@alpaca-software/40kdc-data 0.5.12 → 0.5.14

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"]}