@faircopy/rules-nlp 1.1.1

package/dist/index.d.ts

@@ -0,0 +1,16 @@
+import { Rule } from '@faircopy/core';
+
+interface NoFilterWordsOptions {
+    phrases?: string[];
+}
+declare const noFilterWords: Rule<NoFilterWordsOptions>;
+
+interface NoPassiveVoiceOptions {
+    allowedAuxiliaries?: string[];
+}
+declare const noPassiveVoice: Rule<NoPassiveVoiceOptions>;
+
+/** All NLP rules keyed by their rule ID. */
+declare const ruleRegistry: Map<string, Rule>;
+
+export { type NoFilterWordsOptions, type NoPassiveVoiceOptions, noFilterWords, noPassiveVoice, ruleRegistry };

package/dist/index.js

@@ -0,0 +1,117 @@
+// src/utils.ts
+import nlp from "compromise";
+function createDoc(text) {
+  return nlp(text);
+}
+function getMatchOccurrences(text, matches) {
+  const json = matches.json({ offset: true, text: true, terms: { offset: true } });
+  return json.flatMap((entry) => {
+    const start = entry.offset?.start ?? entry.terms?.[0]?.offset?.start;
+    const length = entry.offset?.length ?? sumTermLengths(entry.terms);
+    if (typeof start !== "number" || typeof length !== "number" || length <= 0) {
+      return [];
+    }
+    return [{
+      text: entry.text ?? text.slice(start, start + length),
+      start,
+      end: start + length
+    }];
+  });
+}
+function sumTermLengths(terms) {
+  if (!terms?.length) return void 0;
+  let total = 0;
+  for (const term of terms) {
+    const length = term.offset?.length;
+    if (typeof length !== "number") return void 0;
+    total += length;
+  }
+  return total;
+}
+
+// src/no-filter-words.ts
+var DEFAULT_PHRASES = [
+  "I think",
+  "it seems",
+  "basically",
+  "in order to"
+];
+var noFilterWords = {
+  id: "no-filter-words",
+  description: "Ban filter phrases that distance the claim from the reader",
+  defaults: { phrases: DEFAULT_PHRASES },
+  help: "Filter phrases announce a perspective or pad the sentence instead of making the point. Delete the phrase or rewrite the sentence so the claim stands on its own.",
+  check({ text, sourceMap, options }) {
+    const diagnostics = [];
+    const phrases = options.phrases?.length ? options.phrases : DEFAULT_PHRASES;
+    const doc = createDoc(text);
+    for (const phrase of phrases) {
+      const matches = doc.match(phrase);
+      for (const occurrence of getMatchOccurrences(text, matches)) {
+        const start = sourceMap[occurrence.start];
+        const end = sourceMap[occurrence.end - 1];
+        if (start === void 0 || end === void 0) continue;
+        diagnostics.push({
+          ruleId: "no-filter-words",
+          severity: "error",
+          message: `remove "${occurrence.text.toLowerCase()}" \u2014 state the claim directly`,
+          range: { start, end: end + 1 },
+          help: noFilterWords.help
+        });
+      }
+    }
+    return diagnostics;
+  }
+};
+
+// src/no-passive-voice.ts
+var DEFAULT_ALLOWED_AUXILIARIES = ["is", "are", "was", "were", "be", "been", "being"];
+var noPassiveVoice = {
+  id: "no-passive-voice",
+  description: "Flag likely passive-voice constructions using POS tagging patterns",
+  defaults: { allowedAuxiliaries: DEFAULT_ALLOWED_AUXILIARIES },
+  help: "Passive voice often hides the actor and adds drag. Prefer naming who did the action unless the actor genuinely does not matter.",
+  check({ text, sourceMap, options }) {
+    const diagnostics = [];
+    const auxiliaries = new Set((options.allowedAuxiliaries?.length ? options.allowedAuxiliaries : DEFAULT_ALLOWED_AUXILIARIES).map((value) => value.toLowerCase()));
+    const doc = createDoc(text);
+    const matches = doc.match("(#Copula|#Auxiliary) #PastTense");
+    for (const occurrence of getMatchOccurrences(text, matches)) {
+      const words = occurrence.text.trim().split(/\s+/);
+      if (words.length < 2) continue;
+      if (!auxiliaries.has(words[0].toLowerCase())) continue;
+      const start = sourceMap[occurrence.start];
+      const end = sourceMap[occurrence.end - 1];
+      if (start === void 0 || end === void 0) continue;
+      diagnostics.push({
+        ruleId: "no-passive-voice",
+        severity: "warn",
+        message: `rewrite passive construction "${occurrence.text}" with a named actor`,
+        range: { start, end: end + 1 },
+        help: noPassiveVoice.help
+      });
+    }
+    return dedupeDiagnostics(diagnostics);
+  }
+};
+function dedupeDiagnostics(diagnostics) {
+  const seen = /* @__PURE__ */ new Set();
+  return diagnostics.filter((diagnostic) => {
+    const key = `${diagnostic.range.start}:${diagnostic.range.end}`;
+    if (seen.has(key)) return false;
+    seen.add(key);
+    return true;
+  });
+}
+
+// src/index.ts
+var ruleRegistry = /* @__PURE__ */ new Map([
+  ["no-filter-words", noFilterWords],
+  ["no-passive-voice", noPassiveVoice]
+]);
+export {
+  noFilterWords,
+  noPassiveVoice,
+  ruleRegistry
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/utils.ts","../src/no-filter-words.ts","../src/no-passive-voice.ts","../src/index.ts"],"sourcesContent":["import nlp from 'compromise'\nimport type { DocView, JsonOffsetEntry, JsonOffsetTerm, MatchView } from './types.js'\n\nexport interface MatchOccurrence {\n  text: string\n  start: number\n  end: number\n}\n\nexport function createDoc(text: string): DocView {\n  return nlp(text) as unknown as DocView\n}\n\nexport function getMatchOccurrences(text: string, matches: MatchView): MatchOccurrence[] {\n  const json = matches.json({ offset: true, text: true, terms: { offset: true } }) as JsonOffsetEntry[]\n\n  return json.flatMap((entry) => {\n    const start = entry.offset?.start ?? entry.terms?.[0]?.offset?.start\n    const length = entry.offset?.length ?? sumTermLengths(entry.terms)\n\n    if (typeof start !== 'number' || typeof length !== 'number' || length <= 0) {\n      return []\n    }\n\n    return [{\n      text: entry.text ?? text.slice(start, start + length),\n      start,\n      end: start + length,\n    }]\n  })\n}\n\nfunction sumTermLengths(terms: JsonOffsetTerm[] | undefined): number | undefined {\n  if (!terms?.length) return undefined\n\n  let total = 0\n  for (const term of terms) {\n    const length = term.offset?.length\n    if (typeof length !== 'number') return undefined\n    total += length\n  }\n\n  return total\n}\n","import type { Diagnostic, Rule, RuleInput } from '@faircopy/core'\nimport { createDoc, getMatchOccurrences } from './utils.js'\n\nexport interface NoFilterWordsOptions {\n  phrases?: string[]\n}\n\nconst DEFAULT_PHRASES = [\n  'I think',\n  'it seems',\n  'basically',\n  'in order to',\n]\n\nexport const noFilterWords: Rule<NoFilterWordsOptions> = {\n  id: 'no-filter-words',\n  description: 'Ban filter phrases that distance the claim from the reader',\n  defaults: { phrases: DEFAULT_PHRASES },\n  help: 'Filter phrases announce a perspective or pad the sentence instead of making the point. Delete the phrase or rewrite the sentence so the claim stands on its own.',\n\n  check({ text, sourceMap, options }: RuleInput<NoFilterWordsOptions>): Diagnostic[] {\n    const diagnostics: Diagnostic[] = []\n    const phrases = options.phrases?.length ? options.phrases : DEFAULT_PHRASES\n    const doc = createDoc(text)\n\n    for (const phrase of phrases) {\n      const matches = doc.match(phrase)\n      for (const occurrence of getMatchOccurrences(text, matches)) {\n        const start = sourceMap[occurrence.start]\n        const end = sourceMap[occurrence.end - 1]\n        if (start === undefined || end === undefined) continue\n\n        diagnostics.push({\n          ruleId: 'no-filter-words',\n          severity: 'error',\n          message: `remove \"${occurrence.text.toLowerCase()}\" — state the claim directly`,\n          range: { start, end: end + 1 },\n          help: noFilterWords.help,\n        })\n      }\n    }\n\n    return diagnostics\n  },\n}\n","import type { Diagnostic, Rule, RuleInput } from '@faircopy/core'\nimport { createDoc, getMatchOccurrences } from './utils.js'\n\nexport interface NoPassiveVoiceOptions {\n  allowedAuxiliaries?: string[]\n}\n\nconst DEFAULT_ALLOWED_AUXILIARIES = ['is', 'are', 'was', 'were', 'be', 'been', 'being']\n\nexport const noPassiveVoice: Rule<NoPassiveVoiceOptions> = {\n  id: 'no-passive-voice',\n  description: 'Flag likely passive-voice constructions using POS tagging patterns',\n  defaults: { allowedAuxiliaries: DEFAULT_ALLOWED_AUXILIARIES },\n  help: 'Passive voice often hides the actor and adds drag. Prefer naming who did the action unless the actor genuinely does not matter.',\n\n  check({ text, sourceMap, options }: RuleInput<NoPassiveVoiceOptions>): Diagnostic[] {\n    const diagnostics: Diagnostic[] = []\n    const auxiliaries = new Set((options.allowedAuxiliaries?.length ? options.allowedAuxiliaries : DEFAULT_ALLOWED_AUXILIARIES).map(value => value.toLowerCase()))\n    const doc = createDoc(text)\n    const matches = doc.match('(#Copula|#Auxiliary) #PastTense')\n\n    for (const occurrence of getMatchOccurrences(text, matches)) {\n      const words = occurrence.text.trim().split(/\\s+/)\n      if (words.length < 2) continue\n      if (!auxiliaries.has(words[0]!.toLowerCase())) continue\n\n      const start = sourceMap[occurrence.start]\n      const end = sourceMap[occurrence.end - 1]\n      if (start === undefined || end === undefined) continue\n\n      diagnostics.push({\n        ruleId: 'no-passive-voice',\n        severity: 'warn',\n        message: `rewrite passive construction \"${occurrence.text}\" with a named actor`,\n        range: { start, end: end + 1 },\n        help: noPassiveVoice.help,\n      })\n    }\n\n    return dedupeDiagnostics(diagnostics)\n  },\n}\n\nfunction dedupeDiagnostics(diagnostics: Diagnostic[]): Diagnostic[] {\n  const seen = new Set<string>()\n  return diagnostics.filter((diagnostic) => {\n    const key = `${diagnostic.range.start}:${diagnostic.range.end}`\n    if (seen.has(key)) return false\n    seen.add(key)\n    return true\n  })\n}\n","import type { Rule } from '@faircopy/core'\nimport { noFilterWords } from './no-filter-words.js'\nimport { noPassiveVoice } from './no-passive-voice.js'\n\nexport { noFilterWords } from './no-filter-words.js'\nexport { noPassiveVoice } from './no-passive-voice.js'\nexport type { NoFilterWordsOptions } from './no-filter-words.js'\nexport type { NoPassiveVoiceOptions } from './no-passive-voice.js'\n\n/** All NLP rules keyed by their rule ID. */\nexport const ruleRegistry: Map<string, Rule> = new Map([\n  ['no-filter-words', noFilterWords as Rule],\n  ['no-passive-voice', noPassiveVoice as Rule],\n])\n"],"mappings":";AAAA,OAAO,SAAS;AAST,SAAS,UAAU,MAAuB;AAC/C,SAAO,IAAI,IAAI;AACjB;AAEO,SAAS,oBAAoB,MAAc,SAAuC;AACvF,QAAM,OAAO,QAAQ,KAAK,EAAE,QAAQ,MAAM,MAAM,MAAM,OAAO,EAAE,QAAQ,KAAK,EAAE,CAAC;AAE/E,SAAO,KAAK,QAAQ,CAAC,UAAU;AAC7B,UAAM,QAAQ,MAAM,QAAQ,SAAS,MAAM,QAAQ,CAAC,GAAG,QAAQ;AAC/D,UAAM,SAAS,MAAM,QAAQ,UAAU,eAAe,MAAM,KAAK;AAEjE,QAAI,OAAO,UAAU,YAAY,OAAO,WAAW,YAAY,UAAU,GAAG;AAC1E,aAAO,CAAC;AAAA,IACV;AAEA,WAAO,CAAC;AAAA,MACN,MAAM,MAAM,QAAQ,KAAK,MAAM,OAAO,QAAQ,MAAM;AAAA,MACpD;AAAA,MACA,KAAK,QAAQ;AAAA,IACf,CAAC;AAAA,EACH,CAAC;AACH;AAEA,SAAS,eAAe,OAAyD;AAC/E,MAAI,CAAC,OAAO,OAAQ,QAAO;AAE3B,MAAI,QAAQ;AACZ,aAAW,QAAQ,OAAO;AACxB,UAAM,SAAS,KAAK,QAAQ;AAC5B,QAAI,OAAO,WAAW,SAAU,QAAO;AACvC,aAAS;AAAA,EACX;AAEA,SAAO;AACT;;;ACpCA,IAAM,kBAAkB;AAAA,EACtB;AAAA,EACA;AAAA,EACA;AAAA,EACA;AACF;AAEO,IAAM,gBAA4C;AAAA,EACvD,IAAI;AAAA,EACJ,aAAa;AAAA,EACb,UAAU,EAAE,SAAS,gBAAgB;AAAA,EACrC,MAAM;AAAA,EAEN,MAAM,EAAE,MAAM,WAAW,QAAQ,GAAkD;AACjF,UAAM,cAA4B,CAAC;AACnC,UAAM,UAAU,QAAQ,SAAS,SAAS,QAAQ,UAAU;AAC5D,UAAM,MAAM,UAAU,IAAI;AAE1B,eAAW,UAAU,SAAS;AAC5B,YAAM,UAAU,IAAI,MAAM,MAAM;AAChC,iBAAW,cAAc,oBAAoB,MAAM,OAAO,GAAG;AAC3D,cAAM,QAAQ,UAAU,WAAW,KAAK;AACxC,cAAM,MAAM,UAAU,WAAW,MAAM,CAAC;AACxC,YAAI,UAAU,UAAa,QAAQ,OAAW;AAE9C,oBAAY,KAAK;AAAA,UACf,QAAQ;AAAA,UACR,UAAU;AAAA,UACV,SAAS,WAAW,WAAW,KAAK,YAAY,CAAC;AAAA,UACjD,OAAO,EAAE,OAAO,KAAK,MAAM,EAAE;AAAA,UAC7B,MAAM,cAAc;AAAA,QACtB,CAAC;AAAA,MACH;AAAA,IACF;AAEA,WAAO;AAAA,EACT;AACF;;;ACrCA,IAAM,8BAA8B,CAAC,MAAM,OAAO,OAAO,QAAQ,MAAM,QAAQ,OAAO;AAE/E,IAAM,iBAA8C;AAAA,EACzD,IAAI;AAAA,EACJ,aAAa;AAAA,EACb,UAAU,EAAE,oBAAoB,4BAA4B;AAAA,EAC5D,MAAM;AAAA,EAEN,MAAM,EAAE,MAAM,WAAW,QAAQ,GAAmD;AAClF,UAAM,cAA4B,CAAC;AACnC,UAAM,cAAc,IAAI,KAAK,QAAQ,oBAAoB,SAAS,QAAQ,qBAAqB,6BAA6B,IAAI,WAAS,MAAM,YAAY,CAAC,CAAC;AAC7J,UAAM,MAAM,UAAU,IAAI;AAC1B,UAAM,UAAU,IAAI,MAAM,iCAAiC;AAE3D,eAAW,cAAc,oBAAoB,MAAM,OAAO,GAAG;AAC3D,YAAM,QAAQ,WAAW,KAAK,KAAK,EAAE,MAAM,KAAK;AAChD,UAAI,MAAM,SAAS,EAAG;AACtB,UAAI,CAAC,YAAY,IAAI,MAAM,CAAC,EAAG,YAAY,CAAC,EAAG;AAE/C,YAAM,QAAQ,UAAU,WAAW,KAAK;AACxC,YAAM,MAAM,UAAU,WAAW,MAAM,CAAC;AACxC,UAAI,UAAU,UAAa,QAAQ,OAAW;AAE9C,kBAAY,KAAK;AAAA,QACf,QAAQ;AAAA,QACR,UAAU;AAAA,QACV,SAAS,iCAAiC,WAAW,IAAI;AAAA,QACzD,OAAO,EAAE,OAAO,KAAK,MAAM,EAAE;AAAA,QAC7B,MAAM,eAAe;AAAA,MACvB,CAAC;AAAA,IACH;AAEA,WAAO,kBAAkB,WAAW;AAAA,EACtC;AACF;AAEA,SAAS,kBAAkB,aAAyC;AAClE,QAAM,OAAO,oBAAI,IAAY;AAC7B,SAAO,YAAY,OAAO,CAAC,eAAe;AACxC,UAAM,MAAM,GAAG,WAAW,MAAM,KAAK,IAAI,WAAW,MAAM,GAAG;AAC7D,QAAI,KAAK,IAAI,GAAG,EAAG,QAAO;AAC1B,SAAK,IAAI,GAAG;AACZ,WAAO;AAAA,EACT,CAAC;AACH;;;ACzCO,IAAM,eAAkC,oBAAI,IAAI;AAAA,EACrD,CAAC,mBAAmB,aAAqB;AAAA,EACzC,CAAC,oBAAoB,cAAsB;AAC7C,CAAC;","names":[]}

package/package.json

@@ -0,0 +1,38 @@
+{
+  "name": "@faircopy/rules-nlp",
+  "version": "1.1.1",
+  "description": "Optional NLP-powered ruleset for faircopy using compromise",
+  "type": "module",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js"
+    }
+  },
+  "files": [
+    "dist"
+  ],
+  "dependencies": {
+    "compromise": "^14.15.0",
+    "@faircopy/core": "1.1.1"
+  },
+  "devDependencies": {
+    "@types/bun": "latest",
+    "tsup": "^8.4.0",
+    "typescript": "^5.8.0"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "license": "MIT",
+  "author": "omniaura",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/omniaura/faircopy"
+  },
+  "scripts": {
+    "build": "tsup",
+    "typecheck": "tsc --noEmit",
+    "test": "echo 'no tests yet'"
+  }
+}