@lingjingai/scriptctl 0.35.0 → 0.36.0

package/dist/domain/ingest/jsonl-parser.js

@@ -0,0 +1,187 @@
+// Strict JSONL → typed record parser (§7). Rejects unknown fields, malformed
+// records, missing/invalid source spans, and inline-tag syntax errors. Never
+// repairs or downgrades: a bad line fails the batch (§12.3).
+import { NON_ACTOR_SPEAKER_KINDS, RECORD_KINDS, } from "./records.js";
+export class JsonlParseError extends Error {
+    line;
+    constructor(message, line) {
+        super(`line ${line}: ${message}`);
+        this.line = line;
+    }
+}
+function isObj(v) {
+    return typeof v === "object" && v !== null && !Array.isArray(v);
+}
+function requireKeys(obj, allowed, line) {
+    for (const key of Object.keys(obj)) {
+        if (!allowed.includes(key))
+            throw new JsonlParseError(`unknown field '${key}'`, line);
+    }
+}
+function parseSource(v, line) {
+    if (!isObj(v))
+        throw new JsonlParseError("missing source", line);
+    requireKeys(v, ["chunk", "line_start", "line_end"], line);
+    const { chunk, line_start, line_end } = v;
+    if (![chunk, line_start, line_end].every((n) => typeof n === "number" && Number.isInteger(n) && n >= 0)) {
+        throw new JsonlParseError("source needs integer chunk/line_start/line_end", line);
+    }
+    if (line_start >= line_end)
+        throw new JsonlParseError("source line_start must be < line_end", line);
+    return { chunk: chunk, line_start: line_start, line_end: line_end };
+}
+function str(obj, key, line) {
+    const v = obj[key];
+    if (typeof v !== "string" || v === "")
+        throw new JsonlParseError(`'${key}' must be a non-empty string`, line);
+    return v;
+}
+function optStr(obj, key, line) {
+    if (!(key in obj) || obj[key] === undefined)
+        return undefined;
+    const v = obj[key];
+    if (typeof v !== "string" || v === "")
+        throw new JsonlParseError(`'${key}' must be a non-empty string`, line);
+    return v;
+}
+function parseEntityRefs(v, key, line) {
+    if (!Array.isArray(v))
+        throw new JsonlParseError(`'${key}' must be an array`, line);
+    return v.map((item) => {
+        if (!isObj(item))
+            throw new JsonlParseError(`'${key}' entries must be objects`, line);
+        requireKeys(item, ["name", "state"], line);
+        const ref = { name: str(item, "name", line) };
+        const state = optStr(item, "state", line);
+        if (state !== undefined)
+            ref.state = state;
+        return ref;
+    });
+}
+function parseSpeaker(v, line) {
+    if (!isObj(v))
+        throw new JsonlParseError("dialogue/inner_thought needs a speaker", line);
+    if ("kind" in v) {
+        requireKeys(v, ["kind", "label"], line);
+        const kind = v.kind;
+        if (typeof kind !== "string" || !NON_ACTOR_SPEAKER_KINDS.includes(kind)) {
+            throw new JsonlParseError(`speaker.kind must be one of ${NON_ACTOR_SPEAKER_KINDS.join("|")}`, line);
+        }
+        const ref = { kind: kind };
+        const label = optStr(v, "label", line);
+        if (label !== undefined)
+            ref.label = label;
+        return ref;
+    }
+    requireKeys(v, ["name", "state"], line);
+    const ref = { name: str(v, "name", line) };
+    const state = optStr(v, "state", line);
+    if (state !== undefined)
+        ref.state = state;
+    return ref;
+}
+function parseRecord(obj, line) {
+    const kind = obj.kind;
+    if (typeof kind !== "string" || !RECORD_KINDS.includes(kind)) {
+        throw new JsonlParseError(`unknown or missing kind '${String(kind)}'`, line);
+    }
+    const source = parseSource(obj.source, line);
+    switch (kind) {
+        case "episode": {
+            requireKeys(obj, ["kind", "episode", "title", "source"], line);
+            const episode = obj.episode;
+            if (typeof episode !== "number" || !Number.isInteger(episode))
+                throw new JsonlParseError("episode must be an integer", line);
+            const rec = { kind, episode, source };
+            const title = optStr(obj, "title", line);
+            if (title !== undefined)
+                rec.title = title;
+            return rec;
+        }
+        case "scene": {
+            requireKeys(obj, ["kind", "scene", "environment", "location", "actors", "props", "source"], line);
+            const rec = { kind, source };
+            if ("scene" in obj && obj.scene !== undefined) {
+                if (typeof obj.scene !== "number" || !Number.isInteger(obj.scene))
+                    throw new JsonlParseError("scene must be an integer", line);
+                rec.scene = obj.scene;
+            }
+            if ("environment" in obj && obj.environment !== undefined) {
+                if (!isObj(obj.environment))
+                    throw new JsonlParseError("environment must be an object", line);
+                requireKeys(obj.environment, ["space", "time"], line);
+                rec.environment = obj.environment;
+            }
+            if ("location" in obj && obj.location !== undefined) {
+                if (!isObj(obj.location))
+                    throw new JsonlParseError("location must be an object", line);
+                requireKeys(obj.location, ["name", "state"], line);
+                const loc = { name: str(obj.location, "name", line) };
+                const st = optStr(obj.location, "state", line);
+                if (st !== undefined)
+                    loc.state = st;
+                rec.location = loc;
+            }
+            if ("actors" in obj && obj.actors !== undefined)
+                rec.actors = parseEntityRefs(obj.actors, "actors", line);
+            if ("props" in obj && obj.props !== undefined)
+                rec.props = parseEntityRefs(obj.props, "props", line);
+            return rec;
+        }
+        case "action":
+        case "transition": {
+            requireKeys(obj, ["kind", "content", "source"], line);
+            const content = str(obj, "content", line);
+            return { kind, content, source };
+        }
+        case "dialogue":
+        case "inner_thought": {
+            requireKeys(obj, ["kind", "speaker", "content", "emotion", "source"], line);
+            const content = str(obj, "content", line);
+            const speaker = parseSpeaker(obj.speaker, line);
+            const rec = { kind, speaker, content, source };
+            const emotion = optStr(obj, "emotion", line);
+            if (emotion !== undefined)
+                rec.emotion = emotion;
+            return rec;
+        }
+        case "note": {
+            requireKeys(obj, ["kind", "content", "category", "source"], line);
+            const content = str(obj, "content", line);
+            const rec = { kind, content, source };
+            if ("category" in obj && obj.category !== undefined) {
+                const cat = obj.category;
+                if (cat !== "author_note" && cat !== "format_note" && cat !== "unclassified") {
+                    throw new JsonlParseError("note.category must be author_note|format_note|unclassified", line);
+                }
+                rec.category = cat;
+            }
+            return rec;
+        }
+        default:
+            throw new JsonlParseError(`unhandled kind '${kind}'`, line);
+    }
+}
+// Parse full JSONL text. Blank lines are skipped; every other line must be a
+// valid record. `lineOffset` lets callers report absolute file lines.
+export function parseJsonl(text, lineOffset = 0) {
+    const out = [];
+    const lines = text.split("\n");
+    for (let i = 0; i < lines.length; i++) {
+        const raw = lines[i].trim();
+        if (raw === "")
+            continue;
+        let parsed;
+        try {
+            parsed = JSON.parse(raw);
+        }
+        catch {
+            throw new JsonlParseError("invalid JSON", i + 1 + lineOffset);
+        }
+        if (!isObj(parsed))
+            throw new JsonlParseError("record must be a JSON object", i + 1 + lineOffset);
+        out.push(parseRecord(parsed, i + 1 + lineOffset));
+    }
+    return out;
+}
+//# sourceMappingURL=jsonl-parser.js.map

package/dist/domain/ingest/jsonl-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"jsonl-parser.js","sourceRoot":"","sources":["../../../src/domain/ingest/jsonl-parser.ts"],"names":[],"mappings":"AAAA,6EAA6E;AAC7E,6EAA6E;AAC7E,6DAA6D;AAE7D,OAAO,EAKL,uBAAuB,EACvB,YAAY,GACb,MAAM,cAAc,CAAC;AAEtB,MAAM,OAAO,eAAgB,SAAQ,KAAK;IACF;IAAtC,YAAY,OAAe,EAAW,IAAY;QAChD,KAAK,CAAC,QAAQ,IAAI,KAAK,OAAO,EAAE,CAAC,CAAC;QADE,SAAI,GAAJ,IAAI,CAAQ;IAElD,CAAC;CACF;AAID,SAAS,KAAK,CAAC,CAAU;IACvB,OAAO,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;AAClE,CAAC;AAED,SAAS,WAAW,CAAC,GAAQ,EAAE,OAAiB,EAAE,IAAY;IAC5D,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACnC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC;YAAE,MAAM,IAAI,eAAe,CAAC,kBAAkB,GAAG,GAAG,EAAE,IAAI,CAAC,CAAC;IACxF,CAAC;AACH,CAAC;AAED,SAAS,WAAW,CAAC,CAAU,EAAE,IAAY;IAC3C,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QAAE,MAAM,IAAI,eAAe,CAAC,gBAAgB,EAAE,IAAI,CAAC,CAAC;IACjE,WAAW,CAAC,CAAC,EAAE,CAAC,OAAO,EAAE,YAAY,EAAE,UAAU,CAAC,EAAE,IAAI,CAAC,CAAC;IAC1D,MAAM,EAAE,KAAK,EAAE,UAAU,EAAE,QAAQ,EAAE,GAAG,CAAC,CAAC;IAC1C,IAAI,CAAC,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QACxG,MAAM,IAAI,eAAe,CAAC,gDAAgD,EAAE,IAAI,CAAC,CAAC;IACpF,CAAC;IACD,IAAK,UAAqB,IAAK,QAAmB;QAAE,MAAM,IAAI,eAAe,CAAC,sCAAsC,EAAE,IAAI,CAAC,CAAC;IAC5H,OAAO,EAAE,KAAK,EAAE,KAAe,EAAE,UAAU,EAAE,UAAoB,EAAE,QAAQ,EAAE,QAAkB,EAAE,CAAC;AACpG,CAAC;AAED,SAAS,GAAG,CAAC,GAAQ,EAAE,GAAW,EAAE,IAAY;IAC9C,MAAM,CAAC,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;IACnB,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,EAAE;QAAE,MAAM,IAAI,eAAe,CAAC,IAAI,GAAG,8BAA8B,EAAE,IAAI,CAAC,CAAC;IAC9G,OAAO,CAAC,CAAC;AACX,CAAC;AAED,SAAS,MAAM,CAAC,GAAQ,EAAE,GAAW,EAAE,IAAY;IACjD,IAAI,CAAC,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC9D,MAAM,CAAC,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;IACnB,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,EAAE;QAAE,MAAM,IAAI,eAAe,CAAC,IAAI,GAAG,8BAA8B,EAAE,IAAI,CAAC,CAAC;IAC9G,OAAO,CAAC,CAAC;AACX,CAAC;AAED,SAAS,eAAe,CAAC,CAAU,EAAE,GAAW,EAAE,IAAY;IAC5D,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;QAAE,MAAM,IAAI,eAAe,CAAC,IAAI,GAAG,oBAAoB,EAAE,IAAI,CAAC,CAAC;IACpF,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QACpB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;YAAE,MAAM,IAAI,eAAe,CAAC,IAAI,GAAG,2BAA2B,EAAE,IAAI,CAAC,CAAC;QACtF,WAAW,CAAC,IAAI,EAAE,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC;QAC3C,MAAM,GAAG,GAAiB,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,CAAC;QAC5D,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;QAC1C,IAAI,KAAK,KAAK,SAAS;YAAE,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC;QAC3C,OAAO,GAAG,CAAC;IACb,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,YAAY,CAAC,CAAU,EAAE,IAAY;IAC5C,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QAAE,MAAM,IAAI,eAAe,CAAC,wCAAwC,EAAE,IAAI,CAAC,CAAC;IACzF,IAAI,MAAM,IAAI,CAAC,EAAE,CAAC;QAChB,WAAW,CAAC,CAAC,EAAE,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC;QACxC,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC;QACpB,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,CAAE,uBAA6C,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YAC/F,MAAM,IAAI,eAAe,CAAC,+BAA+B,uBAAuB,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;QACtG,CAAC;QACD,MAAM,GAAG,GAAe,EAAE,IAAI,EAAE,IAAgD,EAAE,CAAC;QACnF,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;QACvC,IAAI,KAAK,KAAK,SAAS;YAAE,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC;QAC3C,OAAO,GAAG,CAAC;IACb,CAAC;IACD,WAAW,CAAC,CAAC,EAAE,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC;IACxC,MAAM,GAAG,GAAe,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,CAAC;IACvD,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;IACvC,IAAI,KAAK,KAAK,SAAS;QAAE,GAAG,CAAC,KAAK,GAAG,KAAK,CAAC;IAC3C,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,WAAW,CAAC,GAAQ,EAAE,IAAY;IACzC,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC;IACtB,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,CAAE,YAAkC,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;QACpF,MAAM,IAAI,eAAe,CAAC,4BAA4B,MAAM,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;IAC/E,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAE7C,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,SAAS,CAAC,CAAC,CAAC;YACf,WAAW,CAAC,GAAG,EAAE,CAAC,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YAC/D,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,CAAC;YAC5B,IAAI,OAAO,OAAO,KAAK,QAAQ,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC;gBAAE,MAAM,IAAI,eAAe,CAAC,4BAA4B,EAAE,IAAI,CAAC,CAAC;YAC7H,MAAM,GAAG,GAAG,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,EAAkB,CAAC;YACtD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;YACzC,IAAI,KAAK,KAAK,SAAS;gBAAG,GAA0B,CAAC,KAAK,GAAG,KAAK,CAAC;YACnE,OAAO,GAAG,CAAC;QACb,CAAC;QACD,KAAK,OAAO,CAAC,CAAC,CAAC;YACb,WAAW,CAAC,GAAG,EAAE,CAAC,MAAM,EAAE,OAAO,EAAE,aAAa,EAAE,UAAU,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YAClG,MAAM,GAAG,GAA4B,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC;YACtD,IAAI,OAAO,IAAI,GAAG,IAAI,GAAG,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;gBAC9C,IAAI,OAAO,GAAG,CAAC,KAAK,KAAK,QAAQ,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC;oBAAE,MAAM,IAAI,eAAe,CAAC,0BAA0B,EAAE,IAAI,CAAC,CAAC;gBAC/H,GAAG,CAAC,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC;YACxB,CAAC;YACD,IAAI,aAAa,IAAI,GAAG,IAAI,GAAG,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;gBAC1D,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC;oBAAE,MAAM,IAAI,eAAe,CAAC,+BAA+B,EAAE,IAAI,CAAC,CAAC;gBAC9F,WAAW,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,IAAI,CAAC,CAAC;gBACtD,GAAG,CAAC,WAAW,GAAG,GAAG,CAAC,WAAW,CAAC;YACpC,CAAC;YACD,IAAI,UAAU,IAAI,GAAG,IAAI,GAAG,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;gBACpD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC;oBAAE,MAAM,IAAI,eAAe,CAAC,4BAA4B,EAAE,IAAI,CAAC,CAAC;gBACxF,WAAW,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC;gBACnD,MAAM,GAAG,GAA4B,EAAE,IAAI,EAAE,GAAG,CAAC,GAAG,CAAC,QAAQ,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,CAAC;gBAC/E,MAAM,EAAE,GAAG,MAAM,CAAC,GAAG,CAAC,QAAQ,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;gBAC/C,IAAI,EAAE,KAAK,SAAS;oBAAE,GAAG,CAAC,KAAK,GAAG,EAAE,CAAC;gBACrC,GAAG,CAAC,QAAQ,GAAG,GAAG,CAAC;YACrB,CAAC;YACD,IAAI,QAAQ,IAAI,GAAG,IAAI,GAAG,CAAC,MAAM,KAAK,SAAS;gBAAE,GAAG,CAAC,MAAM,GAAG,eAAe,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,IAAI,CAAC,CAAC;YAC1G,IAAI,OAAO,IAAI,GAAG,IAAI,GAAG,CAAC,KAAK,KAAK,SAAS;gBAAE,GAAG,CAAC,KAAK,GAAG,eAAe,CAAC,GAAG,CAAC,KAAK,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC;YACrG,OAAO,GAA8B,CAAC;QACxC,CAAC;QACD,KAAK,QAAQ,CAAC;QACd,KAAK,YAAY,CAAC,CAAC,CAAC;YAClB,WAAW,CAAC,GAAG,EAAE,CAAC,MAAM,EAAE,SAAS,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YACtD,MAAM,OAAO,GAAG,GAAG,CAAC,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC;YAC1C,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,EAAkB,CAAC;QACnD,CAAC;QACD,KAAK,UAAU,CAAC;QAChB,KAAK,eAAe,CAAC,CAAC,CAAC;YACrB,WAAW,CAAC,GAAG,EAAE,CAAC,MAAM,EAAE,SAAS,EAAE,SAAS,EAAE,SAAS,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YAC5E,MAAM,OAAO,GAAG,GAAG,CAAC,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC;YAC1C,MAAM,OAAO,GAAG,YAAY,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;YAChD,MAAM,GAAG,GAA4B,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC;YACxE,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC;YAC7C,IAAI,OAAO,KAAK,SAAS;gBAAE,GAAG,CAAC,OAAO,GAAG,OAAO,CAAC;YACjD,OAAO,GAA8B,CAAC;QACxC,CAAC;QACD,KAAK,MAAM,CAAC,CAAC,CAAC;YACZ,WAAW,CAAC,GAAG,EAAE,CAAC,MAAM,EAAE,SAAS,EAAE,UAAU,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YAClE,MAAM,OAAO,GAAG,GAAG,CAAC,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,CAAC;YAC1C,MAAM,GAAG,GAA4B,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC;YAC/D,IAAI,UAAU,IAAI,GAAG,IAAI,GAAG,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;gBACpD,MAAM,GAAG,GAAG,GAAG,CAAC,QAAQ,CAAC;gBACzB,IAAI,GAAG,KAAK,aAAa,IAAI,GAAG,KAAK,aAAa,IAAI,GAAG,KAAK,cAAc,EAAE,CAAC;oBAC7E,MAAM,IAAI,eAAe,CAAC,4DAA4D,EAAE,IAAI,CAAC,CAAC;gBAChG,CAAC;gBACD,GAAG,CAAC,QAAQ,GAAG,GAAG,CAAC;YACrB,CAAC;YACD,OAAO,GAA8B,CAAC;QACxC,CAAC;QACD;YACE,MAAM,IAAI,eAAe,CAAC,mBAAmB,IAAI,GAAG,EAAE,IAAI,CAAC,CAAC;IAChE,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,sEAAsE;AACtE,MAAM,UAAU,UAAU,CAAC,IAAY,EAAE,UAAU,GAAG,CAAC;IACrD,MAAM,GAAG,GAAmB,EAAE,CAAC;IAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC7B,IAAI,GAAG,KAAK,EAAE;YAAE,SAAS;QACzB,IAAI,MAAe,CAAC;QACpB,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC3B,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,eAAe,CAAC,cAAc,EAAE,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC;QAChE,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;YAAE,MAAM,IAAI,eAAe,CAAC,8BAA8B,EAAE,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC;QAClG,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC,CAAC;IACpD,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/domain/ingest/markdown-parser.d.ts

@@ -0,0 +1,6 @@
+import type { Chunk } from "./chunker.js";
+import { type IngestRecord, type NameStateRef } from "./records.js";
+export declare class MarkdownParseError extends Error {
+}
+export declare function parseEntityList(raw: string): NameStateRef[];
+export declare function parseNormalizeMarkdown(md: string, chunk: Chunk): IngestRecord[];

package/dist/domain/ingest/markdown-parser.js

@@ -0,0 +1,186 @@
+// Deterministic normalize-markdown → records parser (§pass2). The cheap model
+// rewrites one source chunk as a lean plain-text screenplay markdown
+// (episode/scene headers + `- 时空/地点/人物/道具` scene bullets + `> [anchor]
+// body` lines). Bodies are PLAIN TEXT — no inline entity tags. Scene cast/props
+// declared via `- 人物`/`- 道具` are captured onto the scene record; entity
+// binding is done later by pass5-extract. Source is chunk-granular.
+import { NON_ACTOR_SPEAKER_KINDS, } from "./records.js";
+export class MarkdownParseError extends Error {
+}
+const EP_RE = /^#\s*第?\s*(\d+)\s*集?\s*(.*)$/;
+const SCENE_LABEL_RE = /^##\s*(.*)$/;
+const ENV_RE = /^-\s*时空\s*[:：]\s*(.*)$/;
+const LOC_RE = /^-\s*地点\s*[:：]\s*(.*)$/;
+const CAST_RE = /^-\s*人物\s*[:：]\s*(.*)$/;
+const PROP_RE = /^-\s*道具\s*[:：]\s*(.*)$/;
+// dialogue/inner_thought/transition/note carry a `[label]`; everything else is
+// a bare action line. `> ` and `[act]` are accepted but not required.
+const ANCHOR_RE = /^(?:>\s*)?\[([^\]]*)\]\s?([\s\S]*)$/;
+const KNOWN_LABELS = new Set(["act", "dlg", "think", "trans", "note"]);
+const SPACE_MAP = { "内": "interior", "外": "exterior" };
+const TIME_MAP = { "日": "day", "夜": "night" };
+// Split a declared cast/prop list on top-level separators, keeping 【…】 (which
+// may itself contain commas) intact, then split each item into name + 【state】.
+export function parseEntityList(raw) {
+    const items = [];
+    let depth = 0;
+    let buf = "";
+    for (const ch of raw) {
+        if (ch === "【")
+            depth++;
+        else if (ch === "】")
+            depth = Math.max(0, depth - 1);
+        if ((ch === "，" || ch === "," || ch === "、") && depth === 0) {
+            if (buf.trim())
+                items.push(buf.trim());
+            buf = "";
+            continue;
+        }
+        buf += ch;
+    }
+    if (buf.trim())
+        items.push(buf.trim());
+    const refs = [];
+    for (const item of items) {
+        const m = /^([^【]+?)\s*(?:【(.+)】)?\s*$/.exec(item);
+        if (!m)
+            continue;
+        const name = m[1].trim();
+        if (!name)
+            continue;
+        const state = m[2]?.trim();
+        refs.push(state ? { name, state } : { name });
+    }
+    return refs;
+}
+function parseSpeaker(field) {
+    const trimmed = field.trim();
+    const colon = trimmed.indexOf(":");
+    if (colon > 0) {
+        const kind = trimmed.slice(0, colon).trim();
+        if (NON_ACTOR_SPEAKER_KINDS.includes(kind)) {
+            const label = trimmed.slice(colon + 1).trim();
+            return label ? { kind: kind, label } : { kind: kind };
+        }
+    }
+    // A bare kind word (`group` / `system` / …) with no label is that kind, not a
+    // person named "group".
+    if (NON_ACTOR_SPEAKER_KINDS.includes(trimmed))
+        return { kind: trimmed };
+    if (!trimmed)
+        throw new MarkdownParseError("dialogue anchor missing speaker");
+    return { name: trimmed };
+}
+// A dialogue/inner_thought anchor may pack several simultaneous speakers
+// (`李叔,张妈`). Split on punctuation separators only (NOT 和, which can appear
+// inside a name); each becomes its own record with the same content.
+function parseSpeakers(field) {
+    const parts = field.split(/[，,、；;/]/).map((s) => s.trim()).filter(Boolean);
+    if (parts.length <= 1)
+        return [parseSpeaker(field)];
+    return parts.map(parseSpeaker);
+}
+const NOTE_CATEGORY = { author: "author_note", format: "format_note" };
+function anchorRecords(label, body, source) {
+    const content = body.trim();
+    const parts = label.split("|");
+    const kind = parts[0].trim();
+    switch (kind) {
+        case "act":
+            return [{ kind: "action", content, source }];
+        case "trans":
+            return [{ kind: "transition", content, source }];
+        case "dlg":
+        case "think": {
+            const emotion = parts[2]?.trim();
+            // Simultaneous speakers → one record each, same content (overlap 拆多条).
+            return parseSpeakers(parts[1] ?? "").map((speaker) => {
+                const rec = { kind: kind === "dlg" ? "dialogue" : "inner_thought", speaker, content, source };
+                if (emotion)
+                    rec.emotion = emotion;
+                return rec;
+            });
+        }
+        case "note":
+            return [{ kind: "note", content, category: NOTE_CATEGORY[(parts[1] ?? "").trim()] ?? "unclassified", source }];
+        default:
+            throw new MarkdownParseError(`unknown anchor label '[${label}]'`);
+    }
+}
+export function parseNormalizeMarkdown(md, chunk) {
+    const source = { chunk: chunk.index, line_start: chunk.core_line_start, line_end: chunk.core_line_end };
+    const records = [];
+    const lines = stripFences(md).split("\n");
+    let openScene = null;
+    for (const raw of lines) {
+        const t = raw.trim();
+        if (t === "")
+            continue;
+        if (t.startsWith("##")) {
+            const rec = { kind: "scene", source };
+            const num = /(\d+)/.exec(SCENE_LABEL_RE.exec(t)?.[1]?.trim() ?? "")?.[1];
+            if (num)
+                rec.scene = Number(num);
+            records.push(rec);
+            openScene = rec;
+            continue;
+        }
+        if (t.startsWith("#")) {
+            const ep = EP_RE.exec(t);
+            if (ep) {
+                const title = ep[2].trim();
+                const rec = { kind: "episode", episode: Number(ep[1]), source };
+                if (title)
+                    rec.title = title;
+                records.push(rec);
+            }
+            continue;
+        }
+        // `- …` scene metadata bullet (only meaningful while a scene is open).
+        if (t.startsWith("-")) {
+            if (openScene) {
+                const env = ENV_RE.exec(t);
+                if (env) {
+                    const e = {};
+                    for (const tok of env[1].trim().split(/\s+/)) {
+                        if (SPACE_MAP[tok])
+                            e.space = SPACE_MAP[tok];
+                        if (TIME_MAP[tok])
+                            e.time = TIME_MAP[tok];
+                    }
+                    if (e.space || e.time)
+                        openScene.environment = e;
+                }
+                else {
+                    const loc = LOC_RE.exec(t);
+                    const cast = CAST_RE.exec(t);
+                    const prop = PROP_RE.exec(t);
+                    if (loc && loc[1].trim()) {
+                        const [name, state] = loc[1].split("|");
+                        openScene.location = state ? { name: name.trim(), state: state.trim() } : { name: name.trim() };
+                    }
+                    else if (cast && cast[1].trim())
+                        openScene.actors = parseEntityList(cast[1]);
+                    else if (prop && prop[1].trim())
+                        openScene.props = parseEntityList(prop[1]);
+                }
+            }
+            continue;
+        }
+        // A `[dlg|…]`/`[think|…]`/`[trans]`/`[note]`/`[act]` anchor, else a bare
+        // action line (content = the whole line).
+        const anchor = ANCHOR_RE.exec(t);
+        if (anchor && KNOWN_LABELS.has(anchor[1].split("|")[0].trim()))
+            records.push(...anchorRecords(anchor[1], anchor[2] ?? "", source));
+        else
+            records.push({ kind: "action", content: t, source });
+    }
+    const BODY = new Set(["action", "dialogue", "inner_thought", "transition", "note"]);
+    return records.filter((r) => !(BODY.has(r.kind) && "content" in r && r.content.trim() === ""));
+}
+function stripFences(raw) {
+    const trimmed = raw.trim();
+    const fence = trimmed.match(/^```(?:markdown|md)?\n([\s\S]*?)\n```$/);
+    return fence ? fence[1] : trimmed;
+}
+//# sourceMappingURL=markdown-parser.js.map

package/dist/domain/ingest/markdown-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"markdown-parser.js","sourceRoot":"","sources":["../../../src/domain/ingest/markdown-parser.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,qEAAqE;AACrE,uEAAuE;AACvE,gFAAgF;AAChF,wEAAwE;AACxE,oEAAoE;AAGpE,OAAO,EAKL,uBAAuB,GACxB,MAAM,cAAc,CAAC;AAEtB,MAAM,OAAO,kBAAmB,SAAQ,KAAK;CAAG;AAEhD,MAAM,KAAK,GAAG,8BAA8B,CAAC;AAC7C,MAAM,cAAc,GAAG,aAAa,CAAC;AACrC,MAAM,MAAM,GAAG,wBAAwB,CAAC;AACxC,MAAM,MAAM,GAAG,wBAAwB,CAAC;AACxC,MAAM,OAAO,GAAG,wBAAwB,CAAC;AACzC,MAAM,OAAO,GAAG,wBAAwB,CAAC;AACzC,+EAA+E;AAC/E,sEAAsE;AACtE,MAAM,SAAS,GAAG,qCAAqC,CAAC;AACxD,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC;AAEvE,MAAM,SAAS,GAA2B,EAAE,GAAG,EAAE,UAAU,EAAE,GAAG,EAAE,UAAU,EAAE,CAAC;AAC/E,MAAM,QAAQ,GAA2B,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC;AAEtE,8EAA8E;AAC9E,+EAA+E;AAC/E,MAAM,UAAU,eAAe,CAAC,GAAW;IACzC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;QACrB,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,EAAE,CAAC;aACnB,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;QACpD,IAAI,CAAC,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,GAAG,CAAC,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;YAAC,IAAI,GAAG,CAAC,IAAI,EAAE;gBAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;YAAC,GAAG,GAAG,EAAE,CAAC;YAAC,SAAS;QAAC,CAAC;QAC5H,GAAG,IAAI,EAAE,CAAC;IACZ,CAAC;IACD,IAAI,GAAG,CAAC,IAAI,EAAE;QAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;IACvC,MAAM,IAAI,GAAmB,EAAE,CAAC;IAChC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,CAAC,GAAG,6BAA6B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACnD,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC1B,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC3B,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC;IAChD,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,YAAY,CAAC,KAAa;IACjC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAC7B,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IACnC,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;QACd,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5C,IAAK,uBAA6C,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YAClE,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC9C,OAAO,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAgD,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAgD,EAAE,CAAC;QAChJ,CAAC;IACH,CAAC;IACD,8EAA8E;IAC9E,wBAAwB;IACxB,IAAK,uBAA6C,CAAC,QAAQ,CAAC,OAAO,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,OAAmD,EAAE,CAAC;IAC3I,IAAI,CAAC,OAAO;QAAE,MAAM,IAAI,kBAAkB,CAAC,iCAAiC,CAAC,CAAC;IAC9E,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC;AAC3B,CAAC;AAED,yEAAyE;AACzE,2EAA2E;AAC3E,qEAAqE;AACrE,SAAS,aAAa,CAAC,KAAa;IAClC,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAC3E,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC;IACpD,OAAO,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;AACjC,CAAC;AAED,MAAM,aAAa,GAAmE,EAAE,MAAM,EAAE,aAAa,EAAE,MAAM,EAAE,aAAa,EAAE,CAAC;AAEvI,SAAS,aAAa,CAAC,KAAa,EAAE,IAAY,EAAE,MAAkB;IACpE,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IAC9B,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,KAAK;YACR,OAAO,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC;QAC/C,KAAK,OAAO;YACV,OAAO,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC;QACnD,KAAK,KAAK,CAAC;QACX,KAAK,OAAO,CAAC,CAAC,CAAC;YACb,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;YACjC,uEAAuE;YACvE,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE;gBACnD,MAAM,GAAG,GAAiB,EAAE,IAAI,EAAE,IAAI,KAAK,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,eAAe,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAkB,CAAC;gBAC5H,IAAI,OAAO;oBAAG,GAA4B,CAAC,OAAO,GAAG,OAAO,CAAC;gBAC7D,OAAO,GAAG,CAAC;YACb,CAAC,CAAC,CAAC;QACL,CAAC;QACD,KAAK,MAAM;YACT,OAAO,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,cAAc,EAAE,MAAM,EAAE,CAAC,CAAC;QACjH;YACE,MAAM,IAAI,kBAAkB,CAAC,0BAA0B,KAAK,IAAI,CAAC,CAAC;IACtE,CAAC;AACH,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,EAAU,EAAE,KAAY;IAC7D,MAAM,MAAM,GAAe,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,UAAU,EAAE,KAAK,CAAC,eAAe,EAAE,QAAQ,EAAE,KAAK,CAAC,aAAa,EAAE,CAAC;IACpH,MAAM,OAAO,GAAmB,EAAE,CAAC;IACnC,MAAM,KAAK,GAAG,WAAW,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAE1C,IAAI,SAAS,GAAoD,IAAI,CAAC;IAEtE,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACxB,MAAM,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;QACrB,IAAI,CAAC,KAAK,EAAE;YAAE,SAAS;QAEvB,IAAI,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACvB,MAAM,GAAG,GAAG,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,EAA8C,CAAC;YAClF,MAAM,GAAG,GAAG,OAAO,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;YACzE,IAAI,GAAG;gBAAE,GAAG,CAAC,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;YACjC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAClB,SAAS,GAAG,GAAG,CAAC;YAChB,SAAS;QACX,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,EAAE,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACzB,IAAI,EAAE,EAAE,CAAC;gBACP,MAAM,KAAK,GAAG,EAAE,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;gBAC5B,MAAM,GAAG,GAAiB,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAkB,CAAC;gBAC9F,IAAI,KAAK;oBAAG,GAA0B,CAAC,KAAK,GAAG,KAAK,CAAC;gBACrD,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpB,CAAC;YACD,SAAS;QACX,CAAC;QACD,uEAAuE;QACvE,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,IAAI,SAAS,EAAE,CAAC;gBACd,MAAM,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBAC3B,IAAI,GAAG,EAAE,CAAC;oBACR,MAAM,CAAC,GAAsC,EAAE,CAAC;oBAChD,KAAK,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC;wBAAC,IAAI,SAAS,CAAC,GAAG,CAAC;4BAAE,CAAC,CAAC,KAAK,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;wBAAC,IAAI,QAAQ,CAAC,GAAG,CAAC;4BAAE,CAAC,CAAC,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;oBAAC,CAAC;oBAC3I,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI;wBAAE,SAAS,CAAC,WAAW,GAAG,CAAC,CAAC;gBACnD,CAAC;qBAAM,CAAC;oBACN,MAAM,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;oBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;oBAC7B,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;oBAC7B,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,EAAE,CAAC;wBAAC,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;wBAAC,SAAS,CAAC,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAK,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAK,CAAC,IAAI,EAAE,EAAE,CAAC;oBAAC,CAAC;yBACtK,IAAI,IAAI,IAAI,IAAI,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE;wBAAE,SAAS,CAAC,MAAM,GAAG,eAAe,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,CAAC;yBAC1E,IAAI,IAAI,IAAI,IAAI,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE;wBAAE,SAAS,CAAC,KAAK,GAAG,eAAe,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,CAAC;gBAChF,CAAC;YACH,CAAC;YACD,SAAS;QACX,CAAC;QACD,yEAAyE;QACzE,0CAA0C;QAC1C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACjC,IAAI,MAAM,IAAI,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC,CAAE,EAAE,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,MAAM,CAAC,CAAC,CAAC;;YACjI,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC;IAC5D,CAAC;IAED,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,UAAU,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,CAAC,CAAC,CAAC;IACpF,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,SAAS,IAAI,CAAC,IAAK,CAAyB,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;AAC1H,CAAC;AAED,SAAS,WAAW,CAAC,GAAW;IAC9B,MAAM,OAAO,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;IAC3B,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,wCAAwC,CAAC,CAAC;IACtE,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,CAAC,CAAC,OAAO,CAAC;AACrC,CAAC"}

package/dist/domain/ingest/records.d.ts

@@ -0,0 +1,71 @@
+export type SourceSpan = {
+    chunk: number;
+    line_start: number;
+    line_end: number;
+};
+export type NameStateRef = {
+    name: string;
+    state?: string;
+};
+export type SpeakerRef = {
+    name: string;
+    state?: string;
+} | {
+    kind: "system" | "broadcast" | "offscreen" | "group";
+    label?: string;
+};
+export declare const NON_ACTOR_SPEAKER_KINDS: readonly ["system", "broadcast", "offscreen", "group"];
+export type EpisodeRecord = {
+    kind: "episode";
+    episode: number;
+    title?: string;
+    source: SourceSpan;
+};
+export type SceneRecord = {
+    kind: "scene";
+    scene?: number;
+    environment?: {
+        space?: string;
+        time?: string;
+    };
+    location?: {
+        name: string;
+        state?: string;
+    };
+    actors?: NameStateRef[];
+    props?: NameStateRef[];
+    source: SourceSpan;
+};
+export type ActionRecord = {
+    kind: "action";
+    content: string;
+    source: SourceSpan;
+};
+export type DialogueRecord = {
+    kind: "dialogue";
+    speaker: SpeakerRef;
+    content: string;
+    emotion?: string;
+    source: SourceSpan;
+};
+export type InnerThoughtRecord = {
+    kind: "inner_thought";
+    speaker: SpeakerRef;
+    content: string;
+    emotion?: string;
+    source: SourceSpan;
+};
+export type TransitionRecord = {
+    kind: "transition";
+    content: string;
+    source: SourceSpan;
+};
+export type NoteRecord = {
+    kind: "note";
+    content: string;
+    category?: "author_note" | "format_note" | "unclassified";
+    source: SourceSpan;
+};
+export type IngestRecord = EpisodeRecord | SceneRecord | ActionRecord | DialogueRecord | InnerThoughtRecord | TransitionRecord | NoteRecord;
+export declare const RECORD_KINDS: readonly ["episode", "scene", "action", "dialogue", "inner_thought", "transition", "note"];
+export type RecordKind = (typeof RECORD_KINDS)[number];

package/dist/domain/ingest/records.js

@@ -0,0 +1,9 @@
+// Typed JSONL record shapes for the ingest pipeline (§7). Records are name-based
+// and pre-canonicalization; Pass5 assigns final ids. Content is plain text (no
+// inline entity tags); scene-level cast/props are declared on the scene record
+// (from source `人物：`/`道具：` lines) and augmented by the extract pass.
+// Non-actor speaker kinds. Bit-part characters (龙套) are still actors (tagged
+// importance=background), so they speak via actor_id — no special speaker kind.
+export const NON_ACTOR_SPEAKER_KINDS = ["system", "broadcast", "offscreen", "group"];
+export const RECORD_KINDS = ["episode", "scene", "action", "dialogue", "inner_thought", "transition", "note"];
+//# sourceMappingURL=records.js.map

package/dist/domain/ingest/records.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"records.js","sourceRoot":"","sources":["../../../src/domain/ingest/records.ts"],"names":[],"mappings":"AAAA,iFAAiF;AACjF,+EAA+E;AAC/E,+EAA+E;AAC/E,qEAAqE;AAWrE,6EAA6E;AAC7E,gFAAgF;AAChF,MAAM,CAAC,MAAM,uBAAuB,GAAG,CAAC,QAAQ,EAAE,WAAW,EAAE,WAAW,EAAE,OAAO,CAAU,CAAC;AA2B9F,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,SAAS,EAAE,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,CAAU,CAAC"}

package/dist/domain/ingest/renumber.d.ts

@@ -0,0 +1,14 @@
+type Obj = Record<string, unknown>;
+export type IdMap = {
+    episodes: Record<string, string>;
+    scenes: Record<string, string>;
+    actors: Record<string, string>;
+    locations: Record<string, string>;
+    props: Record<string, string>;
+    states: Record<string, string>;
+};
+export declare function renumberIds(input: unknown): {
+    script: Obj;
+    idMap: IdMap;
+};
+export {};

package/dist/domain/ingest/renumber.js

@@ -0,0 +1,130 @@
+// Deterministic final id renumbering (§ ids). Governance leaves survivor ids
+// with gaps (merges/…) and globally-unique state ids; this one-time pass, run at
+// the very end of ingest, renumbers everything to clean, contiguous, readable
+// ids and rewrites all references. ids are opaque stable keys — this is only the
+// clean initial numbering for the freshly-generated product (editing never
+// renumbers). Pure: input never mutated; returns the doc + old→new id map.
+//
+//   episode  ep_NNN   by episodes[] order
+//   scene    scn_NNN  global, by appearance order (scenes can move across eps)
+//   actor    act_NNN  by first appearance in the script (per kind, contiguous)
+//   location loc_NNN
+//   prop     prp_NNN
+//   state    default + st_NNN  per-asset scoped (default kept)
+const isObj = (v) => typeof v === "object" && v !== null && !Array.isArray(v);
+const isArr = (v) => Array.isArray(v);
+const pad = (n) => String(n).padStart(3, "0");
+const KINDS = [
+    { key: "actors", idField: "actor_id", prefix: "act" },
+    { key: "locations", idField: "location_id", prefix: "loc" },
+    { key: "props", idField: "prop_id", prefix: "prp" },
+];
+export function renumberIds(input) {
+    if (!isObj(input))
+        throw new Error("renumberIds expects a script object");
+    const doc = JSON.parse(JSON.stringify(input));
+    const idMap = { episodes: {}, scenes: {}, actors: {}, locations: {}, props: {}, states: {} };
+    // 1. Asset ids by first appearance (document walk).
+    const seen = { actors: [], locations: [], props: [] };
+    const seenSet = { actors: new Set(), locations: new Set(), props: new Set() };
+    const see = (bucket, id) => { if (typeof id === "string" && !seenSet[bucket].has(id)) {
+        seenSet[bucket].add(id);
+        seen[bucket].push(id);
+    } };
+    for (const ep of isArr(doc.episodes) ? doc.episodes : []) {
+        if (!isObj(ep) || !isArr(ep.scenes))
+            continue;
+        for (const sc of ep.scenes) {
+            if (!isObj(sc))
+                continue;
+            if (isObj(sc.location))
+                see("locations", sc.location.location_id);
+            for (const a of isArr(sc.actors) ? sc.actors : [])
+                if (isObj(a))
+                    see("actors", a.actor_id);
+            for (const p of isArr(sc.props) ? sc.props : [])
+                if (isObj(p))
+                    see("props", p.prop_id);
+            for (const ac of isArr(sc.actions) ? sc.actions : [])
+                if (isObj(ac) && isObj(ac.speaker))
+                    see("actors", ac.speaker.actor_id);
+        }
+    }
+    // Append any asset present in the top-level lists but never referenced.
+    for (const { key, idField } of KINDS)
+        for (const a of isArr(doc[key]) ? doc[key] : [])
+            if (isObj(a))
+                see(key, a[idField]);
+    for (const { key, prefix } of KINDS)
+        seen[key].forEach((old, i) => { idMap[key][old] = `${prefix}_${pad(i + 1)}`; });
+    // 2. Rebuild top-level asset lists in new-id order; renumber per-asset states.
+    for (const { key, idField } of KINDS) {
+        const list = (isArr(doc[key]) ? doc[key] : []).filter(isObj);
+        const bucket = idMap[key];
+        const byNew = new Map();
+        for (const a of list) {
+            const oldId = a[idField];
+            const newId = bucket[oldId] ?? oldId;
+            a[idField] = newId;
+            // states: keep default, renumber the rest st_001.. in array order.
+            let n = 0;
+            for (const st of isArr(a.states) ? a.states : []) {
+                if (!isObj(st) || typeof st.state_id !== "string")
+                    continue;
+                if (st.state_id === "default")
+                    continue;
+                const ns = `st_${pad(++n)}`;
+                idMap.states[st.state_id] = ns;
+                st.state_id = ns;
+            }
+            byNew.set(newId, a);
+        }
+        doc[key] = [...byNew.entries()].sort((x, y) => (x[0] < y[0] ? -1 : 1)).map((e) => e[1]);
+    }
+    // 3. Episode / scene ids (in order).
+    let scn = 0;
+    (isArr(doc.episodes) ? doc.episodes : []).forEach((ep, ei) => {
+        if (!isObj(ep))
+            return;
+        if (typeof ep.episode_id === "string") {
+            const nid = `ep_${pad(ei + 1)}`;
+            idMap.episodes[ep.episode_id] = nid;
+            ep.episode_id = nid;
+        }
+        for (const sc of isArr(ep.scenes) ? ep.scenes : []) {
+            if (!isObj(sc))
+                continue;
+            if (typeof sc.scene_id === "string") {
+                const nid = `scn_${pad(++scn)}`;
+                idMap.scenes[sc.scene_id] = nid;
+                sc.scene_id = nid;
+            }
+        }
+    });
+    // 4. Rewrite references (asset ids + state ids) throughout the scenes.
+    const remapRef = (holder, idField, kind) => {
+        if (!isObj(holder) || typeof holder[idField] !== "string")
+            return;
+        holder[idField] = idMap[kind][holder[idField]] ?? holder[idField];
+        if (typeof holder.state_id === "string" && idMap.states[holder.state_id])
+            holder.state_id = idMap.states[holder.state_id];
+    };
+    for (const ep of isArr(doc.episodes) ? doc.episodes : []) {
+        if (!isObj(ep) || !isArr(ep.scenes))
+            continue;
+        for (const sc of ep.scenes) {
+            if (!isObj(sc))
+                continue;
+            remapRef(sc.location, "location_id", "locations");
+            for (const a of isArr(sc.actors) ? sc.actors : [])
+                remapRef(a, "actor_id", "actors");
+            for (const p of isArr(sc.props) ? sc.props : [])
+                remapRef(p, "prop_id", "props");
+            for (const ac of isArr(sc.actions) ? sc.actions : [])
+                if (isObj(ac))
+                    remapRef(ac.speaker, "actor_id", "actors");
+        }
+    }
+    return { script: doc, idMap };
+}
+//# sourceMappingURL=renumber.js.map