, b);\n      const aOnly = a.filter(val => !sharedValues.includes(val));\n      const bOnly = b.filter(val => !sharedValues.includes(val));\n      const hasUnd = [...aOnly, ...bOnly].includes('und');\n\n      if (sharedValues.length < 1) {\n        console.info(`NV: ${sharedValues.join(\", \")}`);\n        if (aOnly.length === 1 && bOnly.length === 1 && hasUnd) {\n          debug(`Both have languages, but none of these match. However, the benefit of doubt is given: '${aOnly[0]}' and '${bOnly[0]}' might mean the same}`);\n          return 0;\n        }\n        debug(`Both have languages, but none of these match.`); // includes 'mul' vs a single language code\n        return -1.0;\n      }\n\n      const {matchingValues, possibleMatchValues, maxValues} = getMatchCounts(a, b);\n\n      debug(`Both have languages, ${matchingValues}/${possibleMatchValues} valid languages match.`);\n      // ignore non-matches if there is mismatching amount of values\n      debug(`Possible matches: ${possibleMatchValues}/${maxValues}`);\n      // we give some kind of penalty for mismatching amount of values instead of simple divide?\n      const missingCount = maxValues - possibleMatchValues;\n      const misMatchCount = possibleMatchValues - matchingValues;\n      debug(`\\t missing: ${missingCount}`);\n      debug(`\\t mismatches: ${misMatchCount}`);\n\n      const penaltyForMissing = 0.02 * (maxValues - possibleMatchValues);\n      const penaltyForMisMatch = 0.05 * (possibleMatchValues - matchingValues);\n      debug(`\\t points: penaltyForMissing: ${penaltyForMissing}`);\n      debug(`\\t points: penaltyForMisMatch: ${penaltyForMisMatch}`);\n\n      const points = Number(Number(0.1 - penaltyForMisMatch - penaltyForMissing).toFixed(2));\n      debug(`Total points: ${points}`);\n\n      return points;\n    }\n  }\n\n});\n\nfunction get008Value(record, label = 'record') {\n  if (!record || !record.fields) {\n    return;\n  }\n  const f008 = record.fields.find(f => f.tag === '008');\n  if (!f008) {\n    return undefined;\n  }\n  const value = f008.value || defaultValue;\n\n  if (!value) {\n    debugData(`${label}: Failed to extact 008/35-37 value from '${value}'`);\n    return defaultValue;\n  }\n\n  const code = value.slice(35, 38);\n  debugData(`${label}: 008 code: '${code}'`);\n  return code;\n}\n\n// Check if a string is a possible, validly formed language code for a single language\n// Currently accept also codes in capitals\nfunction isLangCodeForALanguage(code, label = 'record', encoding = undefined) {\n  if (!code) {\n    return false;\n  }\n\n  if ([undefined, 'ISO 639-2', 'ISO-639-3'].includes(encoding) && code.length !== 3) {\n    debugData(`${label}: Code ${code} is not correct length (3) for a language code.`);\n    return false;\n  }\n  if (!encoding) {\n    // 'mul' should be passed as 'mul' can match 'fin' + 'swe' etc.\n    // 'zxx' should be removed by a validator\n    // '^^^' is Aleph-specific corruption, not supporting it anymore\n    if (code === '|||' || code === '   ' ) { // || code === '^^^' || code === 'mul' || code === 'zxx') {\n      debugData(`${label}: Code ${code} is not code for a spesific language.`);\n      return false;\n    }\n  }\n  // Marc, ISO 639-2 and ISO 639-3 are all three-letters long. Not other language codes seen in Melinda:\n  const langCodePattern = /^[a-z][a-z][a-z]$/ui;\n  if (!langCodePattern.test(code)) {\n    debugData(`${label}: Code ${code} is not valid as a language code`);\n    return false;\n  }\n  return true;\n}\n\n\n\nfunction getLanguageCodesFrom041Fields(record) {\n  // NB! We brutally don't check $2 (language code source) as marc's language codes is practically a subset of ISO 639-2,\n  // and also ISO 639-2 and ISO 639-3 overlap to a degree. If we ever run into a trouble with some ISO 639-2 vs 639-3 mismatch, then we'll work it out.\n  // Also we could write a validator that converts ISO 639-2 to marc if applicable and maybe even partial support for ISO-639-3.\n  // Note that ISO 639-2-B values correspond with marc beteer that ISO 639-2-T. Eg. code for Chinese 'zho' should/code be normalized to 'chi'!\n  const values = record.fields.filter(f => f.tag === '041').flatMap(f => getFieldsLanguageCodes(f));\n  /*\n        // .filter(({ind2}) => ind2 === ' ')\n        .map(({subfields}) => subfields)\n        .flat()\n        .filter(({code}) => code === 'a' || code === 'd')\n        .filter(({value}) => value && isLangCodeForALanguage(value))\n        .map(({value}) => value);\n    */\n  return [...new Set(values)].sort();\n}\n\nfunction getFieldsLanguageCodes(field) {\n  const relevantSubfields = field.subfields.filter(sf => sf.code === 'a' || sf.code === 'd').filter(sf => isLangCodeForALanguage(sf.value, 'field'));\n\n  const values = relevantSubfields.map(sf => sf.value).flat();\n\n  return [...new Set(values)].sort();\n}\n\nfunction getSharedValues(list1, list2) {\n  return list1.filter(val => list2.includes(val));\n}"],
+  "mappings": "AACA,OAAO,uBAAuB;AAE9B,SAAQ,kBAAiB;AAEzB,SAAQ,YAAY,oBAAmB;AAEvC,SAAQ,sBAAqB;AAE7B,MAAM,QAAQ,kBAAkB,qEAAqE;AACrG,MAAM,YAAY,MAAM,OAAO,MAAM;AAerC,eAAe,OAAO;AAAA,EACpB,MAAM;AAAA,EACN,SAAS,CAAC,EAAC,QAAQ,eAAc,MAAM;AACrC,UAAM,QAAQ,kBAAkB,eAAe,QAAQ,eAAe,QAAQ;AAC9E,UAAM,eAAe,IAAI,WAAW,QAAQ,OAAO,kBAAkB;AAErE,eAAW,EAAE,IAAI,YAAY;AAC7B,iBAAa,EAAE,IAAI,YAAY;AAI/B,WAAO,CAAC,EAAC,QAAQ,aAAa,QAAQ,QAAQ,aAAa,OAAM,GAAG,KAAK;AAAA,EAC3E;AAAA;AAAA,EAEA,SAAS,CAAC,IAAI,OAAO;AACnB,UAAM,CAAC,GAAG,MAAM,IAAI;AACpB,UAAM,CAAC,GAAG,MAAM,IAAI;AACpB,cAAU,sBAAsB,KAAK,UAAU,CAAC,CAAC,QAAQ,KAAK,UAAU,CAAC,CAAC,EAAE;AAE5E,UAAM,WAAW,WAAW;AAC5B,UAAM,WAAW,WAAW;AAC5B,WAAO,YAAY,WAAW,QAAQ;AAEtC,aAAS,YAAY,OAAO;AAC1B,UAAI,QAAQ,KAAK;AACf,eAAO;AAAA,MACT;AACA,UAAI,QAAQ,IAAI;AACd,eAAO;AAAA,MACT;AACA,aAAO;AAAA,IACT;AAEA,aAAS,aAAa;AACpB,YAAM,OAAO,YAAY,GAAG,MAAM;AAClC,YAAM,OAAO,YAAY,GAAG,MAAM;AAElC,UAAI,SAAS,UAAa,SAAS,QAAW;AAC5C,eAAO;AAAA,MACT;AAEA,UAAI,eAAe,IAAI,KAAK,eAAe,IAAI,GAAE;AAE/C,eAAO;AAAA,MACT;AAEA,UAAI,SAAS,MAAM;AACjB,eAAO;AAAA,MACT;AAEA,UAAI,SAAS,SAAS,SAAS,OAAO;AACpC,eAAO;AAAA,MACT;AAEA,aAAO;AAAA,IACT;AAMA,aAAS,aAAa;AACpB,YAAM,QAAQ,aAAa,CAAC;AAC5B,YAAM,QAAQ,aAAa,CAAC;AAC5B,UAAI,MAAM,WAAW,KAAK,MAAM,WAAW,GAAG;AAC5C,eAAO;AAAA,MACT;AACA,UAAI,MAAM,WAAW,KAAK,MAAM,WAAW,GAAG;AAE5C,cAAM,gBAAgB,wBAAwB,MAAM,CAAC,CAAC,MAAM,wBAAwB,MAAM,CAAC,CAAC,IAAI,IAAM;AACtG,cAAM,cAAc,yBAAyB,uBAAuB,MAAM,CAAC,CAAC,GAAG,uBAAuB,MAAM,CAAC,CAAC,CAAC;AAC/G,cAAM,YAAY,gBAAgB,IAAM,IAAM,kBAAkB,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC;AAClF,eAAO,cAAc,gBAAgB;AAAA,MACvC;AAEA,aAAO,yBAAyB,8BAA8B,CAAC,GAAG,8BAA8B,CAAC,CAAC;AAElG,eAAS,aAAa,QAAQ;AAC5B,YAAI,CAAC,OAAO,QAAQ;AAClB,iBAAO,CAAC;AAAA,QACV;AACA,eAAO,OAAO,OAAO,OAAO,OAAK,EAAE,QAAQ,KAAK;AAAA,MAClD;AAAA,IAEF;AAEA,aAAS,wBAAwB,OAAO;AACtC,YAAM,MAAM,MAAM,UAAU,KAAK,QAAM,GAAG,SAAS,GAAG;AACtD,UAAI,CAAC,KAAK;AACR,eAAO;AAAA,MACT;AAEA,UAAI,IAAI,MAAM,MAAM,QAAQ,GAAG;AAC7B,eAAO;AAAA,MACT;AACA,UAAI,IAAI,MAAM,MAAM,QAAQ,GAAG;AAC7B,eAAO;AAAA,MACT;AACA,aAAO,IAAI;AAAA,IACb;AAEA,aAAS,kBAAkB,QAAQ,QAAQ;AACzC,UAAI,OAAO,SAAS,OAAO,OAAO,SAAS,OAAM,OAAO,SAAS,OAAO,MAAO;AAC7E,eAAO;AAAA,MACT;AACA,YAAM,yBAA0B,OAAO,IAAI,SAAS,OAAO,IAAI,GAAG;AAClE,aAAO;AAAA,IACT;AAIA,aAAS,eAAe,cAAc;AACpC,aAAO,CAAC,OAAO,OAAO,KAAK,EAAE,SAAS,YAAY;AAAA,IACpD;AAEA,aAAS,yBAAyBA,IAAGC,IAAG;AACtC,UAAID,GAAE,WAAW,KAAKC,GAAE,WAAW,GAAG;AACpC,kBAAU,wBAAwB;AAClC,eAAO;AAAA,MACT;AAGA,YAAM,oBAAoB,CAAC,OAAO,OAAO,OAAO,OAAO,KAAK;AAE5D,UAAID,GAAE,SAAS,KAAK,KAAKC,GAAE,SAAS,KAAK,GAAG;AAC1C,YAAID,GAAE,KAAK,UAAQ,kBAAkB,SAAS,IAAI,CAAC,KAAK,CAACC,GAAE,KAAK,UAAQ,kBAAkB,SAAS,IAAI,CAAC,GAAG;AACzG,iBAAO,yBAAyBD,GAAE,OAAO,SAAO,CAAC,kBAAkB,SAAS,GAAG,CAAC,GAAGC,EAAC;AAAA,QACtF;AACA,YAAIA,GAAE,KAAK,UAAQ,kBAAkB,SAAS,IAAI,CAAC,KAAK,CAACD,GAAE,KAAK,UAAQ,kBAAkB,SAAS,IAAI,CAAC,GAAG;AACzG,iBAAO,yBAAyBA,IAAGC,GAAE,OAAO,SAAO,CAAC,kBAAkB,SAAS,GAAG,CAAC,CAAC;AAAA,QACtF;AAAA,MACF;AAKA,UAAID,GAAE,WAAWC,GAAE,UAAUD,GAAE,MAAM,CAAC,SAAS,UAAU,YAAYC,GAAE,KAAK,CAAC,GAAG;AAC9E,kBAAU,qBAAqB;AAC/B,eAAO;AAAA,MACT;AAIA,UAAID,GAAE,WAAW,KAAKA,GAAE,CAAC,MAAM,SAASC,GAAE,SAAS,GAAG;AACpD,eAAO;AAAA,MACT;AACA,UAAIA,GAAE,WAAW,KAAKA,GAAE,CAAC,MAAM,SAASD,GAAE,SAAS,GAAG;AACpD,eAAO;AAAA,MACT;AAOA,YAAM,eAAe,gBAAgBA,IAAGC,EAAC;AACzC,YAAM,QAAQD,GAAE,OAAO,SAAO,CAAC,aAAa,SAAS,GAAG,CAAC;AACzD,YAAM,QAAQC,GAAE,OAAO,SAAO,CAAC,aAAa,SAAS,GAAG,CAAC;AACzD,YAAM,SAAS,CAAC,GAAG,OAAO,GAAG,KAAK,EAAE,SAAS,KAAK;AAElD,UAAI,aAAa,SAAS,GAAG;AAC3B,gBAAQ,KAAK,OAAO,aAAa,KAAK,IAAI,CAAC,EAAE;AAC7C,YAAI,MAAM,WAAW,KAAK,MAAM,WAAW,KAAK,QAAQ;AACtD,gBAAM,0FAA0F,MAAM,CAAC,CAAC,UAAU,MAAM,CAAC,CAAC,wBAAwB;AAClJ,iBAAO;AAAA,QACT;AACA,cAAM,+CAA+C;AACrD,eAAO;AAAA,MACT;AAEA,YAAM,EAAC,gBAAgB,qBAAqB,UAAS,IAAI,eAAeD,IAAGC,EAAC;AAE5E,YAAM,wBAAwB,cAAc,IAAI,mBAAmB,yBAAyB;AAE5F,YAAM,qBAAqB,mBAAmB,IAAI,SAAS,EAAE;AAE7D,YAAM,eAAe,YAAY;AACjC,YAAM,gBAAgB,sBAAsB;AAC5C,YAAM,cAAe,YAAY,EAAE;AACnC,YAAM,iBAAkB,aAAa,EAAE;AAEvC,YAAM,oBAAoB,QAAQ,YAAY;AAC9C,YAAM,qBAAqB,QAAQ,sBAAsB;AACzD,YAAM,gCAAiC,iBAAiB,EAAE;AAC1D,YAAM,iCAAkC,kBAAkB,EAAE;AAE5D,YAAM,SAAS,OAAO,OAAO,MAAM,qBAAqB,iBAAiB,EAAE,QAAQ,CAAC,CAAC;AACrF,YAAM,iBAAiB,MAAM,EAAE;AAE/B,aAAO;AAAA,IACT;AAAA,EACF;AAEF;AAEA,SAAS,YAAY,QAAQ,QAAQ,UAAU;AAC7C,MAAI,CAAC,UAAU,CAAC,OAAO,QAAQ;AAC7B;AAAA,EACF;AACA,QAAM,OAAO,OAAO,OAAO,KAAK,OAAK,EAAE,QAAQ,KAAK;AACpD,MAAI,CAAC,MAAM;AACT,WAAO;AAAA,EACT;AACA,QAAM,QAAQ,KAAK,SAAS;AAE5B,MAAI,CAAC,OAAO;AACV,cAAU,GAAG,KAAK,4CAA4C,KAAK,GAAG;AACtE,WAAO;AAAA,EACT;AAEA,QAAM,OAAO,MAAM,MAAM,IAAI,EAAE;AAC/B,YAAU,GAAG,KAAK,gBAAgB,IAAI,GAAG;AACzC,SAAO;AACT;AAIA,SAAS,uBAAuB,MAAM,QAAQ,UAAU,WAAW,QAAW;AAC5E,MAAI,CAAC,MAAM;AACT,WAAO;AAAA,EACT;AAEA,MAAI,CAAC,QAAW,aAAa,WAAW,EAAE,SAAS,QAAQ,KAAK,KAAK,WAAW,GAAG;AACjF,cAAU,GAAG,KAAK,UAAU,IAAI,iDAAiD;AACjF,WAAO;AAAA,EACT;AACA,MAAI,CAAC,UAAU;AAIb,QAAI,SAAS,SAAS,SAAS,OAAQ;AACrC,gBAAU,GAAG,KAAK,UAAU,IAAI,uCAAuC;AACvE,aAAO;AAAA,IACT;AAAA,EACF;AAEA,QAAM,kBAAkB;AACxB,MAAI,CAAC,gBAAgB,KAAK,IAAI,GAAG;AAC/B,cAAU,GAAG,KAAK,UAAU,IAAI,kCAAkC;AAClE,WAAO;AAAA,EACT;AACA,SAAO;AACT;AAIA,SAAS,8BAA8B,QAAQ;AAK7C,QAAM,SAAS,OAAO,OAAO,OAAO,OAAK,EAAE,QAAQ,KAAK,EAAE,QAAQ,OAAK,uBAAuB,CAAC,CAAC;AAShG,SAAO,CAAC,GAAG,IAAI,IAAI,MAAM,CAAC,EAAE,KAAK;AACnC;AAEA,SAAS,uBAAuB,OAAO;AACrC,QAAM,oBAAoB,MAAM,UAAU,OAAO,QAAM,GAAG,SAAS,OAAO,GAAG,SAAS,GAAG,EAAE,OAAO,QAAM,uBAAuB,GAAG,OAAO,OAAO,CAAC;AAEjJ,QAAM,SAAS,kBAAkB,IAAI,QAAM,GAAG,KAAK,EAAE,KAAK;AAE1D,SAAO,CAAC,GAAG,IAAI,IAAI,MAAM,CAAC,EAAE,KAAK;AACnC;AAEA,SAAS,gBAAgB,OAAO,OAAO;AACrC,SAAO,MAAM,OAAO,SAAO,MAAM,SAAS,GAAG,CAAC;AAChD;",
+  "names": ["a", "b"]
 }