@alpaca-software/40kdc-data 0.1.2 → 0.1.3

package/dist/scrub-ip.js

@@ -0,0 +1,88 @@
+/**
+ * IP scrub: remove verbatim Games Workshop rules text from enrichment data.
+ *
+ * The original Haiku authoring pass, when it failed to translate a rule into
+ * the ability DSL, left a stub and dumped the source GW `description` into
+ * `community_notes` as `"… Original: <verbatim rules text>"`. CLAUDE.md forbids
+ * committing GW ability text. This codemod replaces any such note with a
+ * non-infringing citation, leaving genuinely-authored notes (analysis,
+ * `"skipped for damage calc"`, etc.) untouched.
+ *
+ * Idempotent — re-running after a scrub is a no-op. Markers: `"Original:"` (the
+ * dump prefix) and `"■"` (a GW bullet glyph that rides along with copied text).
+ *
+ * Usage: npx tsx tools/src/scrub-ip.ts            (writes in place)
+ *        npx tsx tools/src/scrub-ip.ts --check     (report only, non-zero if leaks remain)
+ */
+import { readdirSync, readFileSync, writeFileSync } from "node:fs";
+import { resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+const __dirname = fileURLToPath(new URL(".", import.meta.url));
+const ENRICHMENT_ROOT = resolve(__dirname, "../../data/enrichment");
+/** The note any GW-text-bearing entry is rewritten to. */
+export const SCRUB_CITATION = "auto-generated stub — needs manual authoring (source: 10e Datasheets_abilities; rules text omitted for IP)";
+/** Markers identifying a note that carries verbatim GW text. */
+const LEAK_MARKER = /Original:|■/;
+/** True when this note carries verbatim GW rules text and must be scrubbed. */
+export function isLeak(note) {
+    return typeof note === "string" && LEAK_MARKER.test(note);
+}
+/** Replace a leaking note with the citation; pass authored notes through unchanged. */
+export function scrubNote(note) {
+    return isLeak(note) ? SCRUB_CITATION : note;
+}
+/** Scrub one file's array in place; return how many entries changed. */
+export function scrubAbilities(abilities) {
+    let changed = 0;
+    for (const a of abilities) {
+        if (isLeak(a.community_notes)) {
+            a.community_notes = SCRUB_CITATION;
+            changed++;
+        }
+    }
+    return changed;
+}
+function run(check) {
+    let totalLeaks = 0;
+    let filesTouched = 0;
+    for (const entry of readdirSync(ENRICHMENT_ROOT, { withFileTypes: true })) {
+        if (!entry.isDirectory() || entry.name === "_example")
+            continue;
+        const file = resolve(ENRICHMENT_ROOT, entry.name, "abilities.json");
+        let abilities;
+        try {
+            abilities = JSON.parse(readFileSync(file, "utf-8"));
+        }
+        catch {
+            continue;
+        }
+        if (!Array.isArray(abilities))
+            continue;
+        const leaks = abilities.filter((a) => isLeak(a.community_notes)).length;
+        if (leaks === 0)
+            continue;
+        totalLeaks += leaks;
+        filesTouched++;
+        if (check) {
+            console.log(`  ${String(leaks).padStart(3)}  ${entry.name}`);
+        }
+        else {
+            scrubAbilities(abilities);
+            writeFileSync(file, JSON.stringify(abilities, null, 2) + "\n");
+            console.log(`  ✓ ${entry.name}: scrubbed ${leaks}`);
+        }
+    }
+    if (check) {
+        console.log(totalLeaks === 0 ? "No GW-text leaks." : `\n${totalLeaks} leak(s) across ${filesTouched} file(s).`);
+        if (totalLeaks > 0)
+            process.exit(1);
+    }
+    else {
+        console.log(`\nScrubbed ${totalLeaks} entr${totalLeaks === 1 ? "y" : "ies"} across ${filesTouched} file(s).`);
+    }
+}
+const isMain = process.argv[1] &&
+    resolve(process.argv[1]).replace(/\.\w+$/, "") === fileURLToPath(import.meta.url).replace(/\.\w+$/, "");
+if (isMain)
+    run(process.argv.includes("--check"));
+//# sourceMappingURL=scrub-ip.js.map

package/dist/scrub-ip.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scrub-ip.js","sourceRoot":"","sources":["../src/scrub-ip.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AACH,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AAEzC,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAC/D,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,EAAE,uBAAuB,CAAC,CAAC;AAEpE,0DAA0D;AAC1D,MAAM,CAAC,MAAM,cAAc,GACzB,4GAA4G,CAAC;AAE/G,gEAAgE;AAChE,MAAM,WAAW,GAAG,aAAa,CAAC;AAElC,+EAA+E;AAC/E,MAAM,UAAU,MAAM,CAAC,IAAwB;IAC7C,OAAO,OAAO,IAAI,KAAK,QAAQ,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC5D,CAAC;AAED,uFAAuF;AACvF,MAAM,UAAU,SAAS,CAAC,IAAwB;IAChD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC;AAC9C,CAAC;AAOD,wEAAwE;AACxE,MAAM,UAAU,cAAc,CAAC,SAAyB;IACtD,IAAI,OAAO,GAAG,CAAC,CAAC;IAChB,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;QAC1B,IAAI,MAAM,CAAC,CAAC,CAAC,eAAe,CAAC,EAAE,CAAC;YAC9B,CAAC,CAAC,eAAe,GAAG,cAAc,CAAC;YACnC,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,GAAG,CAAC,KAAc;IACzB,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,eAAe,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;QAC1E,IAAI,CAAC,KAAK,CAAC,WAAW,EAAE,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU;YAAE,SAAS;QAChE,MAAM,IAAI,GAAG,OAAO,CAAC,eAAe,EAAE,KAAK,CAAC,IAAI,EAAE,gBAAgB,CAAC,CAAC;QACpE,IAAI,SAAyB,CAAC;QAC9B,IAAI,CAAC;YACH,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC;QACtD,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC;YAAE,SAAS;QACxC,MAAM,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,MAAM,CAAC;QACxE,IAAI,KAAK,KAAK,CAAC;YAAE,SAAS;QAC1B,UAAU,IAAI,KAAK,CAAC;QACpB,YAAY,EAAE,CAAC;QACf,IAAI,KAAK,EAAE,CAAC;YACV,OAAO,CAAC,GAAG,CAAC,KAAK,MAAM,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;QAC/D,CAAC;aAAM,CAAC;YACN,cAAc,CAAC,SAAS,CAAC,CAAC;YAC1B,aAAa,CAAC,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;YAC/D,OAAO,CAAC,GAAG,CAAC,OAAO,KAAK,CAAC,IAAI,cAAc,KAAK,EAAE,CAAC,CAAC;QACtD,CAAC;IACH,CAAC;IACD,IAAI,KAAK,EAAE,CAAC;QACV,OAAO,CAAC,GAAG,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,KAAK,UAAU,mBAAmB,YAAY,WAAW,CAAC,CAAC;QAChH,IAAI,UAAU,GAAG,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACtC,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,GAAG,CAAC,cAAc,UAAU,QAAQ,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,WAAW,YAAY,WAAW,CAAC,CAAC;IAChH,CAAC;AACH,CAAC;AAED,MAAM,MAAM,GACV,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;IACf,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,KAAK,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;AAC1G,IAAI,MAAM;IAAE,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC","sourcesContent":["/**\n * IP scrub: remove verbatim Games Workshop rules text from enrichment data.\n *\n * The original Haiku authoring pass, when it failed to translate a rule into\n * the ability DSL, left a stub and dumped the source GW `description` into\n * `community_notes` as `\"… Original: <verbatim rules text>\"`. CLAUDE.md forbids\n * committing GW ability text. This codemod replaces any such note with a\n * non-infringing citation, leaving genuinely-authored notes (analysis,\n * `\"skipped for damage calc\"`, etc.) untouched.\n *\n * Idempotent — re-running after a scrub is a no-op. Markers: `\"Original:\"` (the\n * dump prefix) and `\"■\"` (a GW bullet glyph that rides along with copied text).\n *\n * Usage: npx tsx tools/src/scrub-ip.ts            (writes in place)\n *        npx tsx tools/src/scrub-ip.ts --check     (report only, non-zero if leaks remain)\n */\nimport { readdirSync, readFileSync, writeFileSync } from \"node:fs\";\nimport { resolve } from \"node:path\";\nimport { fileURLToPath } from \"node:url\";\n\nconst __dirname = fileURLToPath(new URL(\".\", import.meta.url));\nconst ENRICHMENT_ROOT = resolve(__dirname, \"../../data/enrichment\");\n\n/** The note any GW-text-bearing entry is rewritten to. */\nexport const SCRUB_CITATION =\n  \"auto-generated stub — needs manual authoring (source: 10e Datasheets_abilities; rules text omitted for IP)\";\n\n/** Markers identifying a note that carries verbatim GW text. */\nconst LEAK_MARKER = /Original:|■/;\n\n/** True when this note carries verbatim GW rules text and must be scrubbed. */\nexport function isLeak(note: string | undefined): boolean {\n  return typeof note === \"string\" && LEAK_MARKER.test(note);\n}\n\n/** Replace a leaking note with the citation; pass authored notes through unchanged. */\nexport function scrubNote(note: string | undefined): string | undefined {\n  return isLeak(note) ? SCRUB_CITATION : note;\n}\n\ninterface AbilityEntry {\n  community_notes?: string;\n  [k: string]: unknown;\n}\n\n/** Scrub one file's array in place; return how many entries changed. */\nexport function scrubAbilities(abilities: AbilityEntry[]): number {\n  let changed = 0;\n  for (const a of abilities) {\n    if (isLeak(a.community_notes)) {\n      a.community_notes = SCRUB_CITATION;\n      changed++;\n    }\n  }\n  return changed;\n}\n\nfunction run(check: boolean): void {\n  let totalLeaks = 0;\n  let filesTouched = 0;\n  for (const entry of readdirSync(ENRICHMENT_ROOT, { withFileTypes: true })) {\n    if (!entry.isDirectory() || entry.name === \"_example\") continue;\n    const file = resolve(ENRICHMENT_ROOT, entry.name, \"abilities.json\");\n    let abilities: AbilityEntry[];\n    try {\n      abilities = JSON.parse(readFileSync(file, \"utf-8\"));\n    } catch {\n      continue;\n    }\n    if (!Array.isArray(abilities)) continue;\n    const leaks = abilities.filter((a) => isLeak(a.community_notes)).length;\n    if (leaks === 0) continue;\n    totalLeaks += leaks;\n    filesTouched++;\n    if (check) {\n      console.log(`  ${String(leaks).padStart(3)}  ${entry.name}`);\n    } else {\n      scrubAbilities(abilities);\n      writeFileSync(file, JSON.stringify(abilities, null, 2) + \"\\n\");\n      console.log(`  ✓ ${entry.name}: scrubbed ${leaks}`);\n    }\n  }\n  if (check) {\n    console.log(totalLeaks === 0 ? \"No GW-text leaks.\" : `\\n${totalLeaks} leak(s) across ${filesTouched} file(s).`);\n    if (totalLeaks > 0) process.exit(1);\n  } else {\n    console.log(`\\nScrubbed ${totalLeaks} entr${totalLeaks === 1 ? \"y\" : \"ies\"} across ${filesTouched} file(s).`);\n  }\n}\n\nconst isMain =\n  process.argv[1] &&\n  resolve(process.argv[1]).replace(/\\.\\w+$/, \"\") === fileURLToPath(import.meta.url).replace(/\\.\\w+$/, \"\");\nif (isMain) run(process.argv.includes(\"--check\"));\n"]}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@alpaca-software/40kdc-data",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "type": "module",
   "description": "The 40kdc Warhammer 40K dataset behind a linked, typed API — find units, follow them to their weapons, abilities, phases, and factions. Also validates data against the canonical JSON Schemas.",
   "keywords": [
@@ -37,7 +37,8 @@
     "README.md"
   ],
   "bin": {
-    "40kdc-validate": "./dist/cli.js"
+    "40kdc-validate": "./dist/cli.js",
+    "40kdc-runner": "./dist/runner.js"
   },
   "publishConfig": {
     "access": "public",
@@ -56,6 +57,12 @@
     "validate:core": "tsx src/cli.ts validate-core",
     "validate:enrichment": "tsx src/cli.ts validate-enrichment",
     "translate": "tsx src/cli.ts translate",
+    "audit:coverage": "tsx src/cli.ts audit-coverage --write",
+    "author:input": "tsx src/author-input.ts",
+    "author:propose": "tsx src/author-batch.ts propose",
+    "author:repair": "tsx src/author-batch.ts repair",
+    "author:apply": "tsx src/author-batch.ts apply",
+    "author:review": "tsx src/author-batch.ts review",
     "bundle:schemas": "tsx src/bundle-schemas.ts",
     "codegen:types": "json2ts -i ../crates/wh40kdc/schemas/bundled.schema.json -o src/generated.ts --unreachableDefinitions --bannerComment \"/* Generated from crates/wh40kdc/schemas/bundled.schema.json by 'npm run codegen:types'. DO NOT EDIT BY HAND. */\"",
     "copy:schemas": "rm -rf schemas && cp -R ../schemas schemas",

package/schemas/core/roster.schema.json

@@ -17,7 +17,9 @@
             "newrecruit-json",
             "newrecruit-wtc-compact",
             "newrecruit-wtc-full",
-            "newrecruit-simple"
+            "newrecruit-simple",
+            "rosterizer",
+            "gw"
           ]
         },
         "generated_by": {