@natlibfi/melinda-record-matching 1.0.8 → 2.0.0

package/dist/match-detection/features/bib/standard-identifier-factory.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../../../src/match-detection/features/bib/standard-identifier-factory.js"],"names":["pattern","subfieldCodes","extract","compare","record","field","get","subfields","filter","code","includes","map","value","replace","a","b","length","bothHaveValidIdentifiers","maxValues","matchingValues","getValueCount","aValues","bValues","validOnly","getIdentifiers","aValue","some","bValue","values"],"mappings":";;;;;;;AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;eA4Be,CAAC;AAACA,EAAAA,OAAD;AAAUC,EAAAA;AAAV,CAAD,KAA8B;AAC3C,SAAO;AAACC,IAAAA,OAAD;AAAUC,IAAAA;AAAV,GAAP;;AAEA,WAASD,OAAT,CAAiBE,MAAjB,EAAyB;AACvB,UAAM,CAACC,KAAD,IAAUD,MAAM,CAACE,GAAP,CAAWN,OAAX,CAAhB;;AAEA,QAAIK,KAAJ,EAAW;AACT,aAAOA,KAAK,CAACE,SAAN,CACJC,MADI,CACG,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYR,aAAa,CAACS,QAAd,CAAuBD,IAAvB,CADf,EAEJE,GAFI,CAEA,CAAC;AAACF,QAAAA,IAAD;AAAOG,QAAAA;AAAP,OAAD,MAAoB;AAACH,QAAAA,IAAD;AAAOG,QAAAA,KAAK,EAAEA,KAAK,CAACC,OAAN,CAAc,KAAd,EAAqB,EAArB;AAAd,OAApB,CAFA,CAAP;AAGD;;AAED,WAAO,EAAP;AACD;;AAED,WAASV,OAAT,CAAiBW,CAAjB,EAAoBC,CAApB,EAAuB;AACrB,QAAID,CAAC,CAACE,MAAF,KAAa,CAAb,IAAkBD,CAAC,CAACC,MAAF,KAAa,CAAnC,EAAsC;AACpC,aAAO,CAAP;AACD;;AAED,QAAIC,wBAAwB,EAA5B,EAAgC;AAC9B,YAAM;AAACC,QAAAA,SAAD;AAAYC,QAAAA;AAAZ,UAA8BC,aAAa,CAAC,IAAD,CAAjD;AACA,aAAOD,cAAc,GAAGD,SAAjB,GAA6B,IAApC;AACD;;AAED,UAAM;AAACA,MAAAA,SAAD;AAAYC,MAAAA;AAAZ,QAA8BC,aAAa,EAAjD;AACA,WAAOD,cAAc,GAAGD,SAAjB,GAA6B,GAApC;;AAEA,aAASD,wBAAT,GAAoC;AAClC,YAAMI,OAAO,GAAGP,CAAC,CAACN,MAAF,CAAS,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAK,GAA9B,CAAhB;AACA,YAAMa,OAAO,GAAGR,CAAC,CAACN,MAAF,CAAS,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAK,GAA9B,CAAhB;AACA,aAAOY,OAAO,CAACL,MAAR,GAAiB,CAAjB,IAAsBM,OAAO,CAACN,MAAR,GAAiB,CAA9C;AACD;;AAED,aAASI,aAAT,CAAuBG,SAAS,GAAG,KAAnC,EAA0C;AACxC,YAAMF,OAAO,GAAGG,cAAc,CAACV,CAAD,CAA9B;AACA,YAAMQ,OAAO,GAAGE,cAAc,CAACT,CAAD,CAA9B;AAEA,aAAO;AACLG,QAAAA,SAAS,EAAEG,OAAO,CAACL,MAAR,GAAiBM,OAAO,CAACN,MAAzB,GAAkCK,OAAO,CAACL,MAA1C,GAAmDM,OAAO,CAACN,MADjE;AAELG,QAAAA,cAAc,EAAEE,OAAO,CAACb,MAAR,CAAeiB,MAAM,IAAIH,OAAO,CAACI,IAAR,CAAaC,MAAM,IAAIF,MAAM,KAAKE,MAAlC,CAAzB,EAAoEX;AAF/E,OAAP;;AAKA,eAASQ,cAAT,CAAwBI,MAAxB,EAAgC;AAC9B,YAAIL,SAAJ,EAAe;AACb,iBAAOK,MAAM,CACVpB,MADI,CACG,CAAC;AAACC,YAAAA;AAAD,WAAD,KAAYA,IAAI,KAAK,GADxB,EAEJE,GAFI,CAEA,CAAC;AAACC,YAAAA;AAAD,WAAD,KAAaA,KAFb,CAAP;AAGD;;AAED,eAAOgB,MAAM,CAACjB,GAAP,CAAW,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAaA,KAAxB,CAAP;AACD;AACF;AACF;AACF,C","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nexport default ({pattern, subfieldCodes}) => {\n  return {extract, compare};\n\n  function extract(record) {\n    const [field] = record.get(pattern);\n\n    if (field) {\n      return field.subfields\n        .filter(({code}) => subfieldCodes.includes(code))\n        .map(({code, value}) => ({code, value: value.replace(/-/ug, '')}));\n    }\n\n    return [];\n  }\n\n  function compare(a, b) {\n    if (a.length === 0 || b.length === 0) {\n      return 0;\n    }\n\n    if (bothHaveValidIdentifiers()) {\n      const {maxValues, matchingValues} = getValueCount(true);\n      return matchingValues / maxValues * 0.75;\n    }\n\n    const {maxValues, matchingValues} = getValueCount();\n    return matchingValues / maxValues * 0.2;\n\n    function bothHaveValidIdentifiers() {\n      const aValues = a.filter(({code}) => code === 'a');\n      const bValues = a.filter(({code}) => code === 'a');\n      return aValues.length > 0 && bValues.length > 0;\n    }\n\n    function getValueCount(validOnly = false) {\n      const aValues = getIdentifiers(a);\n      const bValues = getIdentifiers(b);\n\n      return {\n        maxValues: aValues.length > bValues.length ? aValues.length : bValues.length,\n        matchingValues: aValues.filter(aValue => bValues.some(bValue => aValue === bValue)).length\n      };\n\n      function getIdentifiers(values) {\n        if (validOnly) {\n          return values\n            .filter(({code}) => code === 'a')\n            .map(({value}) => value);\n        }\n\n        return values.map(({value}) => value);\n      }\n    }\n  }\n};\n"],"file":"standard-identifier-factory.js"}
+{"version":3,"sources":["../../../../src/match-detection/features/bib/standard-identifier-factory.js"],"names":["debug","debugData","extend","pattern","subfieldCodes","extract","compare","record","field","get","subfields","filter","code","includes","map","value","replace","a","b","length","bothHaveValidIdentifiers","maxValues","matchingValues","getValueCount","aValues","bValues","validOnly","getIdentifiers","aValue","some","bValue","values"],"mappings":";;;;;;;AA4BA;;;;AA5BA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AAIA,MAAMA,KAAK,GAAG,oBAAkB,iFAAlB,CAAd;AACA,MAAMC,SAAS,GAAGD,KAAK,CAACE,MAAN,CAAa,MAAb,CAAlB;;eAEe,CAAC;AAACC,EAAAA,OAAD;AAAUC,EAAAA;AAAV,CAAD,KAA8B;AAC3C,SAAO;AAACC,IAAAA,OAAD;AAAUC,IAAAA;AAAV,GAAP;;AAEA,WAASD,OAAT,CAAiBE,MAAjB,EAAyB;AACvB,UAAM,CAACC,KAAD,IAAUD,MAAM,CAACE,GAAP,CAAWN,OAAX,CAAhB;;AAEA,QAAIK,KAAJ,EAAW;AACT,aAAOA,KAAK,CAACE,SAAN,CACJC,MADI,CACG,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYR,aAAa,CAACS,QAAd,CAAuBD,IAAvB,CADf,EAEJE,GAFI,CAEA,CAAC;AAACF,QAAAA,IAAD;AAAOG,QAAAA;AAAP,OAAD,MAAoB;AAACH,QAAAA,IAAD;AAAOG,QAAAA,KAAK,EAAEA,KAAK,CAACC,OAAN,CAAc,KAAd,EAAqB,EAArB;AAAd,OAApB,CAFA,CAAP;AAGD;;AAED,WAAO,EAAP;AACD;;AAED,WAASV,OAAT,CAAiBW,CAAjB,EAAoBC,CAApB,EAAuB;AACrB,QAAID,CAAC,CAACE,MAAF,KAAa,CAAb,IAAkBD,CAAC,CAACC,MAAF,KAAa,CAAnC,EAAsC;AACpClB,MAAAA,SAAS,CAAE,mCAAF,CAAT;AACA,aAAO,CAAP;AACD;;AAED,QAAImB,wBAAwB,EAA5B,EAAgC;AAC9B,YAAM;AAACC,QAAAA,SAAD;AAAYC,QAAAA;AAAZ,UAA8BC,aAAa,CAAC,IAAD,CAAjD;AACA,aAAOD,cAAc,GAAGD,SAAjB,GAA6B,IAApC;AACD;;AAED,UAAM;AAACA,MAAAA,SAAD;AAAYC,MAAAA;AAAZ,QAA8BC,aAAa,EAAjD;AACA,WAAOD,cAAc,GAAGD,SAAjB,GAA6B,GAApC;;AAEA,aAASD,wBAAT,GAAoC;AAClC,YAAMI,OAAO,GAAGP,CAAC,CAACN,MAAF,CAAS,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAK,GAA9B,CAAhB;AACA,YAAMa,OAAO,GAAGR,CAAC,CAACN,MAAF,CAAS,CAAC;AAACC,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAK,GAA9B,CAAhB;AACA,aAAOY,OAAO,CAACL,MAAR,GAAiB,CAAjB,IAAsBM,OAAO,CAACN,MAAR,GAAiB,CAA9C;AACD;;AAED,aAASI,aAAT,CAAuBG,SAAS,GAAG,KAAnC,EAA0C;AACxC,YAAMF,OAAO,GAAGG,cAAc,CAACV,CAAD,CAA9B;AACA,YAAMQ,OAAO,GAAGE,cAAc,CAACT,CAAD,CAA9B;AAEA,aAAO;AACLG,QAAAA,SAAS,EAAEG,OAAO,CAACL,MAAR,GAAiBM,OAAO,CAACN,MAAzB,GAAkCK,OAAO,CAACL,MAA1C,GAAmDM,OAAO,CAACN,MADjE;AAELG,QAAAA,cAAc,EAAEE,OAAO,CAACb,MAAR,CAAeiB,MAAM,IAAIH,OAAO,CAACI,IAAR,CAAaC,MAAM,IAAIF,MAAM,KAAKE,MAAlC,CAAzB,EAAoEX;AAF/E,OAAP;;AAKA,eAASQ,cAAT,CAAwBI,MAAxB,EAAgC;AAC9B,YAAIL,SAAJ,EAAe;AACb,iBAAOK,MAAM,CACVpB,MADI,CACG,CAAC;AAACC,YAAAA;AAAD,WAAD,KAAYA,IAAI,KAAK,GADxB,EAEJE,GAFI,CAEA,CAAC;AAACC,YAAAA;AAAD,WAAD,KAAaA,KAFb,CAAP;AAGD;;AAED,eAAOgB,MAAM,CAACjB,GAAP,CAAW,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAaA,KAAxB,CAAP;AACD;AACF;AACF;AACF,C","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\n\nconst debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features:standard-identifiers');\nconst debugData = debug.extend('data');\n\nexport default ({pattern, subfieldCodes}) => {\n  return {extract, compare};\n\n  function extract(record) {\n    const [field] = record.get(pattern);\n\n    if (field) {\n      return field.subfields\n        .filter(({code}) => subfieldCodes.includes(code))\n        .map(({code, value}) => ({code, value: value.replace(/-/ug, '')}));\n    }\n\n    return [];\n  }\n\n  function compare(a, b) {\n    if (a.length === 0 || b.length === 0) {\n      debugData(`No standardidentifiers to compare`);\n      return 0;\n    }\n\n    if (bothHaveValidIdentifiers()) {\n      const {maxValues, matchingValues} = getValueCount(true);\n      return matchingValues / maxValues * 0.75;\n    }\n\n    const {maxValues, matchingValues} = getValueCount();\n    return matchingValues / maxValues * 0.2;\n\n    function bothHaveValidIdentifiers() {\n      const aValues = a.filter(({code}) => code === 'a');\n      const bValues = a.filter(({code}) => code === 'a');\n      return aValues.length > 0 && bValues.length > 0;\n    }\n\n    function getValueCount(validOnly = false) {\n      const aValues = getIdentifiers(a);\n      const bValues = getIdentifiers(b);\n\n      return {\n        maxValues: aValues.length > bValues.length ? aValues.length : bValues.length,\n        matchingValues: aValues.filter(aValue => bValues.some(bValue => aValue === bValue)).length\n      };\n\n      function getIdentifiers(values) {\n        if (validOnly) {\n          return values\n            .filter(({code}) => code === 'a')\n            .map(({value}) => value);\n        }\n\n        return values.map(({value}) => value);\n      }\n    }\n  }\n};\n"],"file":"standard-identifier-factory.js"}

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

@@ -1 +1 @@
-{"version":3,"sources":["../../../../src/match-detection/features/bib/title.js"],"names":["treshold","name","extract","record","title","getTitle","replace","toLowerCase","field","get","subfields","filter","code","includes","map","value","join","compare","a","b","debug","distance","maxLength","getMaxLength","percentage","length"],"mappings":";;;;;;;AA4BA;;AACA;;;;AA7BA;;;;;;;;;;;;;;;;;;;;;;;;;;;eA+Be,CAAC;AAACA,EAAAA,QAAQ,GAAG;AAAZ,IAAkB,EAAnB,MAA2B;AACxCC,EAAAA,IAAI,EAAE,OADkC;AAExCC,EAAAA,OAAO,EAAEC,MAAM,IAAI;AACjB,UAAMC,KAAK,GAAGC,QAAQ,EAAtB;;AAEA,QAAID,KAAJ,EAAW;AACT,aAAO,CAACA,KAAK,CAACE,OAAN,CAAc,2BAAd,EAA2C,EAA3C,EAA+CC,WAA/C,EAAD,CAAP;AACD;;AAED,WAAO,EAAP;;AAEA,aAASF,QAAT,GAAoB;AAClB,YAAM,CAACG,KAAD,IAAUL,MAAM,CAACM,GAAP,CAAW,QAAX,CAAhB;;AAEA,UAAID,KAAJ,EAAW;AACT,eAAOA,KAAK,CAACE,SAAN,CACJC,MADI,CACG,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAY,CAAC,GAAD,EAAM,GAAN,EAAWC,QAAX,CAAoBD,IAApB,CADf,EAEJE,GAFI,CAEA,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAaA,KAFb,EAGJC,IAHI,CAGC,EAHD,CAAP;AAID;AACF;AACF,GArBuC;AAsBxCC,EAAAA,OAAO,EAAE,CAACC,CAAD,EAAIC,CAAJ,KAAU;AACjB,UAAMC,KAAK,GAAG,oBAAkB,sEAAlB,CAAd;AACA,UAAMC,QAAQ,GAAG,kCAAMH,CAAC,CAAC,CAAD,CAAP,EAAYC,CAAC,CAAC,CAAD,CAAb,CAAjB;;AAEA,QAAIE,QAAQ,KAAK,CAAjB,EAAoB;AAClB,aAAO,GAAP;AACD;;AAED,UAAMC,SAAS,GAAGC,YAAY,EAA9B;AACA,UAAMC,UAAU,GAAGH,QAAQ,GAAGC,SAAX,GAAuB,GAA1C;AAEAF,IAAAA,KAAK,CAAE,IAAGF,CAAE,SAAQC,CAAE,mBAAkBG,SAAU,gBAAeD,QAAS,kBAAiBG,UAAW,EAAjG,CAAL;;AAEA,QAAIA,UAAU,IAAIxB,QAAlB,EAA4B;AAC1B,aAAO,GAAP;AACD;;AAED,WAAO,CAAC,GAAR;;AAEA,aAASuB,YAAT,GAAwB;AACtB,aAAOL,CAAC,CAAC,CAAD,CAAD,CAAKO,MAAL,GAAcN,CAAC,CAAC,CAAD,CAAD,CAAKM,MAAnB,GAA4BP,CAAC,CAAC,CAAD,CAAD,CAAKO,MAAjC,GAA0CN,CAAC,CAAC,CAAD,CAAD,CAAKM,MAAtD;AACD;AAEF;AA7CuC,CAA3B,C","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\nimport {LevenshteinDistance as leven} from 'natural';\n\nexport default ({treshold = 10} = {}) => ({\n  name: 'Title',\n  extract: record => {\n    const title = getTitle();\n\n    if (title) {\n      return [title.replace(/[^\\p{Letter}\\p{Number}]/gu, '').toLowerCase()];\n    }\n\n    return [];\n\n    function getTitle() {\n      const [field] = record.get(/^245$/u);\n\n      if (field) {\n        return field.subfields\n          .filter(({code}) => ['a', 'b'].includes(code))\n          .map(({value}) => value)\n          .join('');\n      }\n    }\n  },\n  compare: (a, b) => {\n    const debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features/bib/title');\n    const distance = leven(a[0], b[0]);\n\n    if (distance === 0) {\n      return 0.5;\n    }\n\n    const maxLength = getMaxLength();\n    const percentage = distance / maxLength * 100;\n\n    debug(`'${a}' vs '${b}': Max length = ${maxLength}, distance = ${distance}, percentage = ${percentage}`);\n\n    if (percentage <= treshold) {\n      return 0.3;\n    }\n\n    return -0.5;\n\n    function getMaxLength() {\n      return a[0].length > b[0].length ? a[0].length : b[0].length;\n    }\n\n  }\n});\n"],"file":"title.js"}
+{"version":3,"sources":["../../../../src/match-detection/features/bib/title.js"],"names":["treshold","name","extract","record","title","getTitle","replace","toLowerCase","field","get","subfields","filter","code","includes","map","value","join","compare","a","b","debug","distance","maxLength","getMaxLength","percentage","length"],"mappings":";;;;;;;AA4BA;;AACA;;;;AA7BA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;eAKe,CAAC;AAACA,EAAAA,QAAQ,GAAG;AAAZ,IAAkB,EAAnB,MAA2B;AACxCC,EAAAA,IAAI,EAAE,OADkC;AAExCC,EAAAA,OAAO,EAAEC,MAAM,IAAI;AACjB,UAAMC,KAAK,GAAGC,QAAQ,EAAtB;;AAEA,QAAID,KAAJ,EAAW;AACT,aAAO,CAACA,KAAK,CAACE,OAAN,CAAc,2BAAd,EAA2C,EAA3C,EAA+CC,WAA/C,EAAD,CAAP;AACD;;AAED,WAAO,EAAP;;AAEA,aAASF,QAAT,GAAoB;AAClB,YAAM,CAACG,KAAD,IAAUL,MAAM,CAACM,GAAP,CAAW,QAAX,CAAhB;;AAEA,UAAID,KAAJ,EAAW;AACT,eAAOA,KAAK,CAACE,SAAN,CACJC,MADI,CACG,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAY,CAAC,GAAD,EAAM,GAAN,EAAWC,QAAX,CAAoBD,IAApB,CADf,EAEJE,GAFI,CAEA,CAAC;AAACC,UAAAA;AAAD,SAAD,KAAaA,KAFb,EAGJC,IAHI,CAGC,EAHD,CAAP;AAID;AACF;AACF,GArBuC;AAsBxCC,EAAAA,OAAO,EAAE,CAACC,CAAD,EAAIC,CAAJ,KAAU;AACjB,UAAMC,KAAK,GAAG,oBAAkB,sEAAlB,CAAd;AACA,UAAMC,QAAQ,GAAG,kCAAMH,CAAC,CAAC,CAAD,CAAP,EAAYC,CAAC,CAAC,CAAD,CAAb,CAAjB;;AAEA,QAAIE,QAAQ,KAAK,CAAjB,EAAoB;AAClB,aAAO,GAAP;AACD;;AAED,UAAMC,SAAS,GAAGC,YAAY,EAA9B;AACA,UAAMC,UAAU,GAAGH,QAAQ,GAAGC,SAAX,GAAuB,GAA1C;AAEAF,IAAAA,KAAK,CAAE,IAAGF,CAAE,SAAQC,CAAE,mBAAkBG,SAAU,gBAAeD,QAAS,kBAAiBG,UAAW,EAAjG,CAAL;;AAEA,QAAIA,UAAU,IAAIxB,QAAlB,EAA4B;AAC1B,aAAO,GAAP;AACD;;AAED,WAAO,CAAC,GAAR;;AAEA,aAASuB,YAAT,GAAwB;AACtB,aAAOL,CAAC,CAAC,CAAD,CAAD,CAAKO,MAAL,GAAcN,CAAC,CAAC,CAAD,CAAD,CAAKM,MAAnB,GAA4BP,CAAC,CAAC,CAAD,CAAD,CAAKO,MAAjC,GAA0CN,CAAC,CAAC,CAAD,CAAD,CAAKM,MAAtD;AACD;AAEF;AA7CuC,CAA3B,C","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\nimport {LevenshteinDistance as leven} from 'natural';\n\nexport default ({treshold = 10} = {}) => ({\n  name: 'Title',\n  extract: record => {\n    const title = getTitle();\n\n    if (title) {\n      return [title.replace(/[^\\p{Letter}\\p{Number}]/gu, '').toLowerCase()];\n    }\n\n    return [];\n\n    function getTitle() {\n      const [field] = record.get(/^245$/u);\n\n      if (field) {\n        return field.subfields\n          .filter(({code}) => ['a', 'b'].includes(code))\n          .map(({value}) => value)\n          .join('');\n      }\n    }\n  },\n  compare: (a, b) => {\n    const debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection:features/bib/title');\n    const distance = leven(a[0], b[0]);\n\n    if (distance === 0) {\n      return 0.5;\n    }\n\n    const maxLength = getMaxLength();\n    const percentage = distance / maxLength * 100;\n\n    debug(`'${a}' vs '${b}': Max length = ${maxLength}, distance = ${distance}, percentage = ${percentage}`);\n\n    if (percentage <= treshold) {\n      return 0.3;\n    }\n\n    return -0.5;\n\n    function getMaxLength() {\n      return a[0].length > b[0].length ? a[0].length : b[0].length;\n    }\n\n  }\n});\n"],"file":"title.js"}

package/dist/match-detection/features/index.js

@@ -9,7 +9,7 @@ var bib = _interopRequireWildcard(require("./bib"));
 
 exports.bib = bib;
 
-function _getRequireWildcardCache() { if (typeof WeakMap !== "function") return null; var cache = new WeakMap(); _getRequireWildcardCache = function () { return cache; }; return cache; }
+function _getRequireWildcardCache(nodeInterop) { if (typeof WeakMap !== "function") return null; var cacheBabelInterop = new WeakMap(); var cacheNodeInterop = new WeakMap(); return (_getRequireWildcardCache = function (nodeInterop) { return nodeInterop ? cacheNodeInterop : cacheBabelInterop; })(nodeInterop); }
 
-function _interopRequireWildcard(obj) { if (obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
+function _interopRequireWildcard(obj, nodeInterop) { if (!nodeInterop && obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(nodeInterop); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (key !== "default" && Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
 //# sourceMappingURL=index.js.map

package/dist/match-detection/index.js

@@ -11,12 +11,14 @@ var features = _interopRequireWildcard(require("./features"));
 
 exports.features = features;
 
-function _getRequireWildcardCache() { if (typeof WeakMap !== "function") return null; var cache = new WeakMap(); _getRequireWildcardCache = function () { return cache; }; return cache; }
+function _getRequireWildcardCache(nodeInterop) { if (typeof WeakMap !== "function") return null; var cacheBabelInterop = new WeakMap(); var cacheNodeInterop = new WeakMap(); return (_getRequireWildcardCache = function (nodeInterop) { return nodeInterop ? cacheNodeInterop : cacheBabelInterop; })(nodeInterop); }
 
-function _interopRequireWildcard(obj) { if (obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
+function _interopRequireWildcard(obj, nodeInterop) { if (!nodeInterop && obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(nodeInterop); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (key !== "default" && Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
 
 function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
 
+/* eslint-disable no-console */
+
 /**
 *
 * @licstart  The following is the entire license notice for the JavaScript code in this file.
@@ -44,16 +46,18 @@ function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { de
 * for the JavaScript code in this file.
 *
 */
+// eslint-disable-next-line max-statements
 var _default = ({
   strategy,
   treshold = 0.9
 }) => (recordA, recordB) => {
   const minProbabilityQuantifier = 0.5;
   const debug = (0, _debug.default)('@natlibfi/melinda-record-matching:match-detection');
+  const debugData = debug.extend('data');
   const featuresA = extractFeatures(recordA);
   const featuresB = extractFeatures(recordB);
-  debug(`Features (a): ${JSON.stringify(featuresA)}`);
-  debug(`Features (b): ${JSON.stringify(featuresB)}`);
+  debugData(`Features (a): ${JSON.stringify(featuresA)}`);
+  debugData(`Features (b): ${JSON.stringify(featuresB)}`);
   const featurePairs = generateFeaturePairs();
   const similarityVector = generateSimilarityVector();
 
@@ -66,7 +70,7 @@ var _default = ({
     };
   }
 
-  debug(`No feature yielded minimum probability amount of points (${minProbabilityQuantifier})`);
+  debugData(`No feature yielded minimum probability amount of points (${minProbabilityQuantifier})`);
   return {
     match: false,
     probability: 0.0
@@ -104,7 +108,7 @@ var _default = ({
         points
       };
     });
-    debug(`Points: ${JSON.stringify(compared)}`);
+    debugData(`Points: ${JSON.stringify(compared)}`);
     return compared.map(({
       points
     }) => points);

package/dist/match-detection/index.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../src/match-detection/index.js"],"names":["strategy","treshold","recordA","recordB","minProbabilityQuantifier","debug","featuresA","extractFeatures","featuresB","JSON","stringify","featurePairs","generateFeaturePairs","similarityVector","generateSimilarityVector","some","v","probability","calculateprobability","match","reduce","acc","record","name","extract","concat","value","compared","map","a","b","compare","find","featureName","points","pairs","generatePairs","missingFeatures","findMissing","index","filter","length"],"mappings":";;;;;;;AA4BA;;AACA;;;;;;;;;;AA7BA;;;;;;;;;;;;;;;;;;;;;;;;;;;eAiCe,CAAC;AAACA,EAAAA,QAAD;AAAWC,EAAAA,QAAQ,GAAG;AAAtB,CAAD,KAAgC,CAACC,OAAD,EAAUC,OAAV,KAAsB;AACnE,QAAMC,wBAAwB,GAAG,GAAjC;AAEA,QAAMC,KAAK,GAAG,oBAAkB,mDAAlB,CAAd;AACA,QAAMC,SAAS,GAAGC,eAAe,CAACL,OAAD,CAAjC;AACA,QAAMM,SAAS,GAAGD,eAAe,CAACJ,OAAD,CAAjC;AAEAE,EAAAA,KAAK,CAAE,iBAAgBI,IAAI,CAACC,SAAL,CAAeJ,SAAf,CAA0B,EAA5C,CAAL;AACAD,EAAAA,KAAK,CAAE,iBAAgBI,IAAI,CAACC,SAAL,CAAeF,SAAf,CAA0B,EAA5C,CAAL;AAEA,QAAMG,YAAY,GAAGC,oBAAoB,EAAzC;AACA,QAAMC,gBAAgB,GAAGC,wBAAwB,EAAjD;;AAEA,MAAID,gBAAgB,CAACE,IAAjB,CAAsBC,CAAC,IAAIA,CAAC,IAAIZ,wBAAhC,CAAJ,EAA+D;AAC7D,UAAMa,WAAW,GAAGC,oBAAoB,EAAxC;AACAb,IAAAA,KAAK,CAAE,gBAAeY,WAAY,eAAchB,QAAS,GAApD,CAAL;AACA,WAAO;AAACkB,MAAAA,KAAK,EAAEF,WAAW,IAAIhB,QAAvB;AAAiCgB,MAAAA;AAAjC,KAAP;AACD;;AAEDZ,EAAAA,KAAK,CAAE,4DAA2DD,wBAAyB,GAAtF,CAAL;AACA,SAAO;AAACe,IAAAA,KAAK,EAAE,KAAR;AAAeF,IAAAA,WAAW,EAAE;AAA5B,GAAP;;AAEA,WAASC,oBAAT,GAAgC;AAC9B,UAAMD,WAAW,GAAGJ,gBAAgB,CAACO,MAAjB,CAAwB,CAACC,GAAD,EAAML,CAAN,KAAYK,GAAG,GAAGL,CAA1C,EAA6C,GAA7C,CAApB;AACA,WAAOC,WAAW,GAAG,GAAd,GAAoB,GAApB,GAA0BA,WAAjC;AACD;;AAED,WAASV,eAAT,CAAyBe,MAAzB,EAAiC;AAC/B,WAAOtB,QAAQ,CAACoB,MAAT,CAAgB,CAACC,GAAD,EAAM;AAACE,MAAAA,IAAD;AAAOC,MAAAA;AAAP,KAAN,KAA0BH,GAAG,CAACI,MAAJ,CAAW;AAACF,MAAAA,IAAD;AAAOG,MAAAA,KAAK,EAAEF,OAAO,CAACF,MAAD;AAArB,KAAX,CAA1C,EAAsF,EAAtF,CAAP;AACD;;AAED,WAASR,wBAAT,GAAoC;AAClC,UAAMa,QAAQ,GAAGhB,YAAY,CAACiB,GAAb,CAAiB,CAAC;AAACL,MAAAA,IAAD;AAAOM,MAAAA,CAAP;AAAUC,MAAAA;AAAV,KAAD,KAAkB;AAClD,YAAM;AAACC,QAAAA;AAAD,UAAY/B,QAAQ,CAACgC,IAAT,CAAc,CAAC;AAACT,QAAAA,IAAI,EAAEU;AAAP,OAAD,KAAyBV,IAAI,KAAKU,WAAhD,CAAlB;AACA,YAAMC,MAAM,GAAGH,OAAO,CAACF,CAAD,EAAIC,CAAJ,CAAtB;AACA,aAAO;AAACP,QAAAA,IAAD;AAAOW,QAAAA;AAAP,OAAP;AACD,KAJgB,CAAjB;AAMA7B,IAAAA,KAAK,CAAE,WAAUI,IAAI,CAACC,SAAL,CAAeiB,QAAf,CAAyB,EAArC,CAAL;AACA,WAAOA,QAAQ,CAACC,GAAT,CAAa,CAAC;AAACM,MAAAA;AAAD,KAAD,KAAcA,MAA3B,CAAP;AACD;;AAED,WAAStB,oBAAT,GAAgC;AAC9B,UAAMuB,KAAK,GAAGC,aAAa,EAA3B;AACA,UAAMC,eAAe,GAAGC,WAAW,EAAnC;AAEAjC,IAAAA,KAAK,CAAE,2FAA0FI,IAAI,CAACC,SAAL,CAAe2B,eAAf,CAAgC,EAA5H,CAAL;AACA,WAAOF,KAAP;;AAEA,aAASC,aAAT,GAAyB;AACvB,aAAO9B,SAAS,CACbc,MADI,CACG,CAACC,GAAD,EAAM;AAACE,QAAAA,IAAD;AAAOG,QAAAA;AAAP,OAAN,EAAqBa,KAArB,KAA+BlB,GAAG,CAACI,MAAJ,CAAW;AAChDF,QAAAA,IADgD;AAEhDM,QAAAA,CAAC,EAAEH,KAF6C;AAGhDI,QAAAA,CAAC,EAAEtB,SAAS,CAAC+B,KAAD,CAAT,CAAiBb;AAH4B,OAAX,CADlC,EAKD,EALC,EAMJc,MANI,CAMG,CAAC;AAACX,QAAAA,CAAD;AAAIC,QAAAA;AAAJ,OAAD,KAAY;AAClB,YAAID,CAAC,CAACY,MAAF,KAAa,CAAb,IAAkBX,CAAC,CAACW,MAAF,KAAa,CAAnC,EAAsC;AACpC,iBAAO,KAAP;AACD;;AAED,eAAO,IAAP;AACD,OAZI,CAAP;AAaD;;AAED,aAASH,WAAT,GAAuB;AACrB,aAAOhC,SAAS,CACbsB,GADI,CACA,CAAC;AAACL,QAAAA;AAAD,OAAD,KAAYA,IADZ,EAEJiB,MAFI,CAEGxB,CAAC,IAAImB,KAAK,CAACpB,IAAN,CAAW,CAAC;AAACQ,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAKP,CAAhC,MAAuC,KAF/C,CAAP;AAGD;AACF;AACF,C","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\nimport * as features from './features';\n\nexport {features};\n\nexport default ({strategy, treshold = 0.9}) => (recordA, recordB) => {\n  const minProbabilityQuantifier = 0.5;\n\n  const debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection');\n  const featuresA = extractFeatures(recordA);\n  const featuresB = extractFeatures(recordB);\n\n  debug(`Features (a): ${JSON.stringify(featuresA)}`);\n  debug(`Features (b): ${JSON.stringify(featuresB)}`);\n\n  const featurePairs = generateFeaturePairs();\n  const similarityVector = generateSimilarityVector();\n\n  if (similarityVector.some(v => v >= minProbabilityQuantifier)) {\n    const probability = calculateprobability();\n    debug(`probability: ${probability} (Treshold: ${treshold})`);\n    return {match: probability >= treshold, probability};\n  }\n\n  debug(`No feature yielded minimum probability amount of points (${minProbabilityQuantifier})`);\n  return {match: false, probability: 0.0};\n\n  function calculateprobability() {\n    const probability = similarityVector.reduce((acc, v) => acc + v, 0.0);\n    return probability > 1.0 ? 1.0 : probability;\n  }\n\n  function extractFeatures(record) {\n    return strategy.reduce((acc, {name, extract}) => acc.concat({name, value: extract(record)}), []);\n  }\n\n  function generateSimilarityVector() {\n    const compared = featurePairs.map(({name, a, b}) => {\n      const {compare} = strategy.find(({name: featureName}) => name === featureName);\n      const points = compare(a, b);\n      return {name, points};\n    });\n\n    debug(`Points: ${JSON.stringify(compared)}`);\n    return compared.map(({points}) => points);\n  }\n\n  function generateFeaturePairs() {\n    const pairs = generatePairs();\n    const missingFeatures = findMissing();\n\n    debug(`Not comparing the following features because one, or both records are missing features: ${JSON.stringify(missingFeatures)}`);\n    return pairs;\n\n    function generatePairs() {\n      return featuresA\n        .reduce((acc, {name, value}, index) => acc.concat({\n          name,\n          a: value,\n          b: featuresB[index].value\n        }), [])\n        .filter(({a, b}) => {\n          if (a.length === 0 || b.length === 0) {\n            return false;\n          }\n\n          return true;\n        });\n    }\n\n    function findMissing() {\n      return featuresA\n        .map(({name}) => name)\n        .filter(v => pairs.some(({name}) => name === v) === false);\n    }\n  }\n};\n"],"file":"index.js"}
+{"version":3,"sources":["../../src/match-detection/index.js"],"names":["strategy","treshold","recordA","recordB","minProbabilityQuantifier","debug","debugData","extend","featuresA","extractFeatures","featuresB","JSON","stringify","featurePairs","generateFeaturePairs","similarityVector","generateSimilarityVector","some","v","probability","calculateprobability","match","reduce","acc","record","name","extract","concat","value","compared","map","a","b","compare","find","featureName","points","pairs","generatePairs","missingFeatures","findMissing","index","filter","length"],"mappings":";;;;;;;AA6BA;;AACA;;;;;;;;;;AA9BA;;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AAOA;eACe,CAAC;AAACA,EAAAA,QAAD;AAAWC,EAAAA,QAAQ,GAAG;AAAtB,CAAD,KAAgC,CAACC,OAAD,EAAUC,OAAV,KAAsB;AACnE,QAAMC,wBAAwB,GAAG,GAAjC;AAEA,QAAMC,KAAK,GAAG,oBAAkB,mDAAlB,CAAd;AACA,QAAMC,SAAS,GAAGD,KAAK,CAACE,MAAN,CAAa,MAAb,CAAlB;AACA,QAAMC,SAAS,GAAGC,eAAe,CAACP,OAAD,CAAjC;AACA,QAAMQ,SAAS,GAAGD,eAAe,CAACN,OAAD,CAAjC;AAEAG,EAAAA,SAAS,CAAE,iBAAgBK,IAAI,CAACC,SAAL,CAAeJ,SAAf,CAA0B,EAA5C,CAAT;AACAF,EAAAA,SAAS,CAAE,iBAAgBK,IAAI,CAACC,SAAL,CAAeF,SAAf,CAA0B,EAA5C,CAAT;AAEA,QAAMG,YAAY,GAAGC,oBAAoB,EAAzC;AACA,QAAMC,gBAAgB,GAAGC,wBAAwB,EAAjD;;AAEA,MAAID,gBAAgB,CAACE,IAAjB,CAAsBC,CAAC,IAAIA,CAAC,IAAId,wBAAhC,CAAJ,EAA+D;AAC7D,UAAMe,WAAW,GAAGC,oBAAoB,EAAxC;AACAf,IAAAA,KAAK,CAAE,gBAAec,WAAY,eAAclB,QAAS,GAApD,CAAL;AACA,WAAO;AAACoB,MAAAA,KAAK,EAAEF,WAAW,IAAIlB,QAAvB;AAAiCkB,MAAAA;AAAjC,KAAP;AACD;;AAEDb,EAAAA,SAAS,CAAE,4DAA2DF,wBAAyB,GAAtF,CAAT;AACA,SAAO;AAACiB,IAAAA,KAAK,EAAE,KAAR;AAAeF,IAAAA,WAAW,EAAE;AAA5B,GAAP;;AAEA,WAASC,oBAAT,GAAgC;AAC9B,UAAMD,WAAW,GAAGJ,gBAAgB,CAACO,MAAjB,CAAwB,CAACC,GAAD,EAAML,CAAN,KAAYK,GAAG,GAAGL,CAA1C,EAA6C,GAA7C,CAApB;AACA,WAAOC,WAAW,GAAG,GAAd,GAAoB,GAApB,GAA0BA,WAAjC;AACD;;AAED,WAASV,eAAT,CAAyBe,MAAzB,EAAiC;AAC/B,WAAOxB,QAAQ,CAACsB,MAAT,CAAgB,CAACC,GAAD,EAAM;AAACE,MAAAA,IAAD;AAAOC,MAAAA;AAAP,KAAN,KAA0BH,GAAG,CAACI,MAAJ,CAAW;AAACF,MAAAA,IAAD;AAAOG,MAAAA,KAAK,EAAEF,OAAO,CAACF,MAAD;AAArB,KAAX,CAA1C,EAAsF,EAAtF,CAAP;AACD;;AAED,WAASR,wBAAT,GAAoC;AAClC,UAAMa,QAAQ,GAAGhB,YAAY,CAACiB,GAAb,CAAiB,CAAC;AAACL,MAAAA,IAAD;AAAOM,MAAAA,CAAP;AAAUC,MAAAA;AAAV,KAAD,KAAkB;AAClD,YAAM;AAACC,QAAAA;AAAD,UAAYjC,QAAQ,CAACkC,IAAT,CAAc,CAAC;AAACT,QAAAA,IAAI,EAAEU;AAAP,OAAD,KAAyBV,IAAI,KAAKU,WAAhD,CAAlB;AACA,YAAMC,MAAM,GAAGH,OAAO,CAACF,CAAD,EAAIC,CAAJ,CAAtB;AACA,aAAO;AAACP,QAAAA,IAAD;AAAOW,QAAAA;AAAP,OAAP;AACD,KAJgB,CAAjB;AAMA9B,IAAAA,SAAS,CAAE,WAAUK,IAAI,CAACC,SAAL,CAAeiB,QAAf,CAAyB,EAArC,CAAT;AACA,WAAOA,QAAQ,CAACC,GAAT,CAAa,CAAC;AAACM,MAAAA;AAAD,KAAD,KAAcA,MAA3B,CAAP;AACD;;AAED,WAAStB,oBAAT,GAAgC;AAC9B,UAAMuB,KAAK,GAAGC,aAAa,EAA3B;AACA,UAAMC,eAAe,GAAGC,WAAW,EAAnC;AAEAnC,IAAAA,KAAK,CAAE,2FAA0FM,IAAI,CAACC,SAAL,CAAe2B,eAAf,CAAgC,EAA5H,CAAL;AACA,WAAOF,KAAP;;AAEA,aAASC,aAAT,GAAyB;AACvB,aAAO9B,SAAS,CACbc,MADI,CACG,CAACC,GAAD,EAAM;AAACE,QAAAA,IAAD;AAAOG,QAAAA;AAAP,OAAN,EAAqBa,KAArB,KAA+BlB,GAAG,CAACI,MAAJ,CAAW;AAChDF,QAAAA,IADgD;AAEhDM,QAAAA,CAAC,EAAEH,KAF6C;AAGhDI,QAAAA,CAAC,EAAEtB,SAAS,CAAC+B,KAAD,CAAT,CAAiBb;AAH4B,OAAX,CADlC,EAKD,EALC,EAMJc,MANI,CAMG,CAAC;AAACX,QAAAA,CAAD;AAAIC,QAAAA;AAAJ,OAAD,KAAY;AAClB,YAAID,CAAC,CAACY,MAAF,KAAa,CAAb,IAAkBX,CAAC,CAACW,MAAF,KAAa,CAAnC,EAAsC;AACpC,iBAAO,KAAP;AACD;;AAED,eAAO,IAAP;AACD,OAZI,CAAP;AAaD;;AAED,aAASH,WAAT,GAAuB;AACrB,aAAOhC,SAAS,CACbsB,GADI,CACA,CAAC;AAACL,QAAAA;AAAD,OAAD,KAAYA,IADZ,EAEJiB,MAFI,CAEGxB,CAAC,IAAImB,KAAK,CAACpB,IAAN,CAAW,CAAC;AAACQ,QAAAA;AAAD,OAAD,KAAYA,IAAI,KAAKP,CAAhC,MAAuC,KAF/C,CAAP;AAGD;AACF;AACF,C","sourcesContent":["/* eslint-disable no-console */\n/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\nimport * as features from './features';\n\nexport {features};\n\n// eslint-disable-next-line max-statements\nexport default ({strategy, treshold = 0.9}) => (recordA, recordB) => {\n  const minProbabilityQuantifier = 0.5;\n\n  const debug = createDebugLogger('@natlibfi/melinda-record-matching:match-detection');\n  const debugData = debug.extend('data');\n  const featuresA = extractFeatures(recordA);\n  const featuresB = extractFeatures(recordB);\n\n  debugData(`Features (a): ${JSON.stringify(featuresA)}`);\n  debugData(`Features (b): ${JSON.stringify(featuresB)}`);\n\n  const featurePairs = generateFeaturePairs();\n  const similarityVector = generateSimilarityVector();\n\n  if (similarityVector.some(v => v >= minProbabilityQuantifier)) {\n    const probability = calculateprobability();\n    debug(`probability: ${probability} (Treshold: ${treshold})`);\n    return {match: probability >= treshold, probability};\n  }\n\n  debugData(`No feature yielded minimum probability amount of points (${minProbabilityQuantifier})`);\n  return {match: false, probability: 0.0};\n\n  function calculateprobability() {\n    const probability = similarityVector.reduce((acc, v) => acc + v, 0.0);\n    return probability > 1.0 ? 1.0 : probability;\n  }\n\n  function extractFeatures(record) {\n    return strategy.reduce((acc, {name, extract}) => acc.concat({name, value: extract(record)}), []);\n  }\n\n  function generateSimilarityVector() {\n    const compared = featurePairs.map(({name, a, b}) => {\n      const {compare} = strategy.find(({name: featureName}) => name === featureName);\n      const points = compare(a, b);\n      return {name, points};\n    });\n\n    debugData(`Points: ${JSON.stringify(compared)}`);\n    return compared.map(({points}) => points);\n  }\n\n  function generateFeaturePairs() {\n    const pairs = generatePairs();\n    const missingFeatures = findMissing();\n\n    debug(`Not comparing the following features because one, or both records are missing features: ${JSON.stringify(missingFeatures)}`);\n    return pairs;\n\n    function generatePairs() {\n      return featuresA\n        .reduce((acc, {name, value}, index) => acc.concat({\n          name,\n          a: value,\n          b: featuresB[index].value\n        }), [])\n        .filter(({a, b}) => {\n          if (a.length === 0 || b.length === 0) {\n            return false;\n          }\n\n          return true;\n        });\n    }\n\n    function findMissing() {\n      return featuresA\n        .map(({name}) => name)\n        .filter(v => pairs.some(({name}) => name === v) === false);\n    }\n  }\n};\n"],"file":"index.js"}

package/dist/match-detection/index.spec.js

@@ -12,9 +12,9 @@ var features = _interopRequireWildcard(require("./features"));
 
 var _ = _interopRequireDefault(require("."));
 
-function _getRequireWildcardCache() { if (typeof WeakMap !== "function") return null; var cache = new WeakMap(); _getRequireWildcardCache = function () { return cache; }; return cache; }
+function _getRequireWildcardCache(nodeInterop) { if (typeof WeakMap !== "function") return null; var cacheBabelInterop = new WeakMap(); var cacheNodeInterop = new WeakMap(); return (_getRequireWildcardCache = function (nodeInterop) { return nodeInterop ? cacheNodeInterop : cacheBabelInterop; })(nodeInterop); }
 
-function _interopRequireWildcard(obj) { if (obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
+function _interopRequireWildcard(obj, nodeInterop) { if (!nodeInterop && obj && obj.__esModule) { return obj; } if (obj === null || typeof obj !== "object" && typeof obj !== "function") { return { default: obj }; } var cache = _getRequireWildcardCache(nodeInterop); if (cache && cache.has(obj)) { return cache.get(obj); } var newObj = {}; var hasPropertyDescriptor = Object.defineProperty && Object.getOwnPropertyDescriptor; for (var key in obj) { if (key !== "default" && Object.prototype.hasOwnProperty.call(obj, key)) { var desc = hasPropertyDescriptor ? Object.getOwnPropertyDescriptor(obj, key) : null; if (desc && (desc.get || desc.set)) { Object.defineProperty(newObj, key, desc); } else { newObj[key] = obj[key]; } } } newObj.default = obj; if (cache) { cache.set(obj, newObj); } return newObj; }
 
 function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
 
@@ -45,30 +45,37 @@ function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { de
 * for the JavaScript code in this file.
 *
 */
-(0, _fixugen.default)({
-  path: [__dirname, '..', '..', 'test-fixtures', 'match-detection', 'index'],
-  useMetadataFile: true,
-  recurse: false,
-  fixura: {
-    reader: _fixura.READERS.JSON
-  },
-  callback: ({
-    getFixture,
-    options,
-    expectedResults
-  }) => {
-    const detect = (0, _.default)(formatOptions());
-    const recordA = new _marcRecord.MarcRecord(getFixture('recordA.json'));
-    const recordB = new _marcRecord.MarcRecord(getFixture('recordB.json'));
-    const results = detect(recordA, recordB);
-    (0, _chai.expect)(results).to.eql(expectedResults);
+describe('match-detection', () => {
+  (0, _fixugen.default)({
+    path: [__dirname, '..', '..', 'test-fixtures', 'match-detection', 'index'],
+    useMetadataFile: true,
+    recurse: false,
+    fixura: {
+      reader: _fixura.READERS.JSON
+    },
+    callback: ({
+      getFixture,
+      options,
+      expectedResults,
+      enabled = true
+    }) => {
+      if (!enabled) {
+        return;
+      }
 
-    function formatOptions() {
-      const contextFeatures = features[options.strategy.type];
-      return { ...options,
-        strategy: options.strategy.features.map(v => contextFeatures[v]())
-      };
+      const detect = (0, _.default)(formatOptions());
+      const recordA = new _marcRecord.MarcRecord(getFixture('recordA.json'));
+      const recordB = new _marcRecord.MarcRecord(getFixture('recordB.json'));
+      const results = detect(recordA, recordB);
+      (0, _chai.expect)(results).to.eql(expectedResults);
+
+      function formatOptions() {
+        const contextFeatures = features[options.strategy.type];
+        return { ...options,
+          strategy: options.strategy.features.map(v => contextFeatures[v]())
+        };
+      }
     }
-  }
+  });
 });
 //# sourceMappingURL=index.spec.js.map

package/dist/match-detection/index.spec.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../../src/match-detection/index.spec.js"],"names":["path","__dirname","useMetadataFile","recurse","fixura","reader","READERS","JSON","callback","getFixture","options","expectedResults","detect","formatOptions","recordA","MarcRecord","recordB","results","to","eql","contextFeatures","features","strategy","type","map","v"],"mappings":";;AA4BA;;AACA;;AACA;;AACA;;AACA;;AACA;;;;;;;;AAjCA;;;;;;;;;;;;;;;;;;;;;;;;;;;AAmCA,sBAAc;AACZA,EAAAA,IAAI,EAAE,CAACC,SAAD,EAAY,IAAZ,EAAkB,IAAlB,EAAwB,eAAxB,EAAyC,iBAAzC,EAA4D,OAA5D,CADM;AAEZC,EAAAA,eAAe,EAAE,IAFL;AAGZC,EAAAA,OAAO,EAAE,KAHG;AAIZC,EAAAA,MAAM,EAAE;AACNC,IAAAA,MAAM,EAAEC,gBAAQC;AADV,GAJI;AAOZC,EAAAA,QAAQ,EAAE,CAAC;AAACC,IAAAA,UAAD;AAAaC,IAAAA,OAAb;AAAsBC,IAAAA;AAAtB,GAAD,KAA4C;AACpD,UAAMC,MAAM,GAAG,eAAyBC,aAAa,EAAtC,CAAf;AACA,UAAMC,OAAO,GAAG,IAAIC,sBAAJ,CAAeN,UAAU,CAAC,cAAD,CAAzB,CAAhB;AACA,UAAMO,OAAO,GAAG,IAAID,sBAAJ,CAAeN,UAAU,CAAC,cAAD,CAAzB,CAAhB;AACA,UAAMQ,OAAO,GAAGL,MAAM,CAACE,OAAD,EAAUE,OAAV,CAAtB;AAEA,sBAAOC,OAAP,EAAgBC,EAAhB,CAAmBC,GAAnB,CAAuBR,eAAvB;;AAEA,aAASE,aAAT,GAAyB;AACvB,YAAMO,eAAe,GAAGC,QAAQ,CAACX,OAAO,CAACY,QAAR,CAAiBC,IAAlB,CAAhC;AAEA,aAAO,EACL,GAAGb,OADE;AAELY,QAAAA,QAAQ,EAAEZ,OAAO,CAACY,QAAR,CAAiBD,QAAjB,CAA0BG,GAA1B,CAA8BC,CAAC,IAAIL,eAAe,CAACK,CAAD,CAAf,EAAnC;AAFL,OAAP;AAID;AACF;AAvBW,CAAd","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport generateTests from '@natlibfi/fixugen';\nimport {READERS} from '@natlibfi/fixura';\nimport {expect} from 'chai';\nimport {MarcRecord} from '@natlibfi/marc-record';\nimport * as features from './features';\nimport createDetectionInterface from '.';\n\ngenerateTests({\n  path: [__dirname, '..', '..', 'test-fixtures', 'match-detection', 'index'],\n  useMetadataFile: true,\n  recurse: false,\n  fixura: {\n    reader: READERS.JSON\n  },\n  callback: ({getFixture, options, expectedResults}) => {\n    const detect = createDetectionInterface(formatOptions());\n    const recordA = new MarcRecord(getFixture('recordA.json'));\n    const recordB = new MarcRecord(getFixture('recordB.json'));\n    const results = detect(recordA, recordB);\n\n    expect(results).to.eql(expectedResults);\n\n    function formatOptions() {\n      const contextFeatures = features[options.strategy.type];\n\n      return {\n        ...options,\n        strategy: options.strategy.features.map(v => contextFeatures[v]())\n      };\n    }\n  }\n});\n"],"file":"index.spec.js"}
+{"version":3,"sources":["../../src/match-detection/index.spec.js"],"names":["describe","path","__dirname","useMetadataFile","recurse","fixura","reader","READERS","JSON","callback","getFixture","options","expectedResults","enabled","detect","formatOptions","recordA","MarcRecord","recordB","results","to","eql","contextFeatures","features","strategy","type","map","v"],"mappings":";;AA4BA;;AACA;;AACA;;AACA;;AACA;;AACA;;;;;;;;AAjCA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AASAA,QAAQ,CAAC,iBAAD,EAAoB,MAAM;AAChC,wBAAc;AACZC,IAAAA,IAAI,EAAE,CAACC,SAAD,EAAY,IAAZ,EAAkB,IAAlB,EAAwB,eAAxB,EAAyC,iBAAzC,EAA4D,OAA5D,CADM;AAEZC,IAAAA,eAAe,EAAE,IAFL;AAGZC,IAAAA,OAAO,EAAE,KAHG;AAIZC,IAAAA,MAAM,EAAE;AACNC,MAAAA,MAAM,EAAEC,gBAAQC;AADV,KAJI;AAOZC,IAAAA,QAAQ,EAAE,CAAC;AAACC,MAAAA,UAAD;AAAaC,MAAAA,OAAb;AAAsBC,MAAAA,eAAtB;AAAuCC,MAAAA,OAAO,GAAG;AAAjD,KAAD,KAA4D;AAEpE,UAAI,CAACA,OAAL,EAAc;AACZ;AACD;;AAED,YAAMC,MAAM,GAAG,eAAyBC,aAAa,EAAtC,CAAf;AACA,YAAMC,OAAO,GAAG,IAAIC,sBAAJ,CAAeP,UAAU,CAAC,cAAD,CAAzB,CAAhB;AACA,YAAMQ,OAAO,GAAG,IAAID,sBAAJ,CAAeP,UAAU,CAAC,cAAD,CAAzB,CAAhB;AACA,YAAMS,OAAO,GAAGL,MAAM,CAACE,OAAD,EAAUE,OAAV,CAAtB;AAEA,wBAAOC,OAAP,EAAgBC,EAAhB,CAAmBC,GAAnB,CAAuBT,eAAvB;;AAEA,eAASG,aAAT,GAAyB;AACvB,cAAMO,eAAe,GAAGC,QAAQ,CAACZ,OAAO,CAACa,QAAR,CAAiBC,IAAlB,CAAhC;AAEA,eAAO,EACL,GAAGd,OADE;AAELa,UAAAA,QAAQ,EAAEb,OAAO,CAACa,QAAR,CAAiBD,QAAjB,CAA0BG,GAA1B,CAA8BC,CAAC,IAAIL,eAAe,CAACK,CAAD,CAAf,EAAnC;AAFL,SAAP;AAID;AACF;AA5BW,GAAd;AA8BD,CA/BO,CAAR","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport generateTests from '@natlibfi/fixugen';\nimport {READERS} from '@natlibfi/fixura';\nimport {expect} from 'chai';\nimport {MarcRecord} from '@natlibfi/marc-record';\nimport * as features from './features';\nimport createDetectionInterface from '.';\n\ndescribe('match-detection', () => {\n  generateTests({\n    path: [__dirname, '..', '..', 'test-fixtures', 'match-detection', 'index'],\n    useMetadataFile: true,\n    recurse: false,\n    fixura: {\n      reader: READERS.JSON\n    },\n    callback: ({getFixture, options, expectedResults, enabled = true}) => {\n\n      if (!enabled) {\n        return;\n      }\n\n      const detect = createDetectionInterface(formatOptions());\n      const recordA = new MarcRecord(getFixture('recordA.json'));\n      const recordB = new MarcRecord(getFixture('recordB.json'));\n      const results = detect(recordA, recordB);\n\n      expect(results).to.eql(expectedResults);\n\n      function formatOptions() {\n        const contextFeatures = features[options.strategy.type];\n\n        return {\n          ...options,\n          strategy: options.strategy.features.map(v => contextFeatures[v]())\n        };\n      }\n    }\n  });\n});\n"],"file":"index.spec.js"}

package/dist/matching-utils.js

@@ -0,0 +1,95 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.getMelindaIdsF035 = getMelindaIdsF035;
+exports.getSubfieldValues = getSubfieldValues;
+exports.validateSidFieldSubfieldCounts = validateSidFieldSubfieldCounts;
+
+var _debug = _interopRequireDefault(require("debug"));
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+/**
+*
+* @licstart  The following is the entire license notice for the JavaScript code in this file.
+*
+* Melinda record matching modules for Javascript
+*
+* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)
+*
+* This file is part of melinda-record-matching-js
+*
+* melinda-record-matching-js program is free software: you can redistribute it and/or modify
+* it under the terms of the GNU Lesser General Public License as
+* published by the Free Software Foundation, either version 3 of the
+* License, or (at your option) any later version.
+*
+* melinda-record-matching-js is distributed in the hope that it will be useful,
+* but WITHOUT ANY WARRANTY; without even the implied warranty of
+* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+* GNU Lesser General Public License for more details.
+*
+* You should have received a copy of the GNU Affero General Public License
+* along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*
+* @licend  The above is the entire license notice
+* for the JavaScript code in this file.
+*
+*/
+const debug = (0, _debug.default)('@natlibfi/melinda-record-matching:utils');
+const debugData = debug.extend('data'); // eslint-disable-next-line max-statements
+
+function getMelindaIdsF035(record) {
+  const debug = (0, _debug.default)('@natlibfi/melinda-record-matching:melinda-id');
+  const debugData = debug.extend('data');
+  const f035s = record.getFields('035');
+
+  if (f035s.length < 1) {
+    debug(`No f035s found.`);
+    return [];
+  }
+
+  const allF035MelindaIds = [].concat(...f035s.map(toMelindaIds));
+  const f035MelindaIds = [...new Set(allF035MelindaIds)];
+  debugData(`Fields (${f035s.length}): ${JSON.stringify(f035s)}`);
+  debugData(`Ids (${allF035MelindaIds.length}): ${JSON.stringify(allF035MelindaIds)}`);
+  debugData(`Unique ids (${f035MelindaIds.length}): ${JSON.stringify(f035MelindaIds)}`);
+  return f035MelindaIds;
+
+  function toMelindaIds({
+    subfields
+  }) {
+    const melindaIdRegExp = /^(?<prefix>\(FI-MELINDA\)|FCC)(?<id>\d{9})$/u;
+    return subfields.filter(sub => ['a', 'z'].includes(sub.code)).filter(sub => melindaIdRegExp.test(sub.value)).map(({
+      value
+    }) => value.replace(melindaIdRegExp, '$<id>'));
+  }
+}
+
+function validateSidFieldSubfieldCounts(field) {
+  // Valid SID-fields have just one $c and one $b
+  debugData(`Validating SID field ${JSON.stringify(field)}`);
+  const countC = countSubfields(field, 'c');
+  const countB = countSubfields(field, 'b');
+  debug(`Found ${countC} sf $cs and ${countB} sf $bs. IsValid: ${countC === 1 && countB === 1}`);
+  return countC === 1 && countB === 1;
+}
+
+function countSubfields(field, subfieldCode) {
+  // debug(`Counting subfields ${subfieldCode}`);
+  return field.subfields.filter(({
+    code
+  }) => code === subfieldCode).length;
+}
+
+function getSubfieldValues(field, subfieldCode) {
+  debugData(`Get subfield(s) $${subfieldCode} from ${JSON.stringify(field)}`);
+  return field.subfields.filter(({
+    code
+  }) => code === subfieldCode).map(({
+    value
+  }) => value);
+}
+//# sourceMappingURL=matching-utils.js.map

package/dist/matching-utils.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/matching-utils.js"],"names":["debug","debugData","extend","getMelindaIdsF035","record","f035s","getFields","length","allF035MelindaIds","concat","map","toMelindaIds","f035MelindaIds","Set","JSON","stringify","subfields","melindaIdRegExp","filter","sub","includes","code","test","value","replace","validateSidFieldSubfieldCounts","field","countC","countSubfields","countB","subfieldCode","getSubfieldValues"],"mappings":";;;;;;;;;AA4BA;;;;AA5BA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AAIA,MAAMA,KAAK,GAAG,oBAAkB,yCAAlB,CAAd;AACA,MAAMC,SAAS,GAAGD,KAAK,CAACE,MAAN,CAAa,MAAb,CAAlB,C,CAEA;;AACO,SAASC,iBAAT,CAA2BC,MAA3B,EAAmC;AAExC,QAAMJ,KAAK,GAAG,oBAAkB,8CAAlB,CAAd;AACA,QAAMC,SAAS,GAAGD,KAAK,CAACE,MAAN,CAAa,MAAb,CAAlB;AAEA,QAAMG,KAAK,GAAGD,MAAM,CAACE,SAAP,CAAiB,KAAjB,CAAd;;AAEA,MAAID,KAAK,CAACE,MAAN,GAAe,CAAnB,EAAsB;AACpBP,IAAAA,KAAK,CAAE,iBAAF,CAAL;AACA,WAAO,EAAP;AACD;;AAED,QAAMQ,iBAAiB,GAAG,GAAGC,MAAH,CAAU,GAAGJ,KAAK,CAACK,GAAN,CAAUC,YAAV,CAAb,CAA1B;AACA,QAAMC,cAAc,GAAG,CAAC,GAAG,IAAIC,GAAJ,CAAQL,iBAAR,CAAJ,CAAvB;AAEAP,EAAAA,SAAS,CAAE,WAAUI,KAAK,CAACE,MAAO,MAAKO,IAAI,CAACC,SAAL,CAAeV,KAAf,CAAsB,EAApD,CAAT;AACAJ,EAAAA,SAAS,CAAE,QAAOO,iBAAiB,CAACD,MAAO,MAAKO,IAAI,CAACC,SAAL,CAAeP,iBAAf,CAAkC,EAAzE,CAAT;AACAP,EAAAA,SAAS,CAAE,eAAcW,cAAc,CAACL,MAAO,MAAKO,IAAI,CAACC,SAAL,CAAeH,cAAf,CAA+B,EAA1E,CAAT;AAEA,SAAOA,cAAP;;AAEA,WAASD,YAAT,CAAsB;AAACK,IAAAA;AAAD,GAAtB,EAAmC;AACjC,UAAMC,eAAe,GAAG,8CAAxB;AAEA,WAAOD,SAAS,CACbE,MADI,CACGC,GAAG,IAAI,CAAC,GAAD,EAAM,GAAN,EAAWC,QAAX,CAAoBD,GAAG,CAACE,IAAxB,CADV,EAEJH,MAFI,CAEGC,GAAG,IAAIF,eAAe,CAACK,IAAhB,CAAqBH,GAAG,CAACI,KAAzB,CAFV,EAGJb,GAHI,CAGA,CAAC;AAACa,MAAAA;AAAD,KAAD,KAAaA,KAAK,CAACC,OAAN,CAAcP,eAAd,EAA+B,OAA/B,CAHb,CAAP;AAKD;AACF;;AAEM,SAASQ,8BAAT,CAAwCC,KAAxC,EAA+C;AACpD;AACAzB,EAAAA,SAAS,CAAE,wBAAuBa,IAAI,CAACC,SAAL,CAAeW,KAAf,CAAsB,EAA/C,CAAT;AACA,QAAMC,MAAM,GAAGC,cAAc,CAACF,KAAD,EAAQ,GAAR,CAA7B;AACA,QAAMG,MAAM,GAAGD,cAAc,CAACF,KAAD,EAAQ,GAAR,CAA7B;AACA1B,EAAAA,KAAK,CAAE,SAAQ2B,MAAO,eAAcE,MAAO,qBAAoBF,MAAM,KAAK,CAAX,IAAgBE,MAAM,KAAK,CAAE,EAAvF,CAAL;AAEA,SAAOF,MAAM,KAAK,CAAX,IAAgBE,MAAM,KAAK,CAAlC;AACD;;AAED,SAASD,cAAT,CAAwBF,KAAxB,EAA+BI,YAA/B,EAA6C;AAC3C;AACA,SAAOJ,KAAK,CAACV,SAAN,CAAgBE,MAAhB,CAAuB,CAAC;AAACG,IAAAA;AAAD,GAAD,KAAYA,IAAI,KAAKS,YAA5C,EAA0DvB,MAAjE;AACD;;AAEM,SAASwB,iBAAT,CAA2BL,KAA3B,EAAkCI,YAAlC,EAAgD;AACrD7B,EAAAA,SAAS,CAAE,oBAAmB6B,YAAa,SAAQhB,IAAI,CAACC,SAAL,CAAeW,KAAf,CAAsB,EAAhE,CAAT;AACA,SAAOA,KAAK,CAACV,SAAN,CAAgBE,MAAhB,CAAuB,CAAC;AAACG,IAAAA;AAAD,GAAD,KAAYA,IAAI,KAAKS,YAA5C,EAA0DpB,GAA1D,CAA8D,CAAC;AAACa,IAAAA;AAAD,GAAD,KAAaA,KAA3E,CAAP;AACD","sourcesContent":["/**\n*\n* @licstart  The following is the entire license notice for the JavaScript code in this file.\n*\n* Melinda record matching modules for Javascript\n*\n* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)\n*\n* This file is part of melinda-record-matching-js\n*\n* melinda-record-matching-js program is free software: you can redistribute it and/or modify\n* it under the terms of the GNU Lesser General Public License as\n* published by the Free Software Foundation, either version 3 of the\n* License, or (at your option) any later version.\n*\n* melinda-record-matching-js is distributed in the hope that it will be useful,\n* but WITHOUT ANY WARRANTY; without even the implied warranty of\n* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\n* GNU Lesser General Public License for more details.\n*\n* You should have received a copy of the GNU Affero General Public License\n* along with this program.  If not, see <http://www.gnu.org/licenses/>.\n*\n* @licend  The above is the entire license notice\n* for the JavaScript code in this file.\n*\n*/\n\nimport createDebugLogger from 'debug';\n\nconst debug = createDebugLogger('@natlibfi/melinda-record-matching:utils');\nconst debugData = debug.extend('data');\n\n// eslint-disable-next-line max-statements\nexport function getMelindaIdsF035(record) {\n\n  const debug = createDebugLogger('@natlibfi/melinda-record-matching:melinda-id');\n  const debugData = debug.extend('data');\n\n  const f035s = record.getFields('035');\n\n  if (f035s.length < 1) {\n    debug(`No f035s found.`);\n    return [];\n  }\n\n  const allF035MelindaIds = [].concat(...f035s.map(toMelindaIds));\n  const f035MelindaIds = [...new Set(allF035MelindaIds)];\n\n  debugData(`Fields (${f035s.length}): ${JSON.stringify(f035s)}`);\n  debugData(`Ids (${allF035MelindaIds.length}): ${JSON.stringify(allF035MelindaIds)}`);\n  debugData(`Unique ids (${f035MelindaIds.length}): ${JSON.stringify(f035MelindaIds)}`);\n\n  return f035MelindaIds;\n\n  function toMelindaIds({subfields}) {\n    const melindaIdRegExp = /^(?<prefix>\\(FI-MELINDA\\)|FCC)(?<id>\\d{9})$/u;\n\n    return subfields\n      .filter(sub => ['a', 'z'].includes(sub.code))\n      .filter(sub => melindaIdRegExp.test(sub.value))\n      .map(({value}) => value.replace(melindaIdRegExp, '$<id>'));\n\n  }\n}\n\nexport function validateSidFieldSubfieldCounts(field) {\n  // Valid SID-fields have just one $c and one $b\n  debugData(`Validating SID field ${JSON.stringify(field)}`);\n  const countC = countSubfields(field, 'c');\n  const countB = countSubfields(field, 'b');\n  debug(`Found ${countC} sf $cs and ${countB} sf $bs. IsValid: ${countC === 1 && countB === 1}`);\n\n  return countC === 1 && countB === 1;\n}\n\nfunction countSubfields(field, subfieldCode) {\n  // debug(`Counting subfields ${subfieldCode}`);\n  return field.subfields.filter(({code}) => code === subfieldCode).length;\n}\n\nexport function getSubfieldValues(field, subfieldCode) {\n  debugData(`Get subfield(s) $${subfieldCode} from ${JSON.stringify(field)}`);\n  return field.subfields.filter(({code}) => code === subfieldCode).map(({value}) => value);\n}\n\n"],"file":"matching-utils.js"}

package/package.json

@@ -14,10 +14,10 @@
 		"url": "git@github.com:natlibfi/melinda-record-matching-js.git"
 	},
 	"license": "LGPL-3.0+",
-	"version": "1.0.8",
+	"version": "2.0.0",
 	"main": "./dist/index.js",
 	"engines": {
-		"node": ">=12"
+		"node": ">=14"
 	},
 	"publishConfig": {
 		"access": "public"
@@ -35,35 +35,34 @@
 		"watch:test": "cross-env DEBUG=1 NODE_ENV=test nodemon -w src -w test-fixtures --exec 'npm run test:dev'"
 	},
 	"dependencies": {
-		"@natlibfi/marc-record": "^6.1.1",
-		"@natlibfi/marc-record-serializers": "^7.1.6",
-		"@natlibfi/sru-client": "^4.0.0",
-		"debug": "^4.1.1",
-		"moment": "^2.28.0",
-		"mongodb": "^3.6.2",
-		"natural": "^2.1.5",
-		"uuid": "^8.3.0",
+		"@natlibfi/marc-record": "^7.0.0-alpha.1",
+		"@natlibfi/marc-record-serializers": "^8.0.4",
+		"@natlibfi/sru-client": "^4.1.0",
+		"debug": "^4.3.2",
+		"moment": "^2.29.1",
+		"natural": "^5.1.11",
+		"uuid": "^8.3.2",
 		"winston": "^3.3.3"
 	},
 	"devDependencies": {
-		"@babel/cli": "^7.11.6",
-		"@babel/core": "^7.11.6",
-		"@babel/node": "^7.10.5",
-		"@babel/preset-env": "^7.11.5",
-		"@babel/register": "^7.11.5",
-		"@natlibfi/eslint-config-melinda-backend": "^1.0.5",
-		"@natlibfi/fixugen": "^1.0.0",
-		"@natlibfi/fixugen-http-client": "^1.1.1",
-		"@natlibfi/fixura": "^2.1.3",
-		"babel-eslint": "^10.1.0",
-		"babel-plugin-istanbul": "^6.0.0",
+		"@babel/cli": "^7.16.0",
+		"@babel/core": "^7.16.0",
+		"@babel/node": "^7.16.0",
+		"@babel/preset-env": "^7.16.4",
+		"@babel/register": "^7.16.0",
+		"@babel/eslint-parser": "^7.16.3",
+		"@natlibfi/eslint-config-melinda-backend": "^1.1.1",
+		"@natlibfi/fixugen": "^1.0.2",
+		"@natlibfi/fixugen-http-client": "^1.1.3",
+		"@natlibfi/fixura": "^2.2.1",
+		"babel-plugin-istanbul": "^6.1.1",
 		"babel-plugin-rewire": "^1.2.0",
-		"chai": "^4.2.0",
+		"chai": "^4.3.4",
 		"chai-as-promised": "^7.1.1",
-		"cross-env": "^7.0.2",
-		"eslint": "^7.9.0",
-		"mocha": "^8.1.3",
-		"nodemon": "^2.0.4",
+		"cross-env": "^7.0.3",
+		"eslint": "^7.32.0",
+		"mocha": "^9.1.3",
+		"nodemon": "^2.0.15",
 		"nyc": "^15.1.0"
 	},
 	"eslintConfig": {
@@ -74,9 +73,7 @@
 			[
 				"@babel/preset-env",
 				{
-					"targets": {
-						"node": "12"
-					}
+					"targets": "maintained node versions"
 				}
 			]
 		],

package/src/candidate-search/candidate-search-utils.js

@@ -0,0 +1,53 @@
+/**
+*
+* @licstart  The following is the entire license notice for the JavaScript code in this file.
+*
+* Melinda record matching modules for Javascript
+*
+* Copyright (C) 2020 University Of Helsinki (The National Library Of Finland)
+*
+* This file is part of melinda-record-matching-js
+*
+* melinda-record-matching-js program is free software: you can redistribute it and/or modify
+* it under the terms of the GNU Lesser General Public License as
+* published by the Free Software Foundation, either version 3 of the
+* License, or (at your option) any later version.
+*
+* melinda-record-matching-js is distributed in the hope that it will be useful,
+* but WITHOUT ANY WARRANTY; without even the implied warranty of
+* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+* GNU Lesser General Public License for more details.
+*
+* You should have received a copy of the GNU Affero General Public License
+* along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*
+* @licend  The above is the entire license notice
+* for the JavaScript code in this file.
+*
+*/
+
+import createDebugLogger from 'debug';
+
+export function toQueries(identifiers, queryString) {
+
+  const debug = createDebugLogger('@natlibfi/melinda-record-matching:toQueries');
+  const debugData = debug.extend('data');
+
+  const quotedIdentifiers = identifiers.map(identifier => identifier.match(/\//u) ? `"${identifier}"` : `${identifier}`);
+
+  // Aleph supports only two queries with or -operator (This is not actually true)
+  const pairs = toPairs(quotedIdentifiers);
+  const queries = pairs.map(([a, b]) => b ? `${queryString}=${a} or ${queryString}=${b}` : `${queryString}=${a}`);
+
+  debugData(`Pairs (${pairs.length}): ${JSON.stringify(pairs)}`);
+  debugData(`Queries (${queries.length}): ${JSON.stringify(queries)}`);
+
+  return queries;
+}
+
+function toPairs(array) {
+  if (array.length === 0) {
+    return [];
+  }
+  return [array.slice(0, 2)].concat(toPairs(array.slice(2), 2));
+}

package/src/candidate-search/index.js

@@ -48,6 +48,7 @@ export default ({record, searchSpec, url, maxRecordsPerRequest = 50}) => {
     retrieveAll: false
   });
 
+  debug(`Searching matches for ${inputRecordId}`);
   debug(`Generated queryList ${JSON.stringify(queryList)}`);
   if (queryList.length === 0) { // eslint-disable-line functional/no-conditional-statement
     throw new CandidateSearchError(`Generated query list contains no queries`);
@@ -101,6 +102,9 @@ export default ({record, searchSpec, url, maxRecordsPerRequest = 50}) => {
                 const foundRecordMarc = await MARCXML.from(foundRecord, {subfieldValues: false});
                 const foundRecordId = getRecordId(foundRecordMarc);
 
+                // This does not work and might cause problems:
+                // Record *should* match itself AND in REST the input record is given id 000000001 always
+                debug(`Checking ${inputRecordId} vs ${foundRecordId}`);
                 if (inputRecordId === foundRecordId) {
                   debug(`Input and candidate are the same record per 001. Discarding candidate`);
                   return;