@alpaca-software/40kdc-data 0.5.12 → 0.5.15

package/dist/audit-phrasing.js

@@ -0,0 +1,208 @@
+/**
+ * Phrasing audit: catalogue the generated English text for *every* authored
+ * ability and flag mechanical phrasing defects, so an across-the-board prose
+ * cleanup has a worklist.
+ *
+ * The DSL→English describer (`describeAbility`, a.k.a. `ability.print()`) is the
+ * user-facing rules readout in downstream consumers. This tool runs that *real*
+ * describer over each `data/enrichment/<faction>/abilities.json` entry and runs
+ * a set of regex heuristics over the output, flagging the defect classes a
+ * manual audit turned up: leftover `?` placeholders (modifier field-name
+ * mismatches), unmapped `[type]` fallbacks, raw `snake_case` identifiers that
+ * leaked into prose, doubled prepositions ("at on model destroyed"), `0"` noise
+ * on movement grants, `(s)` pluralization stubs, and empty output.
+ *
+ * This is a TS-only analysis utility — it *consumes* the conformance-pinned
+ * describer, it does not change it — so it needs no SPEC_VERSION bump and no
+ * Rust/Python port (mirrors `audit-coverage.ts`).
+ *
+ * Usage:
+ *   npx tsx tools/src/audit-phrasing.ts            (pretty report to stdout)
+ *   npx tsx tools/src/audit-phrasing.ts --json     (machine-readable)
+ *   npx tsx tools/src/audit-phrasing.ts --write     (also emit data/_audit/phrasing.*)
+ *
+ * Wired as `40kdc-validate audit-phrasing` and `npm run audit:phrasing`.
+ */
+import { readdirSync, readFileSync, writeFileSync, mkdirSync } from "node:fs";
+import { resolve, basename } from "node:path";
+import { fileURLToPath } from "node:url";
+import chalk from "chalk";
+import { describeAbility } from "./translate/index.js";
+const __dirname = fileURLToPath(new URL(".", import.meta.url));
+const DATA_ROOT = resolve(__dirname, "../../data");
+const ENRICHMENT_ROOT = resolve(DATA_ROOT, "enrichment");
+const AUDIT_DIR = resolve(DATA_ROOT, "_audit");
+const PREP = "(?:at|on|in|of|to|by|for|from|with)";
+const DOUBLED_PREP = new RegExp(`\\b${PREP}\\s+${PREP}\\b`, "i");
+// The unknown-effect-type fallback emits a lowercase kebab token in brackets
+// (`[some-type]`); intentional GW weapon keywords are bracketed UPPERCASE
+// (`[LETHAL HITS]`) and must NOT be flagged.
+const TYPE_FALLBACK = /\[[a-z][a-z0-9-]*\]/;
+const SNAKE_CASE = /[a-z]+_[a-z]+/;
+const ZERO_INCH = /\b0"/;
+/**
+ * Classify one piece of generated text. Pure and order-stable so it is directly
+ * unit-testable. The order of the returned flags follows the {@link PhrasingFlag}
+ * union for deterministic output.
+ */
+export function classifyPhrasing(text) {
+    const flags = [];
+    if (text.trim() === "") {
+        flags.push("empty");
+        return flags;
+    }
+    if (text.includes("?"))
+        flags.push("placeholder");
+    if (TYPE_FALLBACK.test(text))
+        flags.push("type_fallback");
+    if (SNAKE_CASE.test(text))
+        flags.push("snake_case_leftover");
+    if (DOUBLED_PREP.test(text))
+        flags.push("doubled_preposition");
+    if (ZERO_INCH.test(text))
+        flags.push("zero_inch");
+    if (text.includes("(s)"))
+        flags.push("paren_plural");
+    return flags;
+}
+/** Compute the phrasing report for a set of factions. Pure — IO lives in the command. */
+export function auditPhrasing(input) {
+    const rows = [];
+    const byFaction = [];
+    const flagCounts = new Map();
+    for (const { faction, abilities } of input) {
+        let factionFlagged = 0;
+        for (const a of abilities) {
+            const text = describeAbility({
+                ...(a.effect ? { effect: a.effect } : {}),
+                ...(a.scope ? { scope: a.scope } : {}),
+                ...(a.applies_to ? { applies_to: a.applies_to } : {}),
+            });
+            const flags = classifyPhrasing(text);
+            if (flags.length > 0) {
+                factionFlagged++;
+                for (const f of flags)
+                    flagCounts.set(f, (flagCounts.get(f) ?? 0) + 1);
+            }
+            rows.push({
+                faction,
+                ability_id: a.ability_id,
+                name: a.name ?? a.ability_id,
+                unit_ids: a.unit_ids ?? [],
+                text,
+                flags,
+            });
+        }
+        byFaction.push({ faction, total: abilities.length, flagged: factionFlagged });
+    }
+    byFaction.sort((a, b) => a.faction.localeCompare(b.faction));
+    rows.sort((a, b) => a.faction.localeCompare(b.faction) || a.ability_id.localeCompare(b.ability_id));
+    const byFlag = [...flagCounts.entries()]
+        .map(([flag, count]) => ({ flag, count }))
+        .sort((a, b) => b.count - a.count || a.flag.localeCompare(b.flag));
+    return {
+        rows,
+        total: rows.length,
+        flagged: rows.filter((r) => r.flags.length > 0).length,
+        byFlag,
+        byFaction,
+    };
+}
+/** Read every `data/enrichment/<faction>/abilities.json` (skips `_`-prefixed dirs). */
+export function loadFactions() {
+    const out = [];
+    for (const entry of readdirSync(ENRICHMENT_ROOT, { withFileTypes: true })) {
+        if (!entry.isDirectory() || entry.name.startsWith("_"))
+            continue;
+        const file = resolve(ENRICHMENT_ROOT, entry.name, "abilities.json");
+        let abilities;
+        try {
+            abilities = JSON.parse(readFileSync(file, "utf-8"));
+        }
+        catch {
+            continue; // no abilities.json in this folder
+        }
+        if (!Array.isArray(abilities))
+            continue;
+        out.push({ faction: entry.name, abilities });
+    }
+    return out;
+}
+const pct = (n, d) => (d === 0 ? "—" : `${Math.round((100 * n) / d)}%`);
+/** CSV-escape one field (RFC-4180: wrap in quotes, double internal quotes). */
+function csvField(v) {
+    return `"${v.replace(/"/g, '""')}"`;
+}
+/**
+ * One row per ability. The multi-line describer output is flattened to ` / ` so
+ * each catalogued ability stays on a single CSV line (greppable); the JSON
+ * artifact keeps the original newlines.
+ */
+function csvReport(r) {
+    const lines = ["faction,ability_id,name,unit_ids,flags,text"];
+    for (const row of r.rows) {
+        lines.push([
+            csvField(row.faction),
+            csvField(row.ability_id),
+            csvField(row.name),
+            csvField(row.unit_ids.join(";")),
+            csvField(row.flags.join(";")),
+            csvField(row.text.replace(/\n/g, " / ")),
+        ].join(","));
+    }
+    return lines.join("\n") + "\n";
+}
+function prettyReport(r) {
+    const lines = [];
+    lines.push("");
+    lines.push(chalk.bold("40kdc Ability Phrasing Audit"));
+    lines.push(chalk.gray("─".repeat(60)));
+    lines.push(`Abilities catalogued: ${chalk.cyan(String(r.total))}`);
+    lines.push(`Flagged for review:   ${chalk.yellow(String(r.flagged))} (${pct(r.flagged, r.total)})`);
+    lines.push("");
+    lines.push(chalk.bold("Defect flags:"));
+    if (r.byFlag.length === 0) {
+        lines.push(chalk.green("  none — every ability's generated text is clean"));
+    }
+    else {
+        for (const { flag, count } of r.byFlag) {
+            lines.push(`  ${String(count).padStart(5)}  ${chalk.gray(flag)}`);
+        }
+    }
+    lines.push("");
+    lines.push(chalk.bold("Most-flagged factions:"));
+    const worst = [...r.byFaction].filter((f) => f.flagged > 0).sort((a, b) => b.flagged - a.flagged);
+    for (const f of worst.slice(0, 15)) {
+        lines.push(`  ${String(f.flagged).padStart(4)}/${String(f.total).padEnd(4)}  ${chalk.gray(f.faction)}`);
+    }
+    lines.push("");
+    return lines.join("\n");
+}
+/** Commander action: run the phrasing audit over the repo's enrichment data. */
+export function auditPhrasingCommand(opts = {}) {
+    const report = auditPhrasing(loadFactions());
+    if (opts.write) {
+        mkdirSync(AUDIT_DIR, { recursive: true });
+        writeFileSync(resolve(AUDIT_DIR, "phrasing.json"), JSON.stringify(report, null, 2) + "\n");
+        writeFileSync(resolve(AUDIT_DIR, "phrasing.csv"), csvReport(report));
+    }
+    if (opts.reporter === "json") {
+        console.log(JSON.stringify(report, null, 2));
+    }
+    else {
+        console.log(prettyReport(report));
+        if (opts.write)
+            console.log(chalk.gray(`Wrote ${basename(AUDIT_DIR)}/phrasing.csv + phrasing.json`));
+    }
+}
+// Direct-invocation entry point (`npx tsx tools/src/audit-phrasing.ts [--json] [--write]`).
+const isMain = process.argv[1] &&
+    resolve(process.argv[1]).replace(/\.\w+$/, "") === fileURLToPath(import.meta.url).replace(/\.\w+$/, "");
+if (isMain) {
+    const argv = process.argv.slice(2);
+    auditPhrasingCommand({
+        reporter: argv.includes("--json") ? "json" : "pretty",
+        write: argv.includes("--write"),
+    });
+}
+//# sourceMappingURL=audit-phrasing.js.map

package/dist/audit-phrasing.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"audit-phrasing.js","sourceRoot":"","sources":["../src/audit-phrasing.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AAC9E,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,WAAW,CAAC;AAC9C,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,eAAe,EAAyD,MAAM,sBAAsB,CAAC;AAE9G,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAC/D,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,EAAE,YAAY,CAAC,CAAC;AACnD,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,EAAE,YAAY,CAAC,CAAC;AACzD,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,EAAE,QAAQ,CAAC,CAAC;AA6C/C,MAAM,IAAI,GAAG,qCAAqC,CAAC;AACnD,MAAM,YAAY,GAAG,IAAI,MAAM,CAAC,MAAM,IAAI,OAAO,IAAI,KAAK,EAAE,GAAG,CAAC,CAAC;AACjE,6EAA6E;AAC7E,0EAA0E;AAC1E,6CAA6C;AAC7C,MAAM,aAAa,GAAG,qBAAqB,CAAC;AAC5C,MAAM,UAAU,GAAG,eAAe,CAAC;AACnC,MAAM,SAAS,GAAG,MAAM,CAAC;AAEzB;;;;GAIG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,MAAM,KAAK,GAAmB,EAAE,CAAC;IACjC,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;QACvB,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACpB,OAAO,KAAK,CAAC;IACf,CAAC;IACD,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;IAClD,IAAI,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;IAC1D,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC;IAC7D,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC;IAC/D,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;IAClD,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IACrD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,yFAAyF;AACzF,MAAM,UAAU,aAAa,CAC3B,KAAuD;IAEvD,MAAM,IAAI,GAAkB,EAAE,CAAC;IAC/B,MAAM,SAAS,GAA0D,EAAE,CAAC;IAC5E,MAAM,UAAU,GAAG,IAAI,GAAG,EAAwB,CAAC;IAEnD,KAAK,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,IAAI,KAAK,EAAE,CAAC;QAC3C,IAAI,cAAc,GAAG,CAAC,CAAC;QACvB,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;YAC1B,MAAM,IAAI,GAAG,eAAe,CAAC;gBAC3B,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACzC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACtC,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACtD,CAAC,CAAC;YACH,MAAM,KAAK,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;YACrC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,cAAc,EAAE,CAAC;gBACjB,KAAK,MAAM,CAAC,IAAI,KAAK;oBAAE,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACzE,CAAC;YACD,IAAI,CAAC,IAAI,CAAC;gBACR,OAAO;gBACP,UAAU,EAAE,CAAC,CAAC,UAAU;gBACxB,IAAI,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,UAAU;gBAC5B,QAAQ,EAAE,CAAC,CAAC,QAAQ,IAAI,EAAE;gBAC1B,IAAI;gBACJ,KAAK;aACN,CAAC,CAAC;QACL,CAAC;QACD,SAAS,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,SAAS,CAAC,MAAM,EAAE,OAAO,EAAE,cAAc,EAAE,CAAC,CAAC;IAChF,CAAC;IAED,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;IAC7D,IAAI,CAAC,IAAI,CACP,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC,UAAU,CAAC,CACzF,CAAC;IAEF,MAAM,MAAM,GAAG,CAAC,GAAG,UAAU,CAAC,OAAO,EAAE,CAAC;SACrC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;SACzC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAErE,OAAO;QACL,IAAI;QACJ,KAAK,EAAE,IAAI,CAAC,MAAM;QAClB,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM;QACtD,MAAM;QACN,SAAS;KACV,CAAC;AACJ,CAAC;AAED,uFAAuF;AACvF,MAAM,UAAU,YAAY;IAC1B,MAAM,GAAG,GAAqD,EAAE,CAAC;IACjE,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,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QACjE,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,CAAC,mCAAmC;QAC/C,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC;YAAE,SAAS;QACxC,GAAG,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,CAAC,CAAC;IAC/C,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,GAAG,GAAG,CAAC,CAAS,EAAE,CAAS,EAAU,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;AAEhG,+EAA+E;AAC/E,SAAS,QAAQ,CAAC,CAAS;IACzB,OAAO,IAAI,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC;AACtC,CAAC;AAED;;;;GAIG;AACH,SAAS,SAAS,CAAC,CAAiB;IAClC,MAAM,KAAK,GAAG,CAAC,6CAA6C,CAAC,CAAC;IAC9D,KAAK,MAAM,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CACR;YACE,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC;YACrB,QAAQ,CAAC,GAAG,CAAC,UAAU,CAAC;YACxB,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC;YAClB,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAChC,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAC7B,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;SACzC,CAAC,IAAI,CAAC,GAAG,CAAC,CACZ,CAAC;IACJ,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC;AACjC,CAAC;AAED,SAAS,YAAY,CAAC,CAAiB;IACrC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACf,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,8BAA8B,CAAC,CAAC,CAAC;IACvD,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACvC,KAAK,CAAC,IAAI,CAAC,yBAAyB,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;IACnE,KAAK,CAAC,IAAI,CAAC,yBAAyB,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACf,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,CAAC;IACxC,IAAI,CAAC,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC1B,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,kDAAkD,CAAC,CAAC,CAAC;IAC9E,CAAC;SAAM,CAAC;QACN,KAAK,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC;YACvC,KAAK,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACpE,CAAC;IACH,CAAC;IACD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACf,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,wBAAwB,CAAC,CAAC,CAAC;IACjD,MAAM,KAAK,GAAG,CAAC,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC;IAClG,KAAK,MAAM,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC;QACnC,KAAK,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IAC1G,CAAC;IACD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACf,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAOD,gFAAgF;AAChF,MAAM,UAAU,oBAAoB,CAAC,OAA6B,EAAE;IAClE,MAAM,MAAM,GAAG,aAAa,CAAC,YAAY,EAAE,CAAC,CAAC;IAE7C,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;QACf,SAAS,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1C,aAAa,CAAC,OAAO,CAAC,SAAS,EAAE,eAAe,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;QAC3F,aAAa,CAAC,OAAO,CAAC,SAAS,EAAE,cAAc,CAAC,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;IACvE,CAAC;IAED,IAAI,IAAI,CAAC,QAAQ,KAAK,MAAM,EAAE,CAAC;QAC7B,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IAC/C,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC;QAClC,IAAI,IAAI,CAAC,KAAK;YAAE,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,QAAQ,CAAC,SAAS,CAAC,+BAA+B,CAAC,CAAC,CAAC;IACvG,CAAC;AACH,CAAC;AAED,4FAA4F;AAC5F,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,EAAE,CAAC;IACX,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACnC,oBAAoB,CAAC;QACnB,QAAQ,EAAE,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ;QACrD,KAAK,EAAE,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC;KAChC,CAAC,CAAC;AACL,CAAC","sourcesContent":["/**\n * Phrasing audit: catalogue the generated English text for *every* authored\n * ability and flag mechanical phrasing defects, so an across-the-board prose\n * cleanup has a worklist.\n *\n * The DSL→English describer (`describeAbility`, a.k.a. `ability.print()`) is the\n * user-facing rules readout in downstream consumers. This tool runs that *real*\n * describer over each `data/enrichment/<faction>/abilities.json` entry and runs\n * a set of regex heuristics over the output, flagging the defect classes a\n * manual audit turned up: leftover `?` placeholders (modifier field-name\n * mismatches), unmapped `[type]` fallbacks, raw `snake_case` identifiers that\n * leaked into prose, doubled prepositions (\"at on model destroyed\"), `0\"` noise\n * on movement grants, `(s)` pluralization stubs, and empty output.\n *\n * This is a TS-only analysis utility — it *consumes* the conformance-pinned\n * describer, it does not change it — so it needs no SPEC_VERSION bump and no\n * Rust/Python port (mirrors `audit-coverage.ts`).\n *\n * Usage:\n *   npx tsx tools/src/audit-phrasing.ts            (pretty report to stdout)\n *   npx tsx tools/src/audit-phrasing.ts --json     (machine-readable)\n *   npx tsx tools/src/audit-phrasing.ts --write     (also emit data/_audit/phrasing.*)\n *\n * Wired as `40kdc-validate audit-phrasing` and `npm run audit:phrasing`.\n */\nimport { readdirSync, readFileSync, writeFileSync, mkdirSync } from \"node:fs\";\nimport { resolve, basename } from \"node:path\";\nimport { fileURLToPath } from \"node:url\";\nimport chalk from \"chalk\";\nimport { describeAbility, type Effect, type AbilityScope, type AbilityAppliesTo } from \"./translate/index.js\";\n\nconst __dirname = fileURLToPath(new URL(\".\", import.meta.url));\nconst DATA_ROOT = resolve(__dirname, \"../../data\");\nconst ENRICHMENT_ROOT = resolve(DATA_ROOT, \"enrichment\");\nconst AUDIT_DIR = resolve(DATA_ROOT, \"_audit\");\n\n/** Shape of an authored ability entry (only the fields the describer reads). */\ninterface AbilityEntry {\n  ability_id: string;\n  name?: string;\n  effect?: Effect;\n  scope?: AbilityScope;\n  applies_to?: AbilityAppliesTo | null;\n  /** Units that carry this ability — surfaced so a worklist row points at the datasheet. */\n  unit_ids?: string[];\n}\n\n/** One catalogued ability: its identity, the generated text, and any defect flags. */\nexport interface PhrasingRow {\n  faction: string;\n  ability_id: string;\n  name: string;\n  /** Slugs of the units carrying this ability (for datasheet/Wahapedia lookup). */\n  unit_ids: string[];\n  /** Full multi-line `describeAbility` output (effect + scope + applies). */\n  text: string;\n  flags: PhrasingFlag[];\n}\n\nexport interface PhrasingReport {\n  rows: PhrasingRow[];\n  total: number;\n  /** Rows carrying ≥1 flag. */\n  flagged: number;\n  /** Per-flag counts, descending. */\n  byFlag: { flag: PhrasingFlag; count: number }[];\n  /** Per-faction totals + flagged counts, faction-sorted. */\n  byFaction: { faction: string; total: number; flagged: number }[];\n}\n\nexport type PhrasingFlag =\n  | \"placeholder\"\n  | \"type_fallback\"\n  | \"snake_case_leftover\"\n  | \"doubled_preposition\"\n  | \"zero_inch\"\n  | \"paren_plural\"\n  | \"empty\";\n\nconst PREP = \"(?:at|on|in|of|to|by|for|from|with)\";\nconst DOUBLED_PREP = new RegExp(`\\\\b${PREP}\\\\s+${PREP}\\\\b`, \"i\");\n// The unknown-effect-type fallback emits a lowercase kebab token in brackets\n// (`[some-type]`); intentional GW weapon keywords are bracketed UPPERCASE\n// (`[LETHAL HITS]`) and must NOT be flagged.\nconst TYPE_FALLBACK = /\\[[a-z][a-z0-9-]*\\]/;\nconst SNAKE_CASE = /[a-z]+_[a-z]+/;\nconst ZERO_INCH = /\\b0\"/;\n\n/**\n * Classify one piece of generated text. Pure and order-stable so it is directly\n * unit-testable. The order of the returned flags follows the {@link PhrasingFlag}\n * union for deterministic output.\n */\nexport function classifyPhrasing(text: string): PhrasingFlag[] {\n  const flags: PhrasingFlag[] = [];\n  if (text.trim() === \"\") {\n    flags.push(\"empty\");\n    return flags;\n  }\n  if (text.includes(\"?\")) flags.push(\"placeholder\");\n  if (TYPE_FALLBACK.test(text)) flags.push(\"type_fallback\");\n  if (SNAKE_CASE.test(text)) flags.push(\"snake_case_leftover\");\n  if (DOUBLED_PREP.test(text)) flags.push(\"doubled_preposition\");\n  if (ZERO_INCH.test(text)) flags.push(\"zero_inch\");\n  if (text.includes(\"(s)\")) flags.push(\"paren_plural\");\n  return flags;\n}\n\n/** Compute the phrasing report for a set of factions. Pure — IO lives in the command. */\nexport function auditPhrasing(\n  input: { faction: string; abilities: AbilityEntry[] }[],\n): PhrasingReport {\n  const rows: PhrasingRow[] = [];\n  const byFaction: { faction: string; total: number; flagged: number }[] = [];\n  const flagCounts = new Map<PhrasingFlag, number>();\n\n  for (const { faction, abilities } of input) {\n    let factionFlagged = 0;\n    for (const a of abilities) {\n      const text = describeAbility({\n        ...(a.effect ? { effect: a.effect } : {}),\n        ...(a.scope ? { scope: a.scope } : {}),\n        ...(a.applies_to ? { applies_to: a.applies_to } : {}),\n      });\n      const flags = classifyPhrasing(text);\n      if (flags.length > 0) {\n        factionFlagged++;\n        for (const f of flags) flagCounts.set(f, (flagCounts.get(f) ?? 0) + 1);\n      }\n      rows.push({\n        faction,\n        ability_id: a.ability_id,\n        name: a.name ?? a.ability_id,\n        unit_ids: a.unit_ids ?? [],\n        text,\n        flags,\n      });\n    }\n    byFaction.push({ faction, total: abilities.length, flagged: factionFlagged });\n  }\n\n  byFaction.sort((a, b) => a.faction.localeCompare(b.faction));\n  rows.sort(\n    (a, b) => a.faction.localeCompare(b.faction) || a.ability_id.localeCompare(b.ability_id),\n  );\n\n  const byFlag = [...flagCounts.entries()]\n    .map(([flag, count]) => ({ flag, count }))\n    .sort((a, b) => b.count - a.count || a.flag.localeCompare(b.flag));\n\n  return {\n    rows,\n    total: rows.length,\n    flagged: rows.filter((r) => r.flags.length > 0).length,\n    byFlag,\n    byFaction,\n  };\n}\n\n/** Read every `data/enrichment/<faction>/abilities.json` (skips `_`-prefixed dirs). */\nexport function loadFactions(): { faction: string; abilities: AbilityEntry[] }[] {\n  const out: { faction: string; abilities: AbilityEntry[] }[] = [];\n  for (const entry of readdirSync(ENRICHMENT_ROOT, { withFileTypes: true })) {\n    if (!entry.isDirectory() || entry.name.startsWith(\"_\")) 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; // no abilities.json in this folder\n    }\n    if (!Array.isArray(abilities)) continue;\n    out.push({ faction: entry.name, abilities });\n  }\n  return out;\n}\n\nconst pct = (n: number, d: number): string => (d === 0 ? \"—\" : `${Math.round((100 * n) / d)}%`);\n\n/** CSV-escape one field (RFC-4180: wrap in quotes, double internal quotes). */\nfunction csvField(v: string): string {\n  return `\"${v.replace(/\"/g, '\"\"')}\"`;\n}\n\n/**\n * One row per ability. The multi-line describer output is flattened to ` / ` so\n * each catalogued ability stays on a single CSV line (greppable); the JSON\n * artifact keeps the original newlines.\n */\nfunction csvReport(r: PhrasingReport): string {\n  const lines = [\"faction,ability_id,name,unit_ids,flags,text\"];\n  for (const row of r.rows) {\n    lines.push(\n      [\n        csvField(row.faction),\n        csvField(row.ability_id),\n        csvField(row.name),\n        csvField(row.unit_ids.join(\";\")),\n        csvField(row.flags.join(\";\")),\n        csvField(row.text.replace(/\\n/g, \" / \")),\n      ].join(\",\"),\n    );\n  }\n  return lines.join(\"\\n\") + \"\\n\";\n}\n\nfunction prettyReport(r: PhrasingReport): string {\n  const lines: string[] = [];\n  lines.push(\"\");\n  lines.push(chalk.bold(\"40kdc Ability Phrasing Audit\"));\n  lines.push(chalk.gray(\"─\".repeat(60)));\n  lines.push(`Abilities catalogued: ${chalk.cyan(String(r.total))}`);\n  lines.push(`Flagged for review:   ${chalk.yellow(String(r.flagged))} (${pct(r.flagged, r.total)})`);\n  lines.push(\"\");\n  lines.push(chalk.bold(\"Defect flags:\"));\n  if (r.byFlag.length === 0) {\n    lines.push(chalk.green(\"  none — every ability's generated text is clean\"));\n  } else {\n    for (const { flag, count } of r.byFlag) {\n      lines.push(`  ${String(count).padStart(5)}  ${chalk.gray(flag)}`);\n    }\n  }\n  lines.push(\"\");\n  lines.push(chalk.bold(\"Most-flagged factions:\"));\n  const worst = [...r.byFaction].filter((f) => f.flagged > 0).sort((a, b) => b.flagged - a.flagged);\n  for (const f of worst.slice(0, 15)) {\n    lines.push(`  ${String(f.flagged).padStart(4)}/${String(f.total).padEnd(4)}  ${chalk.gray(f.faction)}`);\n  }\n  lines.push(\"\");\n  return lines.join(\"\\n\");\n}\n\nexport interface AuditPhrasingOptions {\n  reporter?: \"pretty\" | \"json\";\n  write?: boolean;\n}\n\n/** Commander action: run the phrasing audit over the repo's enrichment data. */\nexport function auditPhrasingCommand(opts: AuditPhrasingOptions = {}): void {\n  const report = auditPhrasing(loadFactions());\n\n  if (opts.write) {\n    mkdirSync(AUDIT_DIR, { recursive: true });\n    writeFileSync(resolve(AUDIT_DIR, \"phrasing.json\"), JSON.stringify(report, null, 2) + \"\\n\");\n    writeFileSync(resolve(AUDIT_DIR, \"phrasing.csv\"), csvReport(report));\n  }\n\n  if (opts.reporter === \"json\") {\n    console.log(JSON.stringify(report, null, 2));\n  } else {\n    console.log(prettyReport(report));\n    if (opts.write) console.log(chalk.gray(`Wrote ${basename(AUDIT_DIR)}/phrasing.csv + phrasing.json`));\n  }\n}\n\n// Direct-invocation entry point (`npx tsx tools/src/audit-phrasing.ts [--json] [--write]`).\nconst isMain =\n  process.argv[1] &&\n  resolve(process.argv[1]).replace(/\\.\\w+$/, \"\") === fileURLToPath(import.meta.url).replace(/\\.\\w+$/, \"\");\nif (isMain) {\n  const argv = process.argv.slice(2);\n  auditPhrasingCommand({\n    reporter: argv.includes(\"--json\") ? \"json\" : \"pretty\",\n    write: argv.includes(\"--write\"),\n  });\n}\n"]}

package/dist/author-input-pack.d.ts

@@ -0,0 +1,27 @@
+import { type Block } from "./pack-blocks.js";
+import type { AuthorInputEntry } from "./author-input.js";
+/** The start of one card within a detachment: its id, name, column, and top y. */
+export interface CardAnchor {
+    id: string;
+    name: string;
+    kind: "detachment" | "stratagem" | "enhancement";
+    x: number;
+    topGy: number;
+}
+/**
+ * Locate every card-start anchor in a detachment's blocks: the detachment rule
+ * (under a DETACHMENT RULE(S) header), each stratagem (its name block above a
+ * "<NAME> – <TYPE> STRATAGEM" label), and each enhancement (a caps header with a
+ * prose body below). Anchors bound each other's bodies by same-column y order.
+ */
+export declare function findAnchors(blocks: Block[]): CardAnchor[];
+/**
+ * Capture a card's rule body: every same-column block strictly below the anchor's
+ * name/header, up to the next same-column anchor (the next card's start). Joined
+ * in reading order. Column-scoping is what keeps the adjacent column's text out —
+ * the whole reason bbox beats `-layout` on these 2-column pages.
+ */
+export declare function captureBody(blocks: Block[], anchor: CardAnchor, anchors: CardAnchor[]): string;
+/** Build author-input entries for one faction's new-detachment stubs from its pack. */
+export declare function buildFactionFromPack(faction: string, pdf: string): AuthorInputEntry[];
+//# sourceMappingURL=author-input-pack.d.ts.map

package/dist/author-input-pack.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"author-input-pack.d.ts","sourceRoot":"","sources":["../src/author-input-pack.ts"],"names":[],"mappings":"AA8BA,OAAO,EACL,KAAK,KAAK,EAOX,MAAM,kBAAkB,CAAC;AAG1B,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAC;AAW1D,kFAAkF;AAClF,MAAM,WAAW,UAAU;IACzB,EAAE,EAAE,MAAM,CAAC;IACX,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,YAAY,GAAG,WAAW,GAAG,aAAa,CAAC;IACjD,CAAC,EAAE,MAAM,CAAC;IACV,KAAK,EAAE,MAAM,CAAC;CACf;AAkBD;;;;;GAKG;AACH,wBAAgB,WAAW,CAAC,MAAM,EAAE,KAAK,EAAE,GAAG,UAAU,EAAE,CAgDzD;AAED;;;;;GAKG;AACH,wBAAgB,WAAW,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,UAAU,EAAE,OAAO,EAAE,UAAU,EAAE,GAAG,MAAM,CAa9F;AAID,uFAAuF;AACvF,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,MAAM,EAAE,GAAG,EAAE,MAAM,GAAG,gBAAgB,EAAE,CAmDrF"}

package/dist/author-input-pack.js

@@ -0,0 +1,198 @@
+/**
+ * Build the DSL-authoring input for *new 11e detachments* from a Faction Pack PDF.
+ *
+ * The stock `author-input.ts` resolves each empty-modifier ability stub to its
+ * source rule in the **10e army-assist archive** (unit→datasheet→ability chain).
+ * That archive predates the 11e Faction Packs, so the new detachments' rules /
+ * stratagems / enhancements aren't in it — the chain returns nothing and the
+ * authoring pipeline has no text to classify.
+ *
+ * This sibling tool fills the gap from the *primary* 11e source: the pack PDF.
+ * It captures each card's rule body by bbox coordinate (column-scoped, so the
+ * packs' 2-column stratagem layout never interleaves — the failure mode that
+ * killed the `-layout` approach) and writes the identical `AuthorInputEntry`
+ * shape that `author-batch.ts propose` consumes. Nothing downstream changes.
+ *
+ * IP firewall: rule prose is captured ONLY into the gitignored authoring worklist
+ * `data/_audit/author-input/<faction>.json` (.gitignore). It is never written to
+ * the committed `faction-pack-input/` staging (that stays names + metadata only).
+ *
+ * Scope: only empty-modifier stubs that belong to a detachment (`detachment_id`
+ * set) — i.e. the new-detachment rules/stratagems/enhancements the pack covers.
+ * Unit-scoped stubs (`detachment_id` null) remain the 10e-archive path's job.
+ *
+ * Requires `pdftotext` (poppler) on PATH — authoring-time only, not CI.
+ *
+ * Usage: npx tsx tools/src/author-input-pack.ts --pdf <pack.pdf> --faction <id>
+ */
+import { readFileSync, writeFileSync, mkdirSync, existsSync } from "node:fs";
+import { resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+import { isCapsHeader, sameColumn, stripCp, slug, titleCase, normCaps, } from "./pack-blocks.js";
+import { detachmentSegments } from "./extract-faction-pack.js";
+import { hasEmptyModifier } from "./audit-coverage.js";
+const __dirname = fileURLToPath(new URL(".", import.meta.url));
+const DATA_ROOT = resolve(__dirname, "../../data");
+const ENRICHMENT_ROOT = resolve(DATA_ROOT, "enrichment");
+const OUT_DIR = resolve(DATA_ROOT, "_audit", "author-input");
+const readJSON = (p) => JSON.parse(readFileSync(p, "utf-8"));
+const STRAT_LABEL = /\bSTRATAGEM$/;
+const STRAT_SECTION = /^STRATAGEMS?$/;
+/** Table headers that masquerade as caps "enhancement" names (mirror of extractor). */
+const NOISE = new Set([
+    "RANGED WEAPONS", "MELEE WEAPONS", "RANGE", "WARGEAR OPTIONS", "UNIT COMPOSITION",
+    "KEYWORDS", "LEADER", "TRANSPORT", "DAMAGED", "SUPREME COMMANDER",
+]);
+/** True when the nearest same-column block below `b` is mixed-case prose. */
+function hasProseBelow(blocks, b) {
+    return blocks
+        .filter((o) => sameColumn(o.x, b.x) && o.gy > b.gy)
+        .sort((p, q) => p.gy - q.gy)
+        .some((o, i) => i === 0 && /[a-z]/.test(o.text));
+}
+/**
+ * Locate every card-start anchor in a detachment's blocks: the detachment rule
+ * (under a DETACHMENT RULE(S) header), each stratagem (its name block above a
+ * "<NAME> – <TYPE> STRATAGEM" label), and each enhancement (a caps header with a
+ * prose body below). Anchors bound each other's bodies by same-column y order.
+ */
+export function findAnchors(blocks) {
+    const anchors = [];
+    const seen = new Set();
+    const push = (a) => {
+        if (a.id && !seen.has(a.id)) {
+            seen.add(a.id);
+            anchors.push(a);
+        }
+    };
+    // detachment rule: first caps header beneath the DETACHMENT RULE(S) marker.
+    const ruleHeader = blocks.filter((b) => /^DETACHMENT RULES?$/.test(b.text)).sort((a, b) => a.gy - b.gy)[0];
+    let ruleSlug = "";
+    if (ruleHeader) {
+        const rn = blocks
+            .filter((b) => sameColumn(b.x, ruleHeader.x) && b.gy > ruleHeader.gy && isCapsHeader(b.text))
+            .sort((a, b) => a.gy - b.gy)[0];
+        if (rn) {
+            ruleSlug = slug(rn.text);
+            push({ id: ruleSlug, name: titleCase(rn.text), kind: "detachment", x: rn.x, topGy: rn.gy });
+        }
+    }
+    // stratagems: name block directly above each "… STRATAGEM" label, same column.
+    const stratIds = new Set();
+    for (const label of blocks.filter((b) => STRAT_LABEL.test(b.text) && !STRAT_SECTION.test(b.text))) {
+        const nameBlock = blocks
+            .filter((b) => sameColumn(b.x, label.x) && b.gy < label.gy && isCapsHeader(stripCp(b.text).name))
+            .sort((a, b) => b.gy - a.gy)[0];
+        if (!nameBlock)
+            continue;
+        const { name } = stripCp(nameBlock.text);
+        const id = slug(name);
+        stratIds.add(id);
+        push({ id, name: titleCase(name), kind: "stratagem", x: nameBlock.x, topGy: nameBlock.gy });
+    }
+    // enhancements: remaining caps headers with a prose body below (subtractive,
+    // mirroring the extractor) — not stratagem names, the rule, or table noise.
+    for (const b of blocks) {
+        if (!isCapsHeader(b.text) || NOISE.has(normCaps(b.text)))
+            continue;
+        const id = slug(b.text.replace(/ UPGRADE$/, "").trim());
+        if (!id || id === ruleSlug || stratIds.has(id) || seen.has(id))
+            continue;
+        if (!hasProseBelow(blocks, b))
+            continue;
+        const name = titleCase(b.text.replace(/ UPGRADE$/, "").trim());
+        push({ id, name, kind: "enhancement", x: b.x, topGy: b.gy });
+    }
+    return anchors;
+}
+/**
+ * Capture a card's rule body: every same-column block strictly below the anchor's
+ * name/header, up to the next same-column anchor (the next card's start). Joined
+ * in reading order. Column-scoping is what keeps the adjacent column's text out —
+ * the whole reason bbox beats `-layout` on these 2-column pages.
+ */
+export function captureBody(blocks, anchor, anchors) {
+    const nextTop = Math.min(...anchors
+        .filter((a) => a !== anchor && sameColumn(a.x, anchor.x) && a.topGy > anchor.topGy)
+        .map((a) => a.topGy), Infinity);
+    return blocks
+        .filter((b) => sameColumn(b.x, anchor.x) && b.gy > anchor.topGy && b.gy < nextTop)
+        .sort((a, b) => a.gy - b.gy)
+        .map((b) => b.text)
+        .join("\n")
+        .trim();
+}
+const normName = (s) => normCaps(s);
+/** Build author-input entries for one faction's new-detachment stubs from its pack. */
+export function buildFactionFromPack(faction, pdf) {
+    const abilitiesPath = resolve(ENRICHMENT_ROOT, faction, "abilities.json");
+    if (!existsSync(abilitiesPath))
+        return [];
+    const abilities = readJSON(abilitiesPath);
+    // In scope: empty-modifier stubs that belong to a detachment (the new-detachment
+    // rules/stratagems/enhancements). Unit-scoped stubs stay the 10e-archive path's.
+    const stubs = abilities.filter((a) => a.detachment_id && hasEmptyModifier(a.effect));
+    const segments = detachmentSegments(pdf);
+    const segById = new Map(segments.map((s) => [s.id, s]));
+    const anchorsBySeg = new Map();
+    const out = [];
+    for (const a of stubs) {
+        const base = {
+            faction,
+            ability_id: a.ability_id,
+            name: a.name,
+            unit_ids: (a.unit_ids ?? []),
+            target: (a.effect && typeof a.effect === "object" && a.effect.target) || null,
+            scope: a.scope ?? null,
+            faction_id: a.faction_id ?? null,
+            ability_type: a.ability_type ?? null,
+        };
+        const seg = segById.get(a.detachment_id);
+        if (!seg) {
+            out.push({ ...base, resolved: false, reason: `detachment "${a.detachment_id}" not found in pack` });
+            continue;
+        }
+        let anchors = anchorsBySeg.get(seg.id);
+        if (!anchors)
+            anchorsBySeg.set(seg.id, (anchors = findAnchors(seg.blocks)));
+        const anchor = anchors.find((c) => c.id === a.ability_id) ??
+            anchors.find((c) => normName(c.name) === normName(a.name));
+        if (!anchor) {
+            out.push({ ...base, resolved: false, reason: `no card matched "${a.name}" in "${a.detachment_id}"` });
+            continue;
+        }
+        const description = captureBody(seg.blocks, anchor, anchors);
+        if (!description) {
+            out.push({ ...base, resolved: false, reason: `empty body for "${a.name}"` });
+            continue;
+        }
+        out.push({
+            ...base,
+            resolved: true,
+            src: { datasheet_id: "", src_type: null, parameter: null, phases: null, description },
+        });
+    }
+    return out;
+}
+function main() {
+    const argv = process.argv.slice(2);
+    const pdf = argv[argv.indexOf("--pdf") + 1];
+    const faction = argv[argv.indexOf("--faction") + 1];
+    if (!argv.includes("--pdf") || !argv.includes("--faction") || !pdf || !faction) {
+        console.error("Usage: npx tsx tools/src/author-input-pack.ts --pdf <pack.pdf> --faction <id>");
+        process.exit(1);
+    }
+    const entries = buildFactionFromPack(faction, resolve(pdf));
+    mkdirSync(OUT_DIR, { recursive: true });
+    const outPath = resolve(OUT_DIR, `${faction}.json`);
+    writeFileSync(outPath, JSON.stringify(entries, null, 2) + "\n");
+    const resolved = entries.filter((e) => e.resolved).length;
+    console.log(`  ${faction}: ${entries.length} stubs, ${resolved} source-resolved → ${outPath}`);
+    for (const e of entries.filter((e) => !e.resolved))
+        console.log(`    unresolved ${e.ability_id}: ${e.reason}`);
+}
+const isMain = process.argv[1] &&
+    resolve(process.argv[1]).replace(/\.\w+$/, "") === fileURLToPath(import.meta.url).replace(/\.\w+$/, "");
+if (isMain)
+    main();
+//# sourceMappingURL=author-input-pack.js.map

package/dist/author-input-pack.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"author-input-pack.js","sourceRoot":"","sources":["../src/author-input-pack.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAC7E,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAEL,YAAY,EACZ,UAAU,EACV,OAAO,EACP,IAAI,EACJ,SAAS,EACT,QAAQ,GACT,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,kBAAkB,EAAE,MAAM,2BAA2B,CAAC;AAC/D,OAAO,EAAE,gBAAgB,EAAE,MAAM,qBAAqB,CAAC;AAGvD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,GAAG,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAC/D,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,EAAE,YAAY,CAAC,CAAC;AACnD,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,EAAE,YAAY,CAAC,CAAC;AACzD,MAAM,OAAO,GAAG,OAAO,CAAC,SAAS,EAAE,QAAQ,EAAE,cAAc,CAAC,CAAC;AAI7D,MAAM,QAAQ,GAAG,CAAC,CAAS,EAAQ,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;AAW3E,MAAM,WAAW,GAAG,cAAc,CAAC;AACnC,MAAM,aAAa,GAAG,eAAe,CAAC;AACtC,uFAAuF;AACvF,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC;IACpB,gBAAgB,EAAE,eAAe,EAAE,OAAO,EAAE,iBAAiB,EAAE,kBAAkB;IACjF,UAAU,EAAE,QAAQ,EAAE,WAAW,EAAE,SAAS,EAAE,mBAAmB;CAClE,CAAC,CAAC;AAEH,6EAA6E;AAC7E,SAAS,aAAa,CAAC,MAAe,EAAE,CAAQ;IAC9C,OAAO,MAAM;SACV,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;SAClD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;SAC3B,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;AACrD,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,WAAW,CAAC,MAAe;IACzC,MAAM,OAAO,GAAiB,EAAE,CAAC;IACjC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,IAAI,GAAG,CAAC,CAAa,EAAQ,EAAE;QACnC,IAAI,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC;YAC5B,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACf,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,CAAC;IACH,CAAC,CAAC;IAEF,4EAA4E;IAC5E,MAAM,UAAU,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,qBAAqB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC3G,IAAI,QAAQ,GAAG,EAAE,CAAC;IAClB,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,EAAE,GAAG,MAAM;aACd,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,GAAG,UAAU,CAAC,EAAE,IAAI,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;aAC5F,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QAClC,IAAI,EAAE,EAAE,CAAC;YACP,QAAQ,GAAG,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;YACzB,IAAI,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,SAAS,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;QAC9F,CAAC;IACH,CAAC;IAED,+EAA+E;IAC/E,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IACnC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QAClG,MAAM,SAAS,GAAG,MAAM;aACrB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,EAAE,IAAI,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;aAChG,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QAClC,IAAI,CAAC,SAAS;YAAE,SAAS;QACzB,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QACzC,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;QACtB,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACjB,IAAI,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,CAAC,EAAE,SAAS,CAAC,CAAC,EAAE,KAAK,EAAE,SAAS,CAAC,EAAE,EAAE,CAAC,CAAC;IAC9F,CAAC;IAED,6EAA6E;IAC7E,4EAA4E;IAC5E,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YAAE,SAAS;QACnE,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACxD,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,QAAQ,IAAI,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAAE,SAAS;QACzE,IAAI,CAAC,aAAa,CAAC,MAAM,EAAE,CAAC,CAAC;YAAE,SAAS;QACxC,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QAC/D,IAAI,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC/D,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,WAAW,CAAC,MAAe,EAAE,MAAkB,EAAE,OAAqB;IACpF,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CACtB,GAAG,OAAO;SACP,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,MAAM,IAAI,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC;SAClF,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,EACtB,QAAQ,CACT,CAAC;IACF,OAAO,MAAM;SACV,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,GAAG,MAAM,CAAC,KAAK,IAAI,CAAC,CAAC,EAAE,GAAG,OAAO,CAAC;SACjF,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;SAC3B,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;SAClB,IAAI,CAAC,IAAI,CAAC;SACV,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,QAAQ,GAAG,CAAC,CAAS,EAAU,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;AAEpD,uFAAuF;AACvF,MAAM,UAAU,oBAAoB,CAAC,OAAe,EAAE,GAAW;IAC/D,MAAM,aAAa,GAAG,OAAO,CAAC,eAAe,EAAE,OAAO,EAAE,gBAAgB,CAAC,CAAC;IAC1E,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC;QAAE,OAAO,EAAE,CAAC;IAC1C,MAAM,SAAS,GAAW,QAAQ,CAAC,aAAa,CAAC,CAAC;IAElD,iFAAiF;IACjF,iFAAiF;IACjF,MAAM,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa,IAAI,gBAAgB,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;IAErF,MAAM,QAAQ,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACxD,MAAM,YAAY,GAAG,IAAI,GAAG,EAAwB,CAAC;IAErD,MAAM,GAAG,GAAuB,EAAE,CAAC;IACnC,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG;YACX,OAAO;YACP,UAAU,EAAE,CAAC,CAAC,UAAoB;YAClC,IAAI,EAAE,CAAC,CAAC,IAAc;YACtB,QAAQ,EAAE,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAa;YACxC,MAAM,EAAE,CAAC,CAAC,CAAC,MAAM,IAAI,OAAO,CAAC,CAAC,MAAM,KAAK,QAAQ,IAAI,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,IAAI;YAC7E,KAAK,EAAE,CAAC,CAAC,KAAK,IAAI,IAAI;YACtB,UAAU,EAAE,CAAC,CAAC,UAAU,IAAI,IAAI;YAChC,YAAY,EAAE,CAAC,CAAC,YAAY,IAAI,IAAI;SACrC,CAAC;QACF,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC;QACzC,IAAI,CAAC,GAAG,EAAE,CAAC;YACT,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,EAAE,eAAe,CAAC,CAAC,aAAa,qBAAqB,EAAE,CAAC,CAAC;YACpG,SAAS;QACX,CAAC;QACD,IAAI,OAAO,GAAG,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACvC,IAAI,CAAC,OAAO;YAAE,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,OAAO,GAAG,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QAC5E,MAAM,MAAM,GACV,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,UAAU,CAAC;YAC1C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QAC7D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,EAAE,oBAAoB,CAAC,CAAC,IAAI,SAAS,CAAC,CAAC,aAAa,GAAG,EAAE,CAAC,CAAC;YACtG,SAAS;QACX,CAAC;QACD,MAAM,WAAW,GAAG,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;QAC7D,IAAI,CAAC,WAAW,EAAE,CAAC;YACjB,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,EAAE,mBAAmB,CAAC,CAAC,IAAI,GAAG,EAAE,CAAC,CAAC;YAC7E,SAAS;QACX,CAAC;QACD,GAAG,CAAC,IAAI,CAAC;YACP,GAAG,IAAI;YACP,QAAQ,EAAE,IAAI;YACd,GAAG,EAAE,EAAE,YAAY,EAAE,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,WAAW,EAAE;SACtF,CAAC,CAAC;IACL,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,IAAI;IACX,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACnC,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5C,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC;IACpD,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,OAAO,EAAE,CAAC;QAC/E,OAAO,CAAC,KAAK,CAAC,+EAA+E,CAAC,CAAC;QAC/F,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IAED,MAAM,OAAO,GAAG,oBAAoB,CAAC,OAAO,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5D,SAAS,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACxC,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,EAAE,GAAG,OAAO,OAAO,CAAC,CAAC;IACpD,aAAa,CAAC,OAAO,EAAE,IAAI,CAAC,SAAS,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;IAChE,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC;IAC1D,OAAO,CAAC,GAAG,CAAC,KAAK,OAAO,KAAK,OAAO,CAAC,MAAM,WAAW,QAAQ,sBAAsB,OAAO,EAAE,CAAC,CAAC;IAC/F,KAAK,MAAM,CAAC,IAAI,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;QAAE,OAAO,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC,UAAU,KAAK,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;AACjH,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,IAAI,EAAE,CAAC","sourcesContent":["/**\n * Build the DSL-authoring input for *new 11e detachments* from a Faction Pack PDF.\n *\n * The stock `author-input.ts` resolves each empty-modifier ability stub to its\n * source rule in the **10e army-assist archive** (unit→datasheet→ability chain).\n * That archive predates the 11e Faction Packs, so the new detachments' rules /\n * stratagems / enhancements aren't in it — the chain returns nothing and the\n * authoring pipeline has no text to classify.\n *\n * This sibling tool fills the gap from the *primary* 11e source: the pack PDF.\n * It captures each card's rule body by bbox coordinate (column-scoped, so the\n * packs' 2-column stratagem layout never interleaves — the failure mode that\n * killed the `-layout` approach) and writes the identical `AuthorInputEntry`\n * shape that `author-batch.ts propose` consumes. Nothing downstream changes.\n *\n * IP firewall: rule prose is captured ONLY into the gitignored authoring worklist\n * `data/_audit/author-input/<faction>.json` (.gitignore). It is never written to\n * the committed `faction-pack-input/` staging (that stays names + metadata only).\n *\n * Scope: only empty-modifier stubs that belong to a detachment (`detachment_id`\n * set) — i.e. the new-detachment rules/stratagems/enhancements the pack covers.\n * Unit-scoped stubs (`detachment_id` null) remain the 10e-archive path's job.\n *\n * Requires `pdftotext` (poppler) on PATH — authoring-time only, not CI.\n *\n * Usage: npx tsx tools/src/author-input-pack.ts --pdf <pack.pdf> --faction <id>\n */\nimport { readFileSync, writeFileSync, mkdirSync, existsSync } from \"node:fs\";\nimport { resolve } from \"node:path\";\nimport { fileURLToPath } from \"node:url\";\nimport {\n  type Block,\n  isCapsHeader,\n  sameColumn,\n  stripCp,\n  slug,\n  titleCase,\n  normCaps,\n} from \"./pack-blocks.js\";\nimport { detachmentSegments } from \"./extract-faction-pack.js\";\nimport { hasEmptyModifier } from \"./audit-coverage.js\";\nimport type { AuthorInputEntry } from \"./author-input.js\";\n\nconst __dirname = fileURLToPath(new URL(\".\", import.meta.url));\nconst DATA_ROOT = resolve(__dirname, \"../../data\");\nconst ENRICHMENT_ROOT = resolve(DATA_ROOT, \"enrichment\");\nconst OUT_DIR = resolve(DATA_ROOT, \"_audit\", \"author-input\");\n\n// eslint-disable-next-line @typescript-eslint/no-explicit-any\ntype Json = any;\nconst readJSON = (p: string): Json => JSON.parse(readFileSync(p, \"utf-8\"));\n\n/** The start of one card within a detachment: its id, name, column, and top y. */\nexport interface CardAnchor {\n  id: string;\n  name: string;\n  kind: \"detachment\" | \"stratagem\" | \"enhancement\";\n  x: number;\n  topGy: number;\n}\n\nconst STRAT_LABEL = /\\bSTRATAGEM$/;\nconst STRAT_SECTION = /^STRATAGEMS?$/;\n/** Table headers that masquerade as caps \"enhancement\" names (mirror of extractor). */\nconst NOISE = new Set([\n  \"RANGED WEAPONS\", \"MELEE WEAPONS\", \"RANGE\", \"WARGEAR OPTIONS\", \"UNIT COMPOSITION\",\n  \"KEYWORDS\", \"LEADER\", \"TRANSPORT\", \"DAMAGED\", \"SUPREME COMMANDER\",\n]);\n\n/** True when the nearest same-column block below `b` is mixed-case prose. */\nfunction hasProseBelow(blocks: Block[], b: Block): boolean {\n  return blocks\n    .filter((o) => sameColumn(o.x, b.x) && o.gy > b.gy)\n    .sort((p, q) => p.gy - q.gy)\n    .some((o, i) => i === 0 && /[a-z]/.test(o.text));\n}\n\n/**\n * Locate every card-start anchor in a detachment's blocks: the detachment rule\n * (under a DETACHMENT RULE(S) header), each stratagem (its name block above a\n * \"<NAME> – <TYPE> STRATAGEM\" label), and each enhancement (a caps header with a\n * prose body below). Anchors bound each other's bodies by same-column y order.\n */\nexport function findAnchors(blocks: Block[]): CardAnchor[] {\n  const anchors: CardAnchor[] = [];\n  const seen = new Set<string>();\n  const push = (a: CardAnchor): void => {\n    if (a.id && !seen.has(a.id)) {\n      seen.add(a.id);\n      anchors.push(a);\n    }\n  };\n\n  // detachment rule: first caps header beneath the DETACHMENT RULE(S) marker.\n  const ruleHeader = blocks.filter((b) => /^DETACHMENT RULES?$/.test(b.text)).sort((a, b) => a.gy - b.gy)[0];\n  let ruleSlug = \"\";\n  if (ruleHeader) {\n    const rn = blocks\n      .filter((b) => sameColumn(b.x, ruleHeader.x) && b.gy > ruleHeader.gy && isCapsHeader(b.text))\n      .sort((a, b) => a.gy - b.gy)[0];\n    if (rn) {\n      ruleSlug = slug(rn.text);\n      push({ id: ruleSlug, name: titleCase(rn.text), kind: \"detachment\", x: rn.x, topGy: rn.gy });\n    }\n  }\n\n  // stratagems: name block directly above each \"… STRATAGEM\" label, same column.\n  const stratIds = new Set<string>();\n  for (const label of blocks.filter((b) => STRAT_LABEL.test(b.text) && !STRAT_SECTION.test(b.text))) {\n    const nameBlock = blocks\n      .filter((b) => sameColumn(b.x, label.x) && b.gy < label.gy && isCapsHeader(stripCp(b.text).name))\n      .sort((a, b) => b.gy - a.gy)[0];\n    if (!nameBlock) continue;\n    const { name } = stripCp(nameBlock.text);\n    const id = slug(name);\n    stratIds.add(id);\n    push({ id, name: titleCase(name), kind: \"stratagem\", x: nameBlock.x, topGy: nameBlock.gy });\n  }\n\n  // enhancements: remaining caps headers with a prose body below (subtractive,\n  // mirroring the extractor) — not stratagem names, the rule, or table noise.\n  for (const b of blocks) {\n    if (!isCapsHeader(b.text) || NOISE.has(normCaps(b.text))) continue;\n    const id = slug(b.text.replace(/ UPGRADE$/, \"\").trim());\n    if (!id || id === ruleSlug || stratIds.has(id) || seen.has(id)) continue;\n    if (!hasProseBelow(blocks, b)) continue;\n    const name = titleCase(b.text.replace(/ UPGRADE$/, \"\").trim());\n    push({ id, name, kind: \"enhancement\", x: b.x, topGy: b.gy });\n  }\n\n  return anchors;\n}\n\n/**\n * Capture a card's rule body: every same-column block strictly below the anchor's\n * name/header, up to the next same-column anchor (the next card's start). Joined\n * in reading order. Column-scoping is what keeps the adjacent column's text out —\n * the whole reason bbox beats `-layout` on these 2-column pages.\n */\nexport function captureBody(blocks: Block[], anchor: CardAnchor, anchors: CardAnchor[]): string {\n  const nextTop = Math.min(\n    ...anchors\n      .filter((a) => a !== anchor && sameColumn(a.x, anchor.x) && a.topGy > anchor.topGy)\n      .map((a) => a.topGy),\n    Infinity,\n  );\n  return blocks\n    .filter((b) => sameColumn(b.x, anchor.x) && b.gy > anchor.topGy && b.gy < nextTop)\n    .sort((a, b) => a.gy - b.gy)\n    .map((b) => b.text)\n    .join(\"\\n\")\n    .trim();\n}\n\nconst normName = (s: string): string => normCaps(s);\n\n/** Build author-input entries for one faction's new-detachment stubs from its pack. */\nexport function buildFactionFromPack(faction: string, pdf: string): AuthorInputEntry[] {\n  const abilitiesPath = resolve(ENRICHMENT_ROOT, faction, \"abilities.json\");\n  if (!existsSync(abilitiesPath)) return [];\n  const abilities: Json[] = readJSON(abilitiesPath);\n\n  // In scope: empty-modifier stubs that belong to a detachment (the new-detachment\n  // rules/stratagems/enhancements). Unit-scoped stubs stay the 10e-archive path's.\n  const stubs = abilities.filter((a) => a.detachment_id && hasEmptyModifier(a.effect));\n\n  const segments = detachmentSegments(pdf);\n  const segById = new Map(segments.map((s) => [s.id, s]));\n  const anchorsBySeg = new Map<string, CardAnchor[]>();\n\n  const out: AuthorInputEntry[] = [];\n  for (const a of stubs) {\n    const base = {\n      faction,\n      ability_id: a.ability_id as string,\n      name: a.name as string,\n      unit_ids: (a.unit_ids ?? []) as string[],\n      target: (a.effect && typeof a.effect === \"object\" && a.effect.target) || null,\n      scope: a.scope ?? null,\n      faction_id: a.faction_id ?? null,\n      ability_type: a.ability_type ?? null,\n    };\n    const seg = segById.get(a.detachment_id);\n    if (!seg) {\n      out.push({ ...base, resolved: false, reason: `detachment \"${a.detachment_id}\" not found in pack` });\n      continue;\n    }\n    let anchors = anchorsBySeg.get(seg.id);\n    if (!anchors) anchorsBySeg.set(seg.id, (anchors = findAnchors(seg.blocks)));\n    const anchor =\n      anchors.find((c) => c.id === a.ability_id) ??\n      anchors.find((c) => normName(c.name) === normName(a.name));\n    if (!anchor) {\n      out.push({ ...base, resolved: false, reason: `no card matched \"${a.name}\" in \"${a.detachment_id}\"` });\n      continue;\n    }\n    const description = captureBody(seg.blocks, anchor, anchors);\n    if (!description) {\n      out.push({ ...base, resolved: false, reason: `empty body for \"${a.name}\"` });\n      continue;\n    }\n    out.push({\n      ...base,\n      resolved: true,\n      src: { datasheet_id: \"\", src_type: null, parameter: null, phases: null, description },\n    });\n  }\n  return out;\n}\n\nfunction main(): void {\n  const argv = process.argv.slice(2);\n  const pdf = argv[argv.indexOf(\"--pdf\") + 1];\n  const faction = argv[argv.indexOf(\"--faction\") + 1];\n  if (!argv.includes(\"--pdf\") || !argv.includes(\"--faction\") || !pdf || !faction) {\n    console.error(\"Usage: npx tsx tools/src/author-input-pack.ts --pdf <pack.pdf> --faction <id>\");\n    process.exit(1);\n  }\n\n  const entries = buildFactionFromPack(faction, resolve(pdf));\n  mkdirSync(OUT_DIR, { recursive: true });\n  const outPath = resolve(OUT_DIR, `${faction}.json`);\n  writeFileSync(outPath, JSON.stringify(entries, null, 2) + \"\\n\");\n  const resolved = entries.filter((e) => e.resolved).length;\n  console.log(`  ${faction}: ${entries.length} stubs, ${resolved} source-resolved → ${outPath}`);\n  for (const e of entries.filter((e) => !e.resolved)) console.log(`    unresolved ${e.ability_id}: ${e.reason}`);\n}\n\nconst isMain =\n  process.argv[1] &&\n  resolve(process.argv[1]).replace(/\\.\\w+$/, \"\") === fileURLToPath(import.meta.url).replace(/\\.\\w+$/, \"\");\nif (isMain) main();\n"]}

package/dist/build-share-registry.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=build-share-registry.d.ts.map

package/dist/build-share-registry.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"build-share-registry.d.ts","sourceRoot":"","sources":["../src/build-share-registry.ts"],"names":[],"mappings":""}

package/dist/build-share-registry.js

@@ -0,0 +1,115 @@
+/**
+ * Regenerates the share-token id registry (`data/share-registry.json`).
+ *
+ * The registry assigns every shareable entity id a stable integer index so the
+ * list-builder share codec can encode a roster as compact varints instead of
+ * full id strings (see `src/share/` and `docs/share-token.md`).
+ *
+ * **Append-only.** This tool never reorders or drops a slot: existing ids keep
+ * their index forever, genuinely-new ids are appended (sorted) at the end, and
+ * an id that has vanished from the dataset keeps its slot and is recorded in
+ * `tombstones`. Renames are author-curated: move the old id from `tombstones`
+ * into `aliases` ({ "old-id": "new-id" }) by hand. These invariants guarantee a
+ * newer registry can always decode an older token; only the reverse (an old
+ * package meeting a newer token) can fail, and it fails safe (stale-registry).
+ *
+ * Run intentionally via `npm run registry:build` and commit the diff — this is
+ * a migration step, not a per-build codegen. `codegen:data` separately inlines
+ * the committed JSON into `src/share/registry.generated.ts`.
+ */
+import { existsSync, readFileSync, writeFileSync } from "node:fs";
+import { dirname, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+import { alliedRules, detachments, enhancements, factions, forceDispositions, units, wargear, weapons, } from "./data/index.js";
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const REGISTRY_PATH = resolve(__dirname, "../../data/share-registry.json");
+/** Registry kinds, in a fixed order. The wire format references each by name. */
+const KINDS = [
+    "faction",
+    "detachment",
+    "unit",
+    "wargear",
+    "enhancement",
+    "ally_rule",
+    "disposition",
+];
+/** Distinct ids currently in the dataset for each kind, sorted lexically. */
+function currentIds() {
+    const distinct = (ids) => new Set(ids);
+    return {
+        faction: distinct(factions.all.map((f) => f.id)),
+        detachment: distinct(detachments.all.map((d) => d.id)),
+        unit: distinct(units.all.map((u) => u.id)),
+        // Loadout refs are either weapons or non-weapon wargear; they share one id
+        // space in the token, so the registry merges both collections.
+        wargear: distinct([...weapons.all.map((w) => w.id), ...wargear.all.map((w) => w.id)]),
+        enhancement: distinct(enhancements.all.map((e) => e.id)),
+        ally_rule: distinct(alliedRules.all.map((r) => r.id)),
+        disposition: distinct(forceDispositions.all.map((d) => d.id)),
+    };
+}
+function emptyRegistry() {
+    return {
+        version: 0,
+        kinds: Object.fromEntries(KINDS.map((k) => [k, []])),
+        aliases: {},
+        tombstones: [],
+    };
+}
+function loadRegistry() {
+    if (!existsSync(REGISTRY_PATH))
+        return emptyRegistry();
+    const parsed = JSON.parse(readFileSync(REGISTRY_PATH, "utf-8"));
+    // Tolerate a registry authored before a kind existed.
+    for (const kind of KINDS)
+        parsed.kinds[kind] ??= [];
+    parsed.aliases ??= {};
+    parsed.tombstones ??= [];
+    return parsed;
+}
+function main() {
+    const prev = loadRegistry();
+    const current = currentIds();
+    const kinds = {};
+    const newlyRemoved = [];
+    let appended = 0;
+    for (const kind of KINDS) {
+        const existing = prev.kinds[kind] ?? [];
+        const have = new Set(existing);
+        // Append-only: new ids (sorted) after the preserved existing order.
+        const added = [...current[kind]].filter((id) => !have.has(id)).sort();
+        kinds[kind] = [...existing, ...added];
+        appended += added.length;
+        // A previously-allocated id absent from the dataset becomes a tombstone
+        // (its slot is retained above, so existing indices never shift).
+        for (const id of existing) {
+            if (!current[kind].has(id))
+                newlyRemoved.push(id);
+        }
+    }
+    // Tombstone anything newly removed that the author hasn't already mapped to a
+    // replacement via `aliases`. Preserve prior tombstones and aliases verbatim.
+    const tombstones = new Set(prev.tombstones);
+    for (const id of newlyRemoved) {
+        if (!(id in prev.aliases))
+            tombstones.add(id);
+    }
+    const changed = appended > 0 || tombstones.size !== prev.tombstones.length;
+    const next = {
+        version: changed ? prev.version + 1 : prev.version,
+        kinds,
+        aliases: prev.aliases,
+        tombstones: [...tombstones].sort(),
+    };
+    writeFileSync(REGISTRY_PATH, JSON.stringify(next, null, 2) + "\n");
+    const counts = KINDS.map((k) => `${k}=${kinds[k].length}`).join(", ");
+    if (changed) {
+        console.log(`Wrote ${REGISTRY_PATH}\n  version ${prev.version} → ${next.version} ` +
+            `(+${appended} ids, ${newlyRemoved.length} tombstoned)\n  ${counts}`);
+    }
+    else {
+        console.log(`No change: ${REGISTRY_PATH} already at version ${next.version}\n  ${counts}`);
+    }
+}
+main();
+//# sourceMappingURL=build-share-registry.js.map