@alpaca-software/40kdc-data 0.3.1 → 0.4.0

package/dist/converters/base-size-guide-extract.js

@@ -0,0 +1,59 @@
+/**
+ * Extract base-size rows from the GW *Chapter Approved Tournament Companion —
+ * Base Size Guide* into a committed numerical-facts table.
+ *
+ * IP: base sizes are numerical facts (same category as stat lines / points, which
+ * the project permits) and unit/model names are entity identifiers already present
+ * in the dataset. We never commit the PDF or any prose/artwork from it — only the
+ * name → size rows. The source document is cited, not reproduced.
+ *
+ * Usage (one-time, from a locally-downloaded PDF; the PDF stays uncommitted):
+ *   pdftotext -layout tournament-companion.pdf tc.txt
+ *   tsx src/converters/base-size-guide-extract.ts tc.txt > src/converters/data/base-size-guide.json
+ *
+ * `-layout` mode keeps each "<unit name>   <base size>" pair on one physical line,
+ * so the two-column tables parse line-by-line. Multi-model datasheets appear as
+ * "Unit: ModelLabel" rows; shared rows as "UnitA/UnitB: ModelLabel".
+ */
+import { readFileSync } from "node:fs";
+/** A base-size token at the end of a guide line: round, oval, Hull, Unique, or a flying base. */
+const SIZE_TOKEN = "(?:\\d+(?:\\.\\d+)?\\s*[xX×]\\s*\\d+(?:\\.\\d+)?\\s*mm(?:\\s*Oval Base)?)" + // oval
+    "|(?:\\d+(?:\\.\\d+)?\\s*mm)" + // round
+    "|Hull|Unique|(?:Large|Small) Flying Base";
+const LINE_RE = new RegExp(`^(.*\\S)\\s{2,}(${SIZE_TOKEN})\\s*$`);
+/** Parse the layout-mode text of the Base Size Guide into structured rows. */
+export function extractGuideRows(layoutText) {
+    const lines = layoutText.split(/\r?\n/);
+    const start = lines.findIndex((l) => /BASE SIZE GUIDE/.test(l));
+    const region = start >= 0 ? lines.slice(start) : lines;
+    const rows = [];
+    for (const line of region) {
+        const m = line.match(LINE_RE);
+        if (!m)
+            continue;
+        const name = m[1].trim();
+        const raw = m[2].trim();
+        const colon = name.indexOf(":");
+        if (colon >= 0) {
+            rows.push({ unit: name.slice(0, colon).trim(), model: name.slice(colon + 1).trim(), raw });
+        }
+        else {
+            rows.push({ unit: name, raw });
+        }
+    }
+    return rows;
+}
+function main() {
+    const path = process.argv[2];
+    if (!path) {
+        console.error("usage: base-size-guide-extract.ts <pdftotext -layout output>");
+        process.exit(2);
+    }
+    const rows = extractGuideRows(readFileSync(path, "utf8"));
+    process.stdout.write(JSON.stringify(rows, null, 2) + "\n");
+    console.error(`extracted ${rows.length} rows`);
+}
+// Run as a script (not when imported by tests).
+if (import.meta.url === `file://${process.argv[1]}`)
+    main();
+//# sourceMappingURL=base-size-guide-extract.js.map

package/dist/converters/base-size-guide-extract.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"base-size-guide-extract.js","sourceRoot":"","sources":["../../src/converters/base-size-guide-extract.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AACH,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AAWvC,iGAAiG;AACjG,MAAM,UAAU,GACd,2EAA2E,GAAG,OAAO;IACrF,6BAA6B,GAAG,QAAQ;IACxC,0CAA0C,CAAC;AAE7C,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,mBAAmB,UAAU,QAAQ,CAAC,CAAC;AAElE,8EAA8E;AAC9E,MAAM,UAAU,gBAAgB,CAAC,UAAkB;IACjD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACxC,MAAM,KAAK,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,MAAM,GAAG,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;IAEvD,MAAM,IAAI,GAAe,EAAE,CAAC;IAC5B,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;QAC1B,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QAC9B,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACzB,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACxB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QAChC,IAAI,KAAK,IAAI,CAAC,EAAE,CAAC;YACf,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;QAC7F,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC,CAAC;QACjC,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,IAAI;IACX,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7B,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,CAAC,KAAK,CAAC,8DAA8D,CAAC,CAAC;QAC9E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,MAAM,IAAI,GAAG,gBAAgB,CAAC,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC;IAC1D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;IAC3D,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,CAAC,MAAM,OAAO,CAAC,CAAC;AACjD,CAAC;AAED,gDAAgD;AAChD,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,UAAU,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE;IAAE,IAAI,EAAE,CAAC","sourcesContent":["/**\n * Extract base-size rows from the GW *Chapter Approved Tournament Companion —\n * Base Size Guide* into a committed numerical-facts table.\n *\n * IP: base sizes are numerical facts (same category as stat lines / points, which\n * the project permits) and unit/model names are entity identifiers already present\n * in the dataset. We never commit the PDF or any prose/artwork from it — only the\n * name → size rows. The source document is cited, not reproduced.\n *\n * Usage (one-time, from a locally-downloaded PDF; the PDF stays uncommitted):\n *   pdftotext -layout tournament-companion.pdf tc.txt\n *   tsx src/converters/base-size-guide-extract.ts tc.txt > src/converters/data/base-size-guide.json\n *\n * `-layout` mode keeps each \"<unit name>   <base size>\" pair on one physical line,\n * so the two-column tables parse line-by-line. Multi-model datasheets appear as\n * \"Unit: ModelLabel\" rows; shared rows as \"UnitA/UnitB: ModelLabel\".\n */\nimport { readFileSync } from \"node:fs\";\n\nexport interface GuideRow {\n  /** Left-of-colon datasheet name(s). May be \"A/B\" for shared rows. */\n  unit: string;\n  /** Right-of-colon per-model label, when the row is model-specific. */\n  model?: string;\n  /** Verbatim base-size string (e.g. \"32mm\", \"60 x 35.5mm Oval Base\", \"Hull\"). */\n  raw: string;\n}\n\n/** A base-size token at the end of a guide line: round, oval, Hull, Unique, or a flying base. */\nconst SIZE_TOKEN =\n  \"(?:\\\\d+(?:\\\\.\\\\d+)?\\\\s*[xX×]\\\\s*\\\\d+(?:\\\\.\\\\d+)?\\\\s*mm(?:\\\\s*Oval Base)?)\" + // oval\n  \"|(?:\\\\d+(?:\\\\.\\\\d+)?\\\\s*mm)\" + // round\n  \"|Hull|Unique|(?:Large|Small) Flying Base\";\n\nconst LINE_RE = new RegExp(`^(.*\\\\S)\\\\s{2,}(${SIZE_TOKEN})\\\\s*$`);\n\n/** Parse the layout-mode text of the Base Size Guide into structured rows. */\nexport function extractGuideRows(layoutText: string): GuideRow[] {\n  const lines = layoutText.split(/\\r?\\n/);\n  const start = lines.findIndex((l) => /BASE SIZE GUIDE/.test(l));\n  const region = start >= 0 ? lines.slice(start) : lines;\n\n  const rows: GuideRow[] = [];\n  for (const line of region) {\n    const m = line.match(LINE_RE);\n    if (!m) continue;\n    const name = m[1].trim();\n    const raw = m[2].trim();\n    const colon = name.indexOf(\":\");\n    if (colon >= 0) {\n      rows.push({ unit: name.slice(0, colon).trim(), model: name.slice(colon + 1).trim(), raw });\n    } else {\n      rows.push({ unit: name, raw });\n    }\n  }\n  return rows;\n}\n\nfunction main(): void {\n  const path = process.argv[2];\n  if (!path) {\n    console.error(\"usage: base-size-guide-extract.ts <pdftotext -layout output>\");\n    process.exit(2);\n  }\n  const rows = extractGuideRows(readFileSync(path, \"utf8\"));\n  process.stdout.write(JSON.stringify(rows, null, 2) + \"\\n\");\n  console.error(`extracted ${rows.length} rows`);\n}\n\n// Run as a script (not when imported by tests).\nif (import.meta.url === `file://${process.argv[1]}`) main();\n"]}

package/dist/converters/option-bridge.d.ts

@@ -0,0 +1,36 @@
+/**
+ * Bridge army-assist's two datasheet id-spaces. The wargear-option table
+ * (`Datasheets_options.json`) and the unit-composition table
+ * (`Datasheets_unit_composition.json`) key on a *numeric* datasheet id
+ * (`000002627`) that never appears in the UUID-keyed `Datasheets.json` /
+ * `Datasheets_models.json`. The link is **model names**: composition
+ * descriptions name a datasheet's models (numeric side), and `Datasheets_models`
+ * names them (UUID side).
+ *
+ * Globally the match is ambiguous (the same model name recurs across factions),
+ * but the converter runs one faction at a time, so candidates are restricted to
+ * that faction's ~50–90 datasheets where model-name sets are effectively unique.
+ * Ties that survive are returned for the caller to report rather than guessed.
+ */
+/** Normalise a model name for matching: lowercase, keep alphanumerics + spaces. */
+export declare function normModelName(s: string): string;
+/**
+ * Extract a model name from a unit-composition description by dropping the
+ * leading count or range ("1 Khorne Berzerker Champion" → "khorne berzerker
+ * champion"; "9-19 Khorne Berzerkers" → "khorne berzerkers").
+ */
+export declare function modelNameFromComposition(desc: string): string;
+export interface BridgeResult {
+    /** numeric datasheet id → UUID datasheet id (unique best model-name match). */
+    byNumeric: Map<string, string>;
+    /** numeric ids that overlap the faction but tie across ≥2 UUIDs. */
+    ambiguous: string[];
+}
+/**
+ * Resolve each numeric datasheet id that carries options to a faction UUID by
+ * best model-name-set overlap. Numeric ids with no overlap are treated as
+ * belonging to another faction and dropped silently; ties (overlap > 0 but no
+ * unique winner) are returned in `ambiguous`.
+ */
+export declare function bridgeOptionsToUnits(factionUuids: readonly string[], modelsByUuid: Map<string, Set<string>>, compByNumeric: Map<string, Set<string>>, optionNumericIds: Iterable<string>): BridgeResult;
+//# sourceMappingURL=option-bridge.d.ts.map

package/dist/converters/option-bridge.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"option-bridge.d.ts","sourceRoot":"","sources":["../../src/converters/option-bridge.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,mFAAmF;AACnF,wBAAgB,aAAa,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,CAO/C;AAED;;;;GAIG;AACH,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAE7D;AAED,MAAM,WAAW,YAAY;IAC3B,+EAA+E;IAC/E,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC/B,oEAAoE;IACpE,SAAS,EAAE,MAAM,EAAE,CAAC;CACrB;AAED;;;;;GAKG;AACH,wBAAgB,oBAAoB,CAClC,YAAY,EAAE,SAAS,MAAM,EAAE,EAC/B,YAAY,EAAE,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,EACtC,aAAa,EAAE,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,EACvC,gBAAgB,EAAE,QAAQ,CAAC,MAAM,CAAC,GACjC,YAAY,CA6Bd"}

package/dist/converters/option-bridge.js

@@ -0,0 +1,72 @@
+/**
+ * Bridge army-assist's two datasheet id-spaces. The wargear-option table
+ * (`Datasheets_options.json`) and the unit-composition table
+ * (`Datasheets_unit_composition.json`) key on a *numeric* datasheet id
+ * (`000002627`) that never appears in the UUID-keyed `Datasheets.json` /
+ * `Datasheets_models.json`. The link is **model names**: composition
+ * descriptions name a datasheet's models (numeric side), and `Datasheets_models`
+ * names them (UUID side).
+ *
+ * Globally the match is ambiguous (the same model name recurs across factions),
+ * but the converter runs one faction at a time, so candidates are restricted to
+ * that faction's ~50–90 datasheets where model-name sets are effectively unique.
+ * Ties that survive are returned for the caller to report rather than guessed.
+ */
+/** Normalise a model name for matching: lowercase, keep alphanumerics + spaces. */
+export function normModelName(s) {
+    return s
+        .normalize("NFD")
+        .replace(/[̀-ͯ]/g, "")
+        .toLowerCase()
+        .replace(/[^a-z0-9]+/g, " ")
+        .trim();
+}
+/**
+ * Extract a model name from a unit-composition description by dropping the
+ * leading count or range ("1 Khorne Berzerker Champion" → "khorne berzerker
+ * champion"; "9-19 Khorne Berzerkers" → "khorne berzerkers").
+ */
+export function modelNameFromComposition(desc) {
+    return normModelName(desc.replace(/^\s*\d+\s*[-–]?\s*\d*\s+/, ""));
+}
+/**
+ * Resolve each numeric datasheet id that carries options to a faction UUID by
+ * best model-name-set overlap. Numeric ids with no overlap are treated as
+ * belonging to another faction and dropped silently; ties (overlap > 0 but no
+ * unique winner) are returned in `ambiguous`.
+ */
+export function bridgeOptionsToUnits(factionUuids, modelsByUuid, compByNumeric, optionNumericIds) {
+    const byNumeric = new Map();
+    const ambiguous = [];
+    for (const numericId of optionNumericIds) {
+        const numModels = compByNumeric.get(numericId);
+        if (!numModels || numModels.size === 0)
+            continue;
+        let bestScore = 0;
+        let winners = [];
+        for (const uuid of factionUuids) {
+            const uModels = modelsByUuid.get(uuid);
+            if (!uModels)
+                continue;
+            let score = 0;
+            for (const m of numModels)
+                if (uModels.has(m))
+                    score++;
+            if (score > bestScore) {
+                bestScore = score;
+                winners = [uuid];
+            }
+            else if (score === bestScore && score > 0) {
+                winners.push(uuid);
+            }
+        }
+        if (bestScore === 0)
+            continue; // another faction's datasheet
+        if (winners.length === 1)
+            byNumeric.set(numericId, winners[0]);
+        else
+            ambiguous.push(numericId);
+    }
+    return { byNumeric, ambiguous };
+}
+//# sourceMappingURL=option-bridge.js.map

package/dist/converters/option-bridge.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"option-bridge.js","sourceRoot":"","sources":["../../src/converters/option-bridge.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,mFAAmF;AACnF,MAAM,UAAU,aAAa,CAAC,CAAS;IACrC,OAAO,CAAC;SACL,SAAS,CAAC,KAAK,CAAC;SAChB,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC;SACrB,WAAW,EAAE;SACb,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC;SAC3B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,wBAAwB,CAAC,IAAY;IACnD,OAAO,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,0BAA0B,EAAE,EAAE,CAAC,CAAC,CAAC;AACrE,CAAC;AASD;;;;;GAKG;AACH,MAAM,UAAU,oBAAoB,CAClC,YAA+B,EAC/B,YAAsC,EACtC,aAAuC,EACvC,gBAAkC;IAElC,MAAM,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC5C,MAAM,SAAS,GAAa,EAAE,CAAC;IAE/B,KAAK,MAAM,SAAS,IAAI,gBAAgB,EAAE,CAAC;QACzC,MAAM,SAAS,GAAG,aAAa,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC/C,IAAI,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC;YAAE,SAAS;QAEjD,IAAI,SAAS,GAAG,CAAC,CAAC;QAClB,IAAI,OAAO,GAAa,EAAE,CAAC;QAC3B,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;YAChC,MAAM,OAAO,GAAG,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACvC,IAAI,CAAC,OAAO;gBAAE,SAAS;YACvB,IAAI,KAAK,GAAG,CAAC,CAAC;YACd,KAAK,MAAM,CAAC,IAAI,SAAS;gBAAE,IAAI,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;oBAAE,KAAK,EAAE,CAAC;YACvD,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;gBACtB,SAAS,GAAG,KAAK,CAAC;gBAClB,OAAO,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;iBAAM,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;gBAC5C,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACrB,CAAC;QACH,CAAC;QAED,IAAI,SAAS,KAAK,CAAC;YAAE,SAAS,CAAC,8BAA8B;QAC7D,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,SAAS,CAAC,GAAG,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;;YAC1D,SAAS,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IACjC,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,CAAC;AAClC,CAAC","sourcesContent":["/**\n * Bridge army-assist's two datasheet id-spaces. The wargear-option table\n * (`Datasheets_options.json`) and the unit-composition table\n * (`Datasheets_unit_composition.json`) key on a *numeric* datasheet id\n * (`000002627`) that never appears in the UUID-keyed `Datasheets.json` /\n * `Datasheets_models.json`. The link is **model names**: composition\n * descriptions name a datasheet's models (numeric side), and `Datasheets_models`\n * names them (UUID side).\n *\n * Globally the match is ambiguous (the same model name recurs across factions),\n * but the converter runs one faction at a time, so candidates are restricted to\n * that faction's ~50–90 datasheets where model-name sets are effectively unique.\n * Ties that survive are returned for the caller to report rather than guessed.\n */\n\n/** Normalise a model name for matching: lowercase, keep alphanumerics + spaces. */\nexport function normModelName(s: string): string {\n  return s\n    .normalize(\"NFD\")\n    .replace(/[̀-ͯ]/g, \"\")\n    .toLowerCase()\n    .replace(/[^a-z0-9]+/g, \" \")\n    .trim();\n}\n\n/**\n * Extract a model name from a unit-composition description by dropping the\n * leading count or range (\"1 Khorne Berzerker Champion\" → \"khorne berzerker\n * champion\"; \"9-19 Khorne Berzerkers\" → \"khorne berzerkers\").\n */\nexport function modelNameFromComposition(desc: string): string {\n  return normModelName(desc.replace(/^\\s*\\d+\\s*[-–]?\\s*\\d*\\s+/, \"\"));\n}\n\nexport interface BridgeResult {\n  /** numeric datasheet id → UUID datasheet id (unique best model-name match). */\n  byNumeric: Map<string, string>;\n  /** numeric ids that overlap the faction but tie across ≥2 UUIDs. */\n  ambiguous: string[];\n}\n\n/**\n * Resolve each numeric datasheet id that carries options to a faction UUID by\n * best model-name-set overlap. Numeric ids with no overlap are treated as\n * belonging to another faction and dropped silently; ties (overlap > 0 but no\n * unique winner) are returned in `ambiguous`.\n */\nexport function bridgeOptionsToUnits(\n  factionUuids: readonly string[],\n  modelsByUuid: Map<string, Set<string>>,\n  compByNumeric: Map<string, Set<string>>,\n  optionNumericIds: Iterable<string>,\n): BridgeResult {\n  const byNumeric = new Map<string, string>();\n  const ambiguous: string[] = [];\n\n  for (const numericId of optionNumericIds) {\n    const numModels = compByNumeric.get(numericId);\n    if (!numModels || numModels.size === 0) continue;\n\n    let bestScore = 0;\n    let winners: string[] = [];\n    for (const uuid of factionUuids) {\n      const uModels = modelsByUuid.get(uuid);\n      if (!uModels) continue;\n      let score = 0;\n      for (const m of numModels) if (uModels.has(m)) score++;\n      if (score > bestScore) {\n        bestScore = score;\n        winners = [uuid];\n      } else if (score === bestScore && score > 0) {\n        winners.push(uuid);\n      }\n    }\n\n    if (bestScore === 0) continue; // another faction's datasheet\n    if (winners.length === 1) byNumeric.set(numericId, winners[0]);\n    else ambiguous.push(numericId);\n  }\n\n  return { byNumeric, ambiguous };\n}\n"]}

package/dist/converters/option-parser.d.ts

@@ -0,0 +1,56 @@
+/**
+ * Parse a Warhammer wargear-option line (army-assist `Datasheets_options.json`
+ * `description` prose) into a structured option, or report why it could not.
+ *
+ * The prose is regular enough to parse with a small staged pipeline:
+ *   1. skip non-options (null / "None" / footnotes / "cannot be replaced");
+ *   2. peel the leading *constraint* clause ("For every 5 models,", "The
+ *      Champion's", "Any number of Boyz", "Up to 2 models", "1 model");
+ *   3. split the core on the swap/add verb (passive "… can be replaced with …",
+ *      active "… can replace its … with …", or add-on "… equipped with / can
+ *      have …") into the replaced weapon(s) and the replacement clause;
+ *   4. parse the replacement clause into a flat group or a "one of the
+ *      following" choice.
+ *
+ * Output carries weapon/wargear **display names**, not ids — the converter
+ * resolves those against each unit's own wargear (see convert-faction.ts).
+ * Anything that does not fit returns `{ ok: false, reason }` so the caller can
+ * surface it in the unparsed report rather than guess. Mechanic-only; no
+ * copyrighted rules text is reproduced.
+ */
+export interface ParsedConstraint {
+    model_name?: string;
+    per_n_models?: number;
+    max_count?: number;
+    any_number?: boolean;
+}
+export interface ParsedOption {
+    kind: "swap" | "addon";
+    constraint: ParsedConstraint;
+    /** Display names of weapons removed; empty for a pure add-on. */
+    replaces: string[];
+    /** Display names added (all of them). Set iff not a choice. */
+    replacement?: string[];
+    /** Choice of groups ("one of the following"); pick one group. Set iff a choice. */
+    replacement_choice?: string[][];
+}
+export type ParseResult = {
+    ok: true;
+    option: ParsedOption;
+}
+/** A real option we failed to parse — goes to the unparsed report. */
+ | {
+    ok: false;
+    reason: string;
+}
+/** Not an option at all (footnote / empty / note) — silently ignored. */
+ | {
+    ok: "skip";
+};
+/**
+ * Parse one option `description`. Returns `{ ok: "skip" }` for non-options,
+ * `{ ok: false, reason }` for a real option we couldn't parse, and
+ * `{ ok: true, option }` otherwise.
+ */
+export declare function parseOption(description: string | null | undefined): ParseResult;
+//# sourceMappingURL=option-parser.d.ts.map

package/dist/converters/option-parser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"option-parser.d.ts","sourceRoot":"","sources":["../../src/converters/option-parser.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,MAAM,WAAW,gBAAgB;IAC/B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,OAAO,CAAC;CACtB;AAED,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC;IACvB,UAAU,EAAE,gBAAgB,CAAC;IAC7B,iEAAiE;IACjE,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,+DAA+D;IAC/D,WAAW,CAAC,EAAE,MAAM,EAAE,CAAC;IACvB,mFAAmF;IACnF,kBAAkB,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;CACjC;AAED,MAAM,MAAM,WAAW,GACnB;IAAE,EAAE,EAAE,IAAI,CAAC;IAAC,MAAM,EAAE,YAAY,CAAA;CAAE;AACpC,sEAAsE;GACpE;IAAE,EAAE,EAAE,KAAK,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE;AAC/B,yEAAyE;GACvE;IAAE,EAAE,EAAE,MAAM,CAAA;CAAE,CAAC;AA6JnB;;;;GAIG;AACH,wBAAgB,WAAW,CAAC,WAAW,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,WAAW,CAkC/E"}

package/dist/converters/option-parser.js

@@ -0,0 +1,209 @@
+/**
+ * Parse a Warhammer wargear-option line (army-assist `Datasheets_options.json`
+ * `description` prose) into a structured option, or report why it could not.
+ *
+ * The prose is regular enough to parse with a small staged pipeline:
+ *   1. skip non-options (null / "None" / footnotes / "cannot be replaced");
+ *   2. peel the leading *constraint* clause ("For every 5 models,", "The
+ *      Champion's", "Any number of Boyz", "Up to 2 models", "1 model");
+ *   3. split the core on the swap/add verb (passive "… can be replaced with …",
+ *      active "… can replace its … with …", or add-on "… equipped with / can
+ *      have …") into the replaced weapon(s) and the replacement clause;
+ *   4. parse the replacement clause into a flat group or a "one of the
+ *      following" choice.
+ *
+ * Output carries weapon/wargear **display names**, not ids — the converter
+ * resolves those against each unit's own wargear (see convert-faction.ts).
+ * Anything that does not fit returns `{ ok: false, reason }` so the caller can
+ * surface it in the unparsed report rather than guess. Mechanic-only; no
+ * copyrighted rules text is reproduced.
+ */
+/** Tidy a captured weapon/wargear name: drop footnote markers, collapse spaces. */
+function cleanName(raw) {
+    return raw
+        .replace(/\*+/g, "") // footnote markers ("blastmaster*")
+        .replace(/\s+/g, " ")
+        .replace(/^[.,;:\s]+|[.,;:\s]+$/g, "")
+        .trim();
+}
+/** Strip a leading per-model quantity ("1 ", "2 ") from a replacement item. */
+function stripLeadingCount(item) {
+    return cleanName(item.replace(/^\s*\d+\s+/, ""));
+}
+/** Split a flat replacement/replaces clause on " and " into individual names. */
+function splitAnd(clause) {
+    return clause
+        .split(/\s+and\s+/i)
+        .map(stripLeadingCount)
+        .filter((s) => /[a-z0-9]/i.test(s)); // drop empties and stray punctuation
+}
+/**
+ * Split the concatenated "one of the following" list into groups. The source
+ * runs items together with no separator ("1 flamer1 grav-gun1 meltagun"), so we
+ * break before each leading count. Each item may itself be "A and B".
+ */
+function parseChoiceList(tail) {
+    // Break before a digit that begins a new item (preceded by a non-digit).
+    const parts = tail
+        .split(/(?<=\D)(?=\d+\s)/)
+        .map((s) => s.trim())
+        .filter((s) => s.length > 0);
+    const groups = parts.map(splitAnd).filter((g) => g.length > 0);
+    return groups;
+}
+/** Parse the text after the swap/add verb into a flat group or a choice. */
+function parseReplacement(rest) {
+    const choiceMatch = rest.match(/one of the following\s*:?\s*(.+)$/is);
+    if (choiceMatch) {
+        const groups = parseChoiceList(choiceMatch[1]);
+        // A "choice" that parsed to a single alternative is just a plain
+        // replacement — the schema reserves replacement_choice for ≥2 options.
+        if (groups.length >= 2)
+            return { replacement_choice: groups };
+        if (groups.length === 1)
+            return { replacement: groups[0] };
+        return {};
+    }
+    return { replacement: splitAnd(rest) };
+}
+/** Normalise known source typos / spacing so the verb regexes match uniformly. */
+function normalize(desc) {
+    return desc
+        .replace(/–|—/g, "-")
+        .replace(/replaced\s+with/gi, "replaced with") // collapse double space
+        .replace(/can be replace with/gi, "can be replaced with") // missing 'd'
+        .replace(/can be replace\b(?! with)/gi, "can be replaced")
+        .replace(/replaced one of the following/gi, "replaced with one of the following")
+        .replace(/following(?=\d)/gi, "following:") // "following1 flamer" -> "following:1 flamer"
+        .replace(/\s+/g, " ")
+        .trim();
+}
+/** Peel the leading constraint clause, returning it plus the remaining core. */
+function extractConstraint(core) {
+    const constraint = {};
+    let s = core;
+    const forEvery = s.match(/^for every (\d+) models?(?: in (?:this unit|the unit))?,?\s*/i);
+    if (forEvery) {
+        constraint.per_n_models = parseInt(forEvery[1], 10);
+        s = s.slice(forEvery[0].length);
+    }
+    if (/^any number of\b/i.test(s)) {
+        constraint.any_number = true;
+        const m = s.match(/^any number of\s+(.+?)\s+can\b/i);
+        if (m && !/^models?$/i.test(m[1]))
+            constraint.model_name = cleanName(m[1]);
+        s = s.replace(/^any number of\s+.+?\s+(?=can\b)/i, "");
+    }
+    else {
+        const upTo = s.match(/^up to (\d+)\s+(.+?)\s+(?=can\b)/i);
+        if (upTo) {
+            constraint.max_count = parseInt(upTo[1], 10);
+            if (!/^models?$/i.test(upTo[2]))
+                constraint.model_name = cleanName(upTo[2]);
+            s = s.slice(upTo[0].length);
+        }
+        else {
+            // "The Champion's …", "The Kill Team Sergeant can …"
+            const theRole = s.match(/^the\s+(.+?)(?:'s|’s)\s+/i) || s.match(/^the\s+(.+?)\s+can\b/i);
+            if (theRole) {
+                constraint.model_name = cleanName(theRole[1]);
+                if (constraint.max_count === undefined && constraint.per_n_models === undefined) {
+                    constraint.max_count = 1;
+                }
+                // Drop only the "The <role>" lead-in; keep the possessive weapon for the verb stage.
+                s = s.replace(/^the\s+/i, "");
+            }
+            else {
+                // "1 model …", "1 Khorne Berzerker's …", "One model …"
+                const leadCount = s.match(/^(\d+|one)\s+(model|.+?)(?=(?:'s|’s)\s+|\s+can\b)/i);
+                if (leadCount) {
+                    const n = /^one$/i.test(leadCount[1]) ? 1 : parseInt(leadCount[1], 10);
+                    if (constraint.per_n_models === undefined && constraint.max_count === undefined) {
+                        constraint.max_count = n;
+                    }
+                    if (!/^models?$/i.test(leadCount[2]))
+                        constraint.model_name = cleanName(leadCount[2]);
+                    s = s.replace(/^(?:\d+|one)\s+(?:model\s+)?/i, "");
+                }
+            }
+        }
+    }
+    // A bare "For every N models, 1 <model>'s …" leaves a leading "1 <model>" —
+    // and "This model"/"This unit"/"it" lead-ins carry no constraint.
+    s = s.replace(/^(\d+)\s+(?=\S+(?:'s|’s)\s)/, "");
+    return { constraint, rest: s.trim() };
+}
+/** Pull the replaced weapon(s) and the replacement clause out of the core. */
+function splitOnVerb(core) {
+    // Active voice: "… (can )?(each )?replace(s|d) (its|their|one of its/their) WEAPON with REST"
+    const active = core.match(/\b(?:can\s+)?(?:each\s+)?replaces?\s+(?:one of\s+)?(?:its|their)\s+(.+?)\s+with\s*:?\s*(.+)$/i);
+    if (active) {
+        return { replaces: splitAnd(active[1]), rest: active[2].trim() };
+    }
+    // "have/has (its|their) WEAPON replaced with REST" ("Any number of Boyz can
+    // each have their slugga and choppa replaced with …").
+    const havePassive = core.match(/\b(?:have|has)\s+(?:its|their)\s+(.+?)\s+replaced\s+with\s*:?\s*(.+)$/i);
+    if (havePassive) {
+        return { replaces: splitAnd(havePassive[1]), rest: havePassive[2].trim() };
+    }
+    // Passive voice: "… WEAPON can [each] be replaced with REST" (WEAPON sits after
+    // a possessive 's, or is the whole lead if none).
+    const passive = core.match(/^(.*?)\s+can (?:each )?be replaced\s+with\s*:?\s*(.+)$/i);
+    if (passive) {
+        const lead = passive[1];
+        const possessive = lead.match(/(?:'s|’s)\s+(.+)$/);
+        const weapon = possessive ? possessive[1] : lead;
+        return { replaces: splitAnd(weapon), rest: passive[2].trim() };
+    }
+    return null;
+}
+/** Pull the added wargear out of an add-on core ("equipped with …" / "can have …"). */
+function splitOnAddVerb(core) {
+    // Allow an immediate colon ("equipped with:1 lobba") as well as a space.
+    const m = core.match(/\b(?:equipped with|can have)\s*:?\s*(.+)$/i);
+    return m ? { rest: m[1].trim() } : null;
+}
+/**
+ * Parse one option `description`. Returns `{ ok: "skip" }` for non-options,
+ * `{ ok: false, reason }` for a real option we couldn't parse, and
+ * `{ ok: true, option }` otherwise.
+ */
+export function parseOption(description) {
+    if (description == null)
+        return { ok: "skip" };
+    const trimmed = description.trim();
+    if (trimmed === "" || trimmed === "None")
+        return { ok: "skip" };
+    if (trimmed.startsWith("*"))
+        return { ok: "skip" };
+    if (/cannot be (replaced|taken|equipped)/i.test(trimmed))
+        return { ok: "skip" };
+    const desc = normalize(trimmed);
+    const isSwap = /\breplace/i.test(desc);
+    const isAddon = !isSwap && /\b(equipped with|can have)\b/i.test(desc);
+    if (!isSwap && !isAddon) {
+        return { ok: false, reason: "no swap/add verb recognised" };
+    }
+    const { constraint, rest: core } = extractConstraint(desc);
+    if (isSwap) {
+        const split = splitOnVerb(core);
+        if (!split)
+            return { ok: false, reason: "could not isolate replaced weapon / replacement" };
+        if (split.replaces.length === 0)
+            return { ok: false, reason: "empty replaced-weapon list" };
+        const repl = parseReplacement(split.rest);
+        if (!repl.replacement?.length && !repl.replacement_choice?.length) {
+            return { ok: false, reason: "empty replacement" };
+        }
+        return { ok: true, option: { kind: "swap", constraint, replaces: split.replaces, ...repl } };
+    }
+    const add = splitOnAddVerb(core);
+    if (!add)
+        return { ok: false, reason: "could not isolate added wargear" };
+    const repl = parseReplacement(add.rest);
+    if (!repl.replacement?.length && !repl.replacement_choice?.length) {
+        return { ok: false, reason: "empty add-on" };
+    }
+    return { ok: true, option: { kind: "addon", constraint, replaces: [], ...repl } };
+}
+//# sourceMappingURL=option-parser.js.map

package/dist/converters/option-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"option-parser.js","sourceRoot":"","sources":["../../src/converters/option-parser.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AA2BH,mFAAmF;AACnF,SAAS,SAAS,CAAC,GAAW;IAC5B,OAAO,GAAG;SACP,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,oCAAoC;SACxD,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,OAAO,CAAC,wBAAwB,EAAE,EAAE,CAAC;SACrC,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,+EAA+E;AAC/E,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC,CAAC;AACnD,CAAC;AAED,iFAAiF;AACjF,SAAS,QAAQ,CAAC,MAAc;IAC9B,OAAO,MAAM;SACV,KAAK,CAAC,YAAY,CAAC;SACnB,GAAG,CAAC,iBAAiB,CAAC;SACtB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,qCAAqC;AAC9E,CAAC;AAED;;;;GAIG;AACH,SAAS,eAAe,CAAC,IAAY;IACnC,yEAAyE;IACzE,MAAM,KAAK,GAAG,IAAI;SACf,KAAK,CAAC,kBAAkB,CAAC;SACzB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC/B,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC/D,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,4EAA4E;AAC5E,SAAS,gBAAgB,CAAC,IAAY;IACpC,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAC;IACtE,IAAI,WAAW,EAAE,CAAC;QAChB,MAAM,MAAM,GAAG,eAAe,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;QAC/C,iEAAiE;QACjE,uEAAuE;QACvE,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC;YAAE,OAAO,EAAE,kBAAkB,EAAE,MAAM,EAAE,CAAC;QAC9D,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,WAAW,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,OAAO,EAAE,CAAC;IACZ,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;AACzC,CAAC;AAED,kFAAkF;AAClF,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI;SACR,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,OAAO,CAAC,mBAAmB,EAAE,eAAe,CAAC,CAAC,wBAAwB;SACtE,OAAO,CAAC,uBAAuB,EAAE,sBAAsB,CAAC,CAAC,cAAc;SACvE,OAAO,CAAC,6BAA6B,EAAE,iBAAiB,CAAC;SACzD,OAAO,CAAC,iCAAiC,EAAE,oCAAoC,CAAC;SAChF,OAAO,CAAC,mBAAmB,EAAE,YAAY,CAAC,CAAC,8CAA8C;SACzF,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,gFAAgF;AAChF,SAAS,iBAAiB,CAAC,IAAY;IACrC,MAAM,UAAU,GAAqB,EAAE,CAAC;IACxC,IAAI,CAAC,GAAG,IAAI,CAAC;IAEb,MAAM,QAAQ,GAAG,CAAC,CAAC,KAAK,CAAC,+DAA+D,CAAC,CAAC;IAC1F,IAAI,QAAQ,EAAE,CAAC;QACb,UAAU,CAAC,YAAY,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QACpD,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;IAClC,CAAC;IAED,IAAI,mBAAmB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;QAChC,UAAU,CAAC,UAAU,GAAG,IAAI,CAAC;QAC7B,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACrD,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAAE,UAAU,CAAC,UAAU,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC3E,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,mCAAmC,EAAE,EAAE,CAAC,CAAC;IACzD,CAAC;SAAM,CAAC;QACN,MAAM,IAAI,GAAG,CAAC,CAAC,KAAK,CAAC,mCAAmC,CAAC,CAAC;QAC1D,IAAI,IAAI,EAAE,CAAC;YACT,UAAU,CAAC,SAAS,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC7C,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBAAE,UAAU,CAAC,UAAU,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YAC5E,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QAC9B,CAAC;aAAM,CAAC;YACN,qDAAqD;YACrD,MAAM,OAAO,GAAG,CAAC,CAAC,KAAK,CAAC,2BAA2B,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,uBAAuB,CAAC,CAAC;YACzF,IAAI,OAAO,EAAE,CAAC;gBACZ,UAAU,CAAC,UAAU,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC9C,IAAI,UAAU,CAAC,SAAS,KAAK,SAAS,IAAI,UAAU,CAAC,YAAY,KAAK,SAAS,EAAE,CAAC;oBAChF,UAAU,CAAC,SAAS,GAAG,CAAC,CAAC;gBAC3B,CAAC;gBACD,qFAAqF;gBACrF,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;YAChC,CAAC;iBAAM,CAAC;gBACN,uDAAuD;gBACvD,MAAM,SAAS,GAAG,CAAC,CAAC,KAAK,CAAC,oDAAoD,CAAC,CAAC;gBAChF,IAAI,SAAS,EAAE,CAAC;oBACd,MAAM,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;oBACvE,IAAI,UAAU,CAAC,YAAY,KAAK,SAAS,IAAI,UAAU,CAAC,SAAS,KAAK,SAAS,EAAE,CAAC;wBAChF,UAAU,CAAC,SAAS,GAAG,CAAC,CAAC;oBAC3B,CAAC;oBACD,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;wBAAE,UAAU,CAAC,UAAU,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC;oBACtF,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,+BAA+B,EAAE,EAAE,CAAC,CAAC;gBACrD,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,6BAA6B,EAAE,EAAE,CAAC,CAAC;IACjD,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;AACxC,CAAC;AAED,8EAA8E;AAC9E,SAAS,WAAW,CAClB,IAAY;IAEZ,8FAA8F;IAC9F,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CACvB,+FAA+F,CAChG,CAAC;IACF,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;IACnE,CAAC;IACD,4EAA4E;IAC5E,uDAAuD;IACvD,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAC5B,wEAAwE,CACzE,CAAC;IACF,IAAI,WAAW,EAAE,CAAC;QAChB,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,WAAW,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;IAC7E,CAAC;IACD,gFAAgF;IAChF,kDAAkD;IAClD,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,yDAAyD,CAAC,CAAC;IACtF,IAAI,OAAO,EAAE,CAAC;QACZ,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACxB,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC;QACnD,MAAM,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QACjD,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;IACjE,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,uFAAuF;AACvF,SAAS,cAAc,CAAC,IAAY;IAClC,yEAAyE;IACzE,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,4CAA4C,CAAC,CAAC;IACnE,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;AAC1C,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,WAAW,CAAC,WAAsC;IAChE,IAAI,WAAW,IAAI,IAAI;QAAE,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC;IAC/C,MAAM,OAAO,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC;IACnC,IAAI,OAAO,KAAK,EAAE,IAAI,OAAO,KAAK,MAAM;QAAE,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC;IAChE,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC;IACnD,IAAI,sCAAsC,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC;IAEhF,MAAM,IAAI,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IAChC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,CAAC,MAAM,IAAI,+BAA+B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACtE,IAAI,CAAC,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;QACxB,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,6BAA6B,EAAE,CAAC;IAC9D,CAAC;IAED,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;IAE3D,IAAI,MAAM,EAAE,CAAC;QACX,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;QAChC,IAAI,CAAC,KAAK;YAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,iDAAiD,EAAE,CAAC;QAC5F,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,4BAA4B,EAAE,CAAC;QAC5F,MAAM,IAAI,GAAG,gBAAgB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC1C,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,MAAM,IAAI,CAAC,IAAI,CAAC,kBAAkB,EAAE,MAAM,EAAE,CAAC;YAClE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,mBAAmB,EAAE,CAAC;QACpD,CAAC;QACD,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,UAAU,EAAE,QAAQ,EAAE,KAAK,CAAC,QAAQ,EAAE,GAAG,IAAI,EAAE,EAAE,CAAC;IAC/F,CAAC;IAED,MAAM,GAAG,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;IACjC,IAAI,CAAC,GAAG;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,iCAAiC,EAAE,CAAC;IAC1E,MAAM,IAAI,GAAG,gBAAgB,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACxC,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,MAAM,IAAI,CAAC,IAAI,CAAC,kBAAkB,EAAE,MAAM,EAAE,CAAC;QAClE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,cAAc,EAAE,CAAC;IAC/C,CAAC;IACD,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,EAAE,EAAE,GAAG,IAAI,EAAE,EAAE,CAAC;AACpF,CAAC","sourcesContent":["/**\n * Parse a Warhammer wargear-option line (army-assist `Datasheets_options.json`\n * `description` prose) into a structured option, or report why it could not.\n *\n * The prose is regular enough to parse with a small staged pipeline:\n *   1. skip non-options (null / \"None\" / footnotes / \"cannot be replaced\");\n *   2. peel the leading *constraint* clause (\"For every 5 models,\", \"The\n *      Champion's\", \"Any number of Boyz\", \"Up to 2 models\", \"1 model\");\n *   3. split the core on the swap/add verb (passive \"… can be replaced with …\",\n *      active \"… can replace its … with …\", or add-on \"… equipped with / can\n *      have …\") into the replaced weapon(s) and the replacement clause;\n *   4. parse the replacement clause into a flat group or a \"one of the\n *      following\" choice.\n *\n * Output carries weapon/wargear **display names**, not ids — the converter\n * resolves those against each unit's own wargear (see convert-faction.ts).\n * Anything that does not fit returns `{ ok: false, reason }` so the caller can\n * surface it in the unparsed report rather than guess. Mechanic-only; no\n * copyrighted rules text is reproduced.\n */\n\nexport interface ParsedConstraint {\n  model_name?: string;\n  per_n_models?: number;\n  max_count?: number;\n  any_number?: boolean;\n}\n\nexport interface ParsedOption {\n  kind: \"swap\" | \"addon\";\n  constraint: ParsedConstraint;\n  /** Display names of weapons removed; empty for a pure add-on. */\n  replaces: string[];\n  /** Display names added (all of them). Set iff not a choice. */\n  replacement?: string[];\n  /** Choice of groups (\"one of the following\"); pick one group. Set iff a choice. */\n  replacement_choice?: string[][];\n}\n\nexport type ParseResult =\n  | { ok: true; option: ParsedOption }\n  /** A real option we failed to parse — goes to the unparsed report. */\n  | { ok: false; reason: string }\n  /** Not an option at all (footnote / empty / note) — silently ignored. */\n  | { ok: \"skip\" };\n\n/** Tidy a captured weapon/wargear name: drop footnote markers, collapse spaces. */\nfunction cleanName(raw: string): string {\n  return raw\n    .replace(/\\*+/g, \"\") // footnote markers (\"blastmaster*\")\n    .replace(/\\s+/g, \" \")\n    .replace(/^[.,;:\\s]+|[.,;:\\s]+$/g, \"\")\n    .trim();\n}\n\n/** Strip a leading per-model quantity (\"1 \", \"2 \") from a replacement item. */\nfunction stripLeadingCount(item: string): string {\n  return cleanName(item.replace(/^\\s*\\d+\\s+/, \"\"));\n}\n\n/** Split a flat replacement/replaces clause on \" and \" into individual names. */\nfunction splitAnd(clause: string): string[] {\n  return clause\n    .split(/\\s+and\\s+/i)\n    .map(stripLeadingCount)\n    .filter((s) => /[a-z0-9]/i.test(s)); // drop empties and stray punctuation\n}\n\n/**\n * Split the concatenated \"one of the following\" list into groups. The source\n * runs items together with no separator (\"1 flamer1 grav-gun1 meltagun\"), so we\n * break before each leading count. Each item may itself be \"A and B\".\n */\nfunction parseChoiceList(tail: string): string[][] {\n  // Break before a digit that begins a new item (preceded by a non-digit).\n  const parts = tail\n    .split(/(?<=\\D)(?=\\d+\\s)/)\n    .map((s) => s.trim())\n    .filter((s) => s.length > 0);\n  const groups = parts.map(splitAnd).filter((g) => g.length > 0);\n  return groups;\n}\n\n/** Parse the text after the swap/add verb into a flat group or a choice. */\nfunction parseReplacement(rest: string): Pick<ParsedOption, \"replacement\" | \"replacement_choice\"> {\n  const choiceMatch = rest.match(/one of the following\\s*:?\\s*(.+)$/is);\n  if (choiceMatch) {\n    const groups = parseChoiceList(choiceMatch[1]);\n    // A \"choice\" that parsed to a single alternative is just a plain\n    // replacement — the schema reserves replacement_choice for ≥2 options.\n    if (groups.length >= 2) return { replacement_choice: groups };\n    if (groups.length === 1) return { replacement: groups[0] };\n    return {};\n  }\n  return { replacement: splitAnd(rest) };\n}\n\n/** Normalise known source typos / spacing so the verb regexes match uniformly. */\nfunction normalize(desc: string): string {\n  return desc\n    .replace(/–|—/g, \"-\")\n    .replace(/replaced\\s+with/gi, \"replaced with\") // collapse double space\n    .replace(/can be replace with/gi, \"can be replaced with\") // missing 'd'\n    .replace(/can be replace\\b(?! with)/gi, \"can be replaced\")\n    .replace(/replaced one of the following/gi, \"replaced with one of the following\")\n    .replace(/following(?=\\d)/gi, \"following:\") // \"following1 flamer\" -> \"following:1 flamer\"\n    .replace(/\\s+/g, \" \")\n    .trim();\n}\n\n/** Peel the leading constraint clause, returning it plus the remaining core. */\nfunction extractConstraint(core: string): { constraint: ParsedConstraint; rest: string } {\n  const constraint: ParsedConstraint = {};\n  let s = core;\n\n  const forEvery = s.match(/^for every (\\d+) models?(?: in (?:this unit|the unit))?,?\\s*/i);\n  if (forEvery) {\n    constraint.per_n_models = parseInt(forEvery[1], 10);\n    s = s.slice(forEvery[0].length);\n  }\n\n  if (/^any number of\\b/i.test(s)) {\n    constraint.any_number = true;\n    const m = s.match(/^any number of\\s+(.+?)\\s+can\\b/i);\n    if (m && !/^models?$/i.test(m[1])) constraint.model_name = cleanName(m[1]);\n    s = s.replace(/^any number of\\s+.+?\\s+(?=can\\b)/i, \"\");\n  } else {\n    const upTo = s.match(/^up to (\\d+)\\s+(.+?)\\s+(?=can\\b)/i);\n    if (upTo) {\n      constraint.max_count = parseInt(upTo[1], 10);\n      if (!/^models?$/i.test(upTo[2])) constraint.model_name = cleanName(upTo[2]);\n      s = s.slice(upTo[0].length);\n    } else {\n      // \"The Champion's …\", \"The Kill Team Sergeant can …\"\n      const theRole = s.match(/^the\\s+(.+?)(?:'s|’s)\\s+/i) || s.match(/^the\\s+(.+?)\\s+can\\b/i);\n      if (theRole) {\n        constraint.model_name = cleanName(theRole[1]);\n        if (constraint.max_count === undefined && constraint.per_n_models === undefined) {\n          constraint.max_count = 1;\n        }\n        // Drop only the \"The <role>\" lead-in; keep the possessive weapon for the verb stage.\n        s = s.replace(/^the\\s+/i, \"\");\n      } else {\n        // \"1 model …\", \"1 Khorne Berzerker's …\", \"One model …\"\n        const leadCount = s.match(/^(\\d+|one)\\s+(model|.+?)(?=(?:'s|’s)\\s+|\\s+can\\b)/i);\n        if (leadCount) {\n          const n = /^one$/i.test(leadCount[1]) ? 1 : parseInt(leadCount[1], 10);\n          if (constraint.per_n_models === undefined && constraint.max_count === undefined) {\n            constraint.max_count = n;\n          }\n          if (!/^models?$/i.test(leadCount[2])) constraint.model_name = cleanName(leadCount[2]);\n          s = s.replace(/^(?:\\d+|one)\\s+(?:model\\s+)?/i, \"\");\n        }\n      }\n    }\n  }\n\n  // A bare \"For every N models, 1 <model>'s …\" leaves a leading \"1 <model>\" —\n  // and \"This model\"/\"This unit\"/\"it\" lead-ins carry no constraint.\n  s = s.replace(/^(\\d+)\\s+(?=\\S+(?:'s|’s)\\s)/, \"\");\n  return { constraint, rest: s.trim() };\n}\n\n/** Pull the replaced weapon(s) and the replacement clause out of the core. */\nfunction splitOnVerb(\n  core: string,\n): { replaces: string[]; rest: string } | null {\n  // Active voice: \"… (can )?(each )?replace(s|d) (its|their|one of its/their) WEAPON with REST\"\n  const active = core.match(\n    /\\b(?:can\\s+)?(?:each\\s+)?replaces?\\s+(?:one of\\s+)?(?:its|their)\\s+(.+?)\\s+with\\s*:?\\s*(.+)$/i,\n  );\n  if (active) {\n    return { replaces: splitAnd(active[1]), rest: active[2].trim() };\n  }\n  // \"have/has (its|their) WEAPON replaced with REST\" (\"Any number of Boyz can\n  // each have their slugga and choppa replaced with …\").\n  const havePassive = core.match(\n    /\\b(?:have|has)\\s+(?:its|their)\\s+(.+?)\\s+replaced\\s+with\\s*:?\\s*(.+)$/i,\n  );\n  if (havePassive) {\n    return { replaces: splitAnd(havePassive[1]), rest: havePassive[2].trim() };\n  }\n  // Passive voice: \"… WEAPON can [each] be replaced with REST\" (WEAPON sits after\n  // a possessive 's, or is the whole lead if none).\n  const passive = core.match(/^(.*?)\\s+can (?:each )?be replaced\\s+with\\s*:?\\s*(.+)$/i);\n  if (passive) {\n    const lead = passive[1];\n    const possessive = lead.match(/(?:'s|’s)\\s+(.+)$/);\n    const weapon = possessive ? possessive[1] : lead;\n    return { replaces: splitAnd(weapon), rest: passive[2].trim() };\n  }\n  return null;\n}\n\n/** Pull the added wargear out of an add-on core (\"equipped with …\" / \"can have …\"). */\nfunction splitOnAddVerb(core: string): { rest: string } | null {\n  // Allow an immediate colon (\"equipped with:1 lobba\") as well as a space.\n  const m = core.match(/\\b(?:equipped with|can have)\\s*:?\\s*(.+)$/i);\n  return m ? { rest: m[1].trim() } : null;\n}\n\n/**\n * Parse one option `description`. Returns `{ ok: \"skip\" }` for non-options,\n * `{ ok: false, reason }` for a real option we couldn't parse, and\n * `{ ok: true, option }` otherwise.\n */\nexport function parseOption(description: string | null | undefined): ParseResult {\n  if (description == null) return { ok: \"skip\" };\n  const trimmed = description.trim();\n  if (trimmed === \"\" || trimmed === \"None\") return { ok: \"skip\" };\n  if (trimmed.startsWith(\"*\")) return { ok: \"skip\" };\n  if (/cannot be (replaced|taken|equipped)/i.test(trimmed)) return { ok: \"skip\" };\n\n  const desc = normalize(trimmed);\n  const isSwap = /\\breplace/i.test(desc);\n  const isAddon = !isSwap && /\\b(equipped with|can have)\\b/i.test(desc);\n  if (!isSwap && !isAddon) {\n    return { ok: false, reason: \"no swap/add verb recognised\" };\n  }\n\n  const { constraint, rest: core } = extractConstraint(desc);\n\n  if (isSwap) {\n    const split = splitOnVerb(core);\n    if (!split) return { ok: false, reason: \"could not isolate replaced weapon / replacement\" };\n    if (split.replaces.length === 0) return { ok: false, reason: \"empty replaced-weapon list\" };\n    const repl = parseReplacement(split.rest);\n    if (!repl.replacement?.length && !repl.replacement_choice?.length) {\n      return { ok: false, reason: \"empty replacement\" };\n    }\n    return { ok: true, option: { kind: \"swap\", constraint, replaces: split.replaces, ...repl } };\n  }\n\n  const add = splitOnAddVerb(core);\n  if (!add) return { ok: false, reason: \"could not isolate added wargear\" };\n  const repl = parseReplacement(add.rest);\n  if (!repl.replacement?.length && !repl.replacement_choice?.length) {\n    return { ok: false, reason: \"empty add-on\" };\n  }\n  return { ok: true, option: { kind: \"addon\", constraint, replaces: [], ...repl } };\n}\n"]}

package/dist/converters/wargear-options.d.ts

@@ -0,0 +1,55 @@
+import { type ParsedConstraint } from "./option-parser.js";
+interface SourceOption {
+    datasheet_id: string;
+    line: string;
+    description: string | null;
+}
+interface SourceComposition {
+    datasheet_id: string;
+    line: string;
+    description: string;
+}
+interface SourceModel {
+    datasheet_id: string;
+    name: string;
+}
+interface GameVersion {
+    edition: string;
+    dataslate: string;
+}
+export interface WargearEntity {
+    id: string;
+    name: string;
+    category?: string;
+    game_version: GameVersion;
+}
+export interface WargearOptionEntity {
+    id: string;
+    unit_id: string;
+    model_constraint?: ParsedConstraint;
+    replaces?: string[];
+    replacement?: string[];
+    replacement_choice?: string[][];
+    is_free: boolean;
+    game_version: GameVersion;
+}
+export interface UnparsedOption {
+    unit_id: string | null;
+    datasheet: string;
+    line: string;
+    description: string | null;
+    reason: string;
+}
+export interface BuildWargearResult {
+    wargearOptions: WargearOptionEntity[];
+    wargear: WargearEntity[];
+    unparsed: UnparsedOption[];
+}
+export declare function buildWargearOptions(factionDatasheets: readonly {
+    id: string;
+    name: string;
+}[], allModels: readonly SourceModel[], allOptions: readonly SourceOption[], allComposition: readonly SourceComposition[], unitWeaponIds: Map<string, Set<string>>, // UUID → weapon ids on that unit
+globalWeaponIds: Set<string>, // every weapon id in the faction
+gameVersion: GameVersion): BuildWargearResult;
+export {};
+//# sourceMappingURL=wargear-options.d.ts.map

package/dist/converters/wargear-options.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"wargear-options.d.ts","sourceRoot":"","sources":["../../src/converters/wargear-options.ts"],"names":[],"mappings":"AAYA,OAAO,EAAe,KAAK,gBAAgB,EAAE,MAAM,oBAAoB,CAAC;AAOxE,UAAU,YAAY;IACpB,YAAY,EAAE,MAAM,CAAC;IACrB,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,GAAG,IAAI,CAAC;CAC5B;AACD,UAAU,iBAAiB;IACzB,YAAY,EAAE,MAAM,CAAC;IACrB,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,CAAC;CACrB;AACD,UAAU,WAAW;IACnB,YAAY,EAAE,MAAM,CAAC;IACrB,IAAI,EAAE,MAAM,CAAC;CACd;AAED,UAAU,WAAW;IACnB,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,aAAa;IAC5B,EAAE,EAAE,MAAM,CAAC;IACX,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,WAAW,CAAC;CAC3B;AAED,MAAM,WAAW,mBAAmB;IAClC,EAAE,EAAE,MAAM,CAAC;IACX,OAAO,EAAE,MAAM,CAAC;IAChB,gBAAgB,CAAC,EAAE,gBAAgB,CAAC;IACpC,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,WAAW,CAAC,EAAE,MAAM,EAAE,CAAC;IACvB,kBAAkB,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;IAChC,OAAO,EAAE,OAAO,CAAC;IACjB,YAAY,EAAE,WAAW,CAAC;CAC3B;AAED,MAAM,WAAW,cAAc;IAC7B,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,SAAS,EAAE,MAAM,CAAC;IAClB,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,kBAAkB;IACjC,cAAc,EAAE,mBAAmB,EAAE,CAAC;IACtC,OAAO,EAAE,aAAa,EAAE,CAAC;IACzB,QAAQ,EAAE,cAAc,EAAE,CAAC;CAC5B;AAqBD,wBAAgB,mBAAmB,CACjC,iBAAiB,EAAE,SAAS;IAAE,EAAE,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,EAAE,EAC1D,SAAS,EAAE,SAAS,WAAW,EAAE,EACjC,UAAU,EAAE,SAAS,YAAY,EAAE,EACnC,cAAc,EAAE,SAAS,iBAAiB,EAAE,EAC5C,aAAa,EAAE,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,iCAAiC;AAC1E,eAAe,EAAE,GAAG,CAAC,MAAM,CAAC,EAAE,iCAAiC;AAC/D,WAAW,EAAE,WAAW,GACvB,kBAAkB,CA0JpB"}