@wanshi-kg/wanshi 0.1.0

package/dist/core/processor/classifier/mergeClassifications.js

@@ -0,0 +1,32 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.mergeChunkClassifications = mergeChunkClassifications;
+/**
+ * Merge per-chunk classification results into one prevalence-weighted ranking.
+ *
+ * Each class's confidence is summed across the chunks it appears in, then divided
+ * by the **total** number of chunks — chunks where a class is absent contribute 0.
+ * So a class present in every chunk outranks a single high-confidence spike, which
+ * matches "what domain is this whole document".
+ *
+ * This replaces the earlier pairwise-reduce that averaged over *present* chunks
+ * only (`sum / resultsByClass.length`): there a class in 1/10 chunks @0.9 beat a
+ * class in 10/10 @0.8, letting one off-topic chunk hijack the file's domain.
+ */
+function mergeChunkClassifications(perChunk) {
+    var _a;
+    const totalChunks = perChunk.length;
+    if (totalChunks === 0)
+        return [];
+    const sums = new Map();
+    for (const chunkResults of perChunk) {
+        for (const { class: cls, confidence } of chunkResults) {
+            sums.set(cls, ((_a = sums.get(cls)) !== null && _a !== void 0 ? _a : 0) + confidence);
+        }
+    }
+    return Array.from(sums, ([cls, sum]) => ({
+        class: cls,
+        confidence: sum / totalChunks,
+    })).sort((a, b) => b.confidence - a.confidence);
+}
+//# sourceMappingURL=mergeClassifications.js.map

package/dist/core/processor/classifier/mergeClassifications.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"mergeClassifications.js","sourceRoot":"","sources":["../../../../src/core/processor/classifier/mergeClassifications.ts"],"names":[],"mappings":";;AAcA,8DAiBC;AA7BD;;;;;;;;;;;GAWG;AACH,SAAgB,yBAAyB,CACvC,QAAkC;;IAElC,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,CAAC;IACpC,IAAI,WAAW,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAwB,CAAC;IAC7C,KAAK,MAAM,YAAY,IAAI,QAAQ,EAAE,CAAC;QACpC,KAAK,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,YAAY,EAAE,CAAC;YACtD,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,MAAA,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,mCAAI,CAAC,CAAC,GAAG,UAAU,CAAC,CAAC;QACnD,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC;QACvC,KAAK,EAAE,GAAG;QACV,UAAU,EAAE,GAAG,GAAG,WAAW;KAC9B,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC;AAClD,CAAC"}

package/dist/core/processor/index.js

@@ -0,0 +1,20 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __exportStar = (this && this.__exportStar) || function(m, exports) {
+    for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+__exportStar(require("./FileProcessor"), exports);
+__exportStar(require("./chunking"), exports);
+__exportStar(require("./readers"), exports);
+//# sourceMappingURL=index.js.map

package/dist/core/processor/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/core/processor/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,kDAAgC;AAChC,6CAA2B;AAC3B,4CAA0B"}

package/dist/core/processor/readers/AudioReader.js

@@ -0,0 +1,462 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.AudioReader = void 0;
+const FileReader_1 = require("./FileReader");
+const path_1 = __importDefault(require("path"));
+const promises_1 = __importDefault(require("fs/promises"));
+const fluent_ffmpeg_1 = __importDefault(require("fluent-ffmpeg"));
+const nodejs_whisper_1 = require("nodejs-whisper");
+const util_1 = require("util");
+const promises_2 = require("fs/promises");
+const child_process_1 = require("child_process");
+const turnPacking_1 = require("./transcript/turnPacking");
+/**
+ * Reader for audio/video files with speech transcription
+ *
+ * This implementation uses:
+ * - fluent-ffmpeg for audio/video preprocessing and metadata extraction
+ * - nodejs-whisper for speech-to-text transcription with timestamps
+ *
+ * Required dependencies:
+ * npm install fluent-ffmpeg nodejs-whisper ffmpeg-static ffprobe-static
+ *
+ * System requirements:
+ * - FFmpeg binary (automatically handled by ffmpeg-static)
+ * - Whisper models (automatically downloaded by nodejs-whisper)
+ *
+ * Supported formats: .mp3, .mp4, .wav, .ogg, .m4a, .flac, .aac, .webm, .mkv, .avi
+ */
+class AudioReader extends FileReader_1.FileReader {
+    constructor(options = {
+        modelName: "medium",
+        language: "auto",
+        translate: false,
+    }, tempDir = "./temp", chunker, logger) {
+        super([
+            ".mp3",
+            ".mp4",
+            ".wav",
+            ".ogg",
+            ".m4a",
+            ".flac",
+            ".aac",
+            ".webm",
+            ".mkv",
+            ".mov",
+            ".avi",
+        ], chunker, logger);
+        this.tempDir = tempDir;
+        this.options = options;
+        this.ensureTempDir();
+    }
+    getName() {
+        return "AudioReader";
+    }
+    read(filePath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            yield this.validateFile(filePath);
+            try {
+                this.logger.debug(`Reading audio/video file: ${filePath}`);
+                const stats = yield (0, promises_2.stat)(filePath);
+                const ext = path_1.default.extname(filePath).toLowerCase();
+                // Step 1: Extract audio metadata using ffprobe
+                const audioMetadata = yield this.extractAudioMetadata(filePath);
+                // Step 1b: dual engine — VAD + dual-STT + diarization via the vendored
+                // Python pipeline. Any failure degrades gracefully to the whisper path.
+                if (this.options.engine === "dual") {
+                    try {
+                        return yield this.transcribeDual(filePath, ext, stats, audioMetadata);
+                    }
+                    catch (error) {
+                        this.logger.warn(`Dual ASR engine failed for ${filePath} (${error.message}); falling back to the whisper engine`);
+                    }
+                }
+                // Step 2: Convert to WAV for Whisper if needed
+                const processedAudioPath = yield this.preprocessAudio(filePath);
+                // Step 3: Transcribe using Whisper
+                const transcriptionResult = yield this.transcribeAudio(processedAudioPath, audioMetadata);
+                // Step 4: Clean up temporary files
+                yield this.cleanup(processedAudioPath, filePath);
+                // Build comprehensive metadata
+                const metadata = Object.assign(Object.assign(Object.assign({ type: this.getMediaType(ext), description: this.getMediaDescription(ext), fileName: path_1.default.basename(filePath), filePath: filePath, fileSize: stats.size, createdAt: stats.birthtime.toISOString(), modifiedAt: stats.mtime.toISOString(), extension: ext }, audioMetadata), transcriptionResult.metadata), { status: "success", processingTime: transcriptionResult.processingTime, transcriptionLength: transcriptionResult.content.length, hasTranscription: transcriptionResult.content.trim().length > 0 });
+                this.logger.debug(`Successfully transcribed ${filePath}: ${transcriptionResult.content.length} characters`);
+                return {
+                    chunks: yield this.chunker.chunk(transcriptionResult.content),
+                    metadata: metadata,
+                };
+            }
+            catch (error) {
+                this.logger.error(`Failed to read audio file ${filePath}: ${error.message}`);
+                return {
+                    chunks: [],
+                    metadata: {
+                        type: this.getMediaType(path_1.default.extname(filePath).toLowerCase()),
+                        description: this.getMediaDescription(path_1.default.extname(filePath)),
+                        fileName: path_1.default.basename(filePath),
+                        filePath: filePath,
+                        status: "error",
+                        error: error.message,
+                        errorType: error.name,
+                        processingStep: error.step || "unknown",
+                    },
+                };
+            }
+        });
+    }
+    /**
+     * Extract audio metadata using ffprobe
+     */
+    extractAudioMetadata(filePath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            var _a, _b, _c, _d, _e, _f;
+            try {
+                const ffprobe = (0, util_1.promisify)(fluent_ffmpeg_1.default.ffprobe);
+                const metadata = (yield ffprobe(filePath));
+                const audioStream = metadata.streams.find(
+                // @ts-expect-error
+                (stream) => stream.codec_type === "audio");
+                const videoStream = metadata.streams.find(
+                // @ts-expect-error
+                (stream) => stream.codec_type === "video");
+                return {
+                    // General file metadata
+                    duration: parseFloat(metadata.format.duration || "0"),
+                    bitrate: parseInt(metadata.format.bit_rate || "0"),
+                    formatName: metadata.format.format_name,
+                    formatLongName: metadata.format.format_long_name,
+                    // Audio stream metadata
+                    audioCodec: (audioStream === null || audioStream === void 0 ? void 0 : audioStream.codec_name) || "unknown",
+                    audioSampleRate: parseInt((audioStream === null || audioStream === void 0 ? void 0 : audioStream.sample_rate) || "0"),
+                    audioChannels: (audioStream === null || audioStream === void 0 ? void 0 : audioStream.channels) || 0,
+                    audioChannelLayout: (audioStream === null || audioStream === void 0 ? void 0 : audioStream.channel_layout) || "unknown",
+                    audioBitrate: parseInt((audioStream === null || audioStream === void 0 ? void 0 : audioStream.bit_rate) || "0"),
+                    // Video metadata (if present)
+                    hasVideo: !!videoStream,
+                    videoCodec: (videoStream === null || videoStream === void 0 ? void 0 : videoStream.codec_name) || null,
+                    videoResolution: videoStream
+                        ? `${videoStream.width}x${videoStream.height}`
+                        : null,
+                    videoFrameRate: (videoStream === null || videoStream === void 0 ? void 0 : videoStream.r_frame_rate) || null,
+                    // Metadata tags
+                    title: ((_a = metadata.format.tags) === null || _a === void 0 ? void 0 : _a.title) || "",
+                    artist: ((_b = metadata.format.tags) === null || _b === void 0 ? void 0 : _b.artist) || "",
+                    album: ((_c = metadata.format.tags) === null || _c === void 0 ? void 0 : _c.album) || "",
+                    genre: ((_d = metadata.format.tags) === null || _d === void 0 ? void 0 : _d.genre) || "",
+                    date: ((_e = metadata.format.tags) === null || _e === void 0 ? void 0 : _e.date) || "",
+                    comment: ((_f = metadata.format.tags) === null || _f === void 0 ? void 0 : _f.comment) || "",
+                };
+            }
+            catch (error) {
+                this.logger.warn(`Could not extract audio metadata: ${error.message}`);
+                return {
+                    duration: 0,
+                    bitrate: 0,
+                    formatName: "unknown",
+                    audioCodec: "unknown",
+                    audioSampleRate: 0,
+                    audioChannels: 0,
+                };
+            }
+        });
+    }
+    /**
+     * Convert audio to optimal format for Whisper
+     */
+    preprocessAudio(filePath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const ext = path_1.default.extname(filePath).toLowerCase();
+            // If it's already a WAV file, check if it needs resampling
+            if (ext === ".wav") {
+                const metadata = yield this.extractAudioMetadata(filePath);
+                if (metadata.audioSampleRate === 16000 && metadata.audioChannels <= 2) {
+                    this.logger.debug("WAV file already in optimal format, skipping preprocessing");
+                    return filePath;
+                }
+            }
+            const outputPath = path_1.default.resolve(path_1.default.join(this.tempDir, `${path_1.default.basename(filePath, ext)}_processed_${Date.now()}.wav`));
+            return new Promise((resolve, reject) => {
+                this.logger.debug(`Converting ${filePath} to WAV format for Whisper`);
+                (0, fluent_ffmpeg_1.default)(filePath)
+                    .audioFrequency(16000) // Whisper loves 16kHz
+                    .audioChannels(1) // Convert to mono for better performance
+                    .audioCodec("pcm_s16le") // Uncompressed PCM
+                    .format("wav")
+                    .on("start", (commandLine) => {
+                    this.logger.debug(`FFmpeg process started: ${commandLine}`);
+                })
+                    .on("progress", (progress) => {
+                    if (progress.percent) {
+                        this.logger.debug(`Audio conversion progress: ${Math.round(progress.percent)}%`);
+                    }
+                })
+                    .on("error", (error) => {
+                    // @ts-expect-error
+                    error.step = "audio_preprocessing";
+                    reject(error);
+                })
+                    .on("end", () => {
+                    this.logger.debug(`Audio conversion completed: ${outputPath}`);
+                    resolve(outputPath);
+                })
+                    .save(outputPath);
+            });
+        });
+    }
+    /**
+     * Transcribe audio using Whisper
+     */
+    transcribeAudio(audioPath, audioMetadata) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const startTime = Date.now();
+            try {
+                this.logger.debug(`Transcribing with Whisper model: ${this.options.modelName}`);
+                const options = {
+                    modelName: this.options.modelName,
+                    removeWavFileAfterTranscription: false, // We'll handle cleanup ourselves
+                    autoDownloadModelName: this.options.modelName,
+                    // logger: {
+                    //   debug: (...args: any[]) => this.logger.debug(args),
+                    //   error: (...args: any[]) => this.logger.error(args),
+                    //   log: (...args: any[]) => this.logger.info(args),
+                    // },
+                    whisperOptions: {
+                        language: this.options.language, // Auto-detect language
+                        outputInText: true,
+                        outputInJson: false, // Get detailed results
+                        outputInSrt: false, // Get timestamps
+                        wordTimestamps: false, // Word-level timestamps
+                        translateToEnglish: this.options.translate,
+                        splitOnWord: false,
+                    },
+                };
+                // Run Whisper transcription
+                const content = yield (0, nodejs_whisper_1.nodewhisper)(audioPath, options);
+                const processingTime = Date.now() - startTime;
+                // Build transcription metadata
+                const transcriptionMetadata = {
+                    whisperModel: this.options.modelName,
+                    detectedLanguage: content.language || "unknown",
+                    transcriptionConfidence: content.confidence || null,
+                    wordCount: content.split(/\s+/).filter((word) => word.length > 0).length,
+                    hasTimestamps: options.whisperOptions.wordTimestamps,
+                    processingTimeMs: processingTime,
+                };
+                return {
+                    content: content.trim(),
+                    metadata: transcriptionMetadata,
+                    processingTime: processingTime,
+                };
+            }
+            catch (error) {
+                error.step = "transcription";
+                throw error;
+            }
+        });
+    }
+    /**
+     * Dual engine: transcribe via the vendored Python `audio-pipeline` subproject
+     * (Silero VAD → diarization → Parakeet/Whisper dual-STT → max-overlap merge).
+     *
+     * The subprocess writes a recua turns JSON sidecar (`<audio>.transcript.json`);
+     * we reuse it if already present (re-runs are cheap), parse it into speaker
+     * turns, and pack them with the shared `packTurns` so the chunks carry the same
+     * speaker provenance as `TranscriptReader`. Throws on any failure so `read()`
+     * can fall back to the whisper engine.
+     */
+    transcribeDual(filePath, ext, stats, audioMetadata) {
+        return __awaiter(this, void 0, void 0, function* () {
+            var _a, _b, _c, _d, _e;
+            const startTime = Date.now();
+            const sidecar = `${filePath}.transcript.json`;
+            let cached = false;
+            if (yield this.sidecarIsFresh(sidecar, filePath)) {
+                this.logger.debug(`Reusing dual-ASR sidecar: ${sidecar}`);
+                cached = true;
+            }
+            else {
+                yield this.runAudioPipeline(filePath, sidecar);
+            }
+            const raw = yield promises_1.default.readFile(sidecar, "utf-8");
+            const parsed = JSON.parse(raw);
+            if (!Array.isArray(parsed)) {
+                throw new Error(`audio-pipeline sidecar is not a turns array: ${sidecar}`);
+            }
+            const turns = parsed
+                .map((t) => { var _a; return ({ speaker: String((_a = t.speaker) !== null && _a !== void 0 ? _a : "UNKNOWN"), text: this.pickTurnText(t) }); })
+                .filter((t) => t.text);
+            if (turns.length === 0) {
+                throw new Error(`audio-pipeline produced no usable turns: ${sidecar}`);
+            }
+            const maxChunkSize = (_a = this.options.maxChunkSize) !== null && _a !== void 0 ? _a : 4000;
+            const chunks = yield (0, turnPacking_1.packTurns)(turns, filePath, maxChunkSize, this.chunker);
+            const speakers = new Set(turns.map((t) => t.speaker)).size;
+            const metadata = Object.assign(Object.assign({ type: this.getMediaType(ext), description: this.getMediaDescription(ext), fileName: path_1.default.basename(filePath), filePath, fileSize: stats.size, createdAt: stats.birthtime.toISOString(), modifiedAt: stats.mtime.toISOString(), extension: ext }, audioMetadata), { asrEngine: "dual", asrModels: (_c = (_b = this.options.dual) === null || _b === void 0 ? void 0 : _b.asr) !== null && _c !== void 0 ? _c : "both", diarized: (_e = (_d = this.options.dual) === null || _d === void 0 ? void 0 : _d.diarize) !== null && _e !== void 0 ? _e : true, speakerCount: speakers, turns: turns.length, sidecar, sidecarCached: cached, status: "success", processingTime: Date.now() - startTime });
+            this.logger.debug(`Dual ASR transcribed ${filePath}: ${turns.length} turns, ${speakers} speaker(s), ${chunks.length} chunk(s)`);
+            return { chunks, metadata };
+        });
+    }
+    /** Prefer corrected > parakeet > whisper > any other string field. */
+    pickTurnText(t) {
+        for (const k of ["corrected", "parakeet", "whisper"]) {
+            const v = t[k];
+            if (typeof v === "string" && v.trim())
+                return v.trim();
+        }
+        for (const [k, v] of Object.entries(t)) {
+            if (k === "start" || k === "end" || k === "speaker")
+                continue;
+            if (typeof v === "string" && v.trim())
+                return v.trim();
+        }
+        return "";
+    }
+    /** A sidecar is reusable if it exists and is newer than the source audio. */
+    sidecarIsFresh(sidecar, filePath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            try {
+                const [s, a] = yield Promise.all([(0, promises_2.stat)(sidecar), (0, promises_2.stat)(filePath)]);
+                return s.mtimeMs >= a.mtimeMs;
+            }
+            catch (_a) {
+                return false;
+            }
+        });
+    }
+    /**
+     * Spawn the Python audio-pipeline (DoclingReader subprocess pattern: captured
+     * stdout/stderr, `PYTHONUNBUFFERED`, hard timeout). Default launcher is
+     * `uv run --project <dir> python -m audio_pipeline …`; override the executable
+     * with `dual.pythonPath`. Rejects on non-zero exit / spawn error / timeout.
+     */
+    runAudioPipeline(filePath, outPath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const dual = this.options.dual;
+            if (!dual)
+                throw new Error("dual engine options missing");
+            const audioAbs = path_1.default.resolve(filePath);
+            const outAbs = path_1.default.resolve(outPath);
+            const cliArgs = ["transcribe", audioAbs, "--out", outAbs, "--asr", dual.asr];
+            if (!dual.diarize)
+                cliArgs.push("--no-diarize");
+            if (dual.numSpeakers)
+                cliArgs.push("--num-speakers", String(dual.numSpeakers));
+            if (dual.device)
+                cliArgs.push("--device", dual.device);
+            let command;
+            let args;
+            if (dual.pythonPath) {
+                command = dual.pythonPath;
+                args = ["-m", "audio_pipeline", ...cliArgs];
+            }
+            else {
+                command = "uv";
+                args = ["run", "--project", dual.projectDir, "python", "-m", "audio_pipeline", ...cliArgs];
+            }
+            this.logger.info(`Dual ASR: ${command} ${args.join(" ")}`);
+            yield new Promise((resolve, reject) => {
+                var _a, _b;
+                const child = (0, child_process_1.spawn)(command, args, {
+                    cwd: path_1.default.resolve(dual.projectDir),
+                    stdio: ["ignore", "pipe", "pipe"],
+                    env: Object.assign(Object.assign({}, process.env), { PYTHONUNBUFFERED: "1" }),
+                });
+                let stderr = "";
+                (_a = child.stdout) === null || _a === void 0 ? void 0 : _a.on("data", (d) => this.logger.debug(`[audio-pipeline] ${d.toString().trim()}`));
+                (_b = child.stderr) === null || _b === void 0 ? void 0 : _b.on("data", (d) => {
+                    stderr += d.toString();
+                    this.logger.debug(`[audio-pipeline] ${d.toString().trim()}`);
+                });
+                const timer = setTimeout(() => {
+                    child.kill("SIGTERM");
+                    reject(new Error(`audio-pipeline timed out after ${dual.timeoutMs}ms`));
+                }, dual.timeoutMs);
+                child.on("error", (err) => {
+                    clearTimeout(timer);
+                    reject(new Error(`failed to launch audio-pipeline (${command}): ${err.message}`));
+                });
+                child.on("close", (code) => {
+                    clearTimeout(timer);
+                    if (code === 0)
+                        resolve();
+                    else
+                        reject(new Error(`audio-pipeline exited ${code}${stderr ? `: ${stderr.trim().slice(-400)}` : ""}`));
+                });
+            });
+        });
+    }
+    /**
+     * Clean up temporary files
+     */
+    cleanup(processedPath, originalPath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            try {
+                // Only delete if it's a temporary file we created
+                if (processedPath !== originalPath &&
+                    processedPath.includes(this.tempDir)) {
+                    yield promises_1.default.unlink(processedPath);
+                    this.logger.debug(`Cleaned up temporary file: ${processedPath}`);
+                }
+            }
+            catch (error) {
+                this.logger.warn(`Could not clean up temporary file ${processedPath}: ${error}`);
+            }
+        });
+    }
+    /**
+     * Ensure temp directory exists
+     */
+    ensureTempDir() {
+        return __awaiter(this, void 0, void 0, function* () {
+            try {
+                yield promises_1.default.mkdir(this.tempDir, { recursive: true });
+            }
+            catch (error) {
+                this.logger.warn(`Could not create temp directory ${this.tempDir}: ${error}`);
+            }
+        });
+    }
+    /**
+     * Get media type based on extension
+     */
+    getMediaType(ext) {
+        const audioExts = [".mp3", ".wav", ".ogg", ".m4a", ".flac", ".aac"];
+        const videoExts = [".mp4", ".webm", ".mkv", ".avi"];
+        if (audioExts.includes(ext))
+            return "audio";
+        if (videoExts.includes(ext))
+            return "video";
+        return "media";
+    }
+    /**
+     * Get human-readable description
+     */
+    getMediaDescription(ext) {
+        const descriptions = {
+            ".mp3": "MP3 Audio File",
+            ".mp4": "MP4 Video File",
+            ".wav": "WAV Audio File",
+            ".ogg": "OGG Audio File",
+            ".m4a": "M4A Audio File",
+            ".flac": "FLAC Audio File",
+            ".aac": "AAC Audio File",
+            ".webm": "WebM Video File",
+            ".mkv": "Matroska Video File",
+            ".avi": "AVI Video File",
+        };
+        return descriptions[ext] || "Media File";
+    }
+}
+exports.AudioReader = AudioReader;
+//# sourceMappingURL=AudioReader.js.map

package/dist/core/processor/readers/AudioReader.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"AudioReader.js","sourceRoot":"","sources":["../../../../src/core/processor/readers/AudioReader.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,6CAA0D;AAC1D,gDAAwB;AACxB,2DAA6B;AAC7B,kEAAmC;AACnC,mDAA6C;AAC7C,+BAAiC;AACjC,0CAAmC;AACnC,iDAAsC;AAGtC,0DAA2D;AAE3D;;;;;;;;;;;;;;;GAeG;AACH,MAAa,WAAY,SAAQ,uBAAU;IAIzC,YACE,UAAkC;QAChC,SAAS,EAAE,QAAQ;QACnB,QAAQ,EAAE,MAAM;QAChB,SAAS,EAAE,KAAK;KACjB,EACD,OAAO,GAAG,QAAQ,EAClB,OAAoB,EACpB,MAAc;QAEd,KAAK,CACH;YACE,MAAM;YACN,MAAM;YACN,MAAM;YACN,MAAM;YACN,MAAM;YACN,OAAO;YACP,MAAM;YACN,OAAO;YACP,MAAM;YACN,MAAM;YACN,MAAM;SACP,EACD,OAAO,EACP,MAAM,CACP,CAAC;QACF,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;QACvB,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;QACvB,IAAI,CAAC,aAAa,EAAE,CAAC;IACvB,CAAC;IAED,OAAO;QACL,OAAO,aAAa,CAAC;IACvB,CAAC;IAEK,IAAI,CAAC,QAAgB;;YACzB,MAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;YAElC,IAAI,CAAC;gBACH,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,6BAA6B,QAAQ,EAAE,CAAC,CAAC;gBAE3D,MAAM,KAAK,GAAG,MAAM,IAAA,eAAI,EAAC,QAAQ,CAAC,CAAC;gBACnC,MAAM,GAAG,GAAG,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;gBAEjD,+CAA+C;gBAC/C,MAAM,aAAa,GAAG,MAAM,IAAI,CAAC,oBAAoB,CAAC,QAAQ,CAAC,CAAC;gBAEhE,uEAAuE;gBACvE,wEAAwE;gBACxE,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,MAAM,EAAE,CAAC;oBACnC,IAAI,CAAC;wBACH,OAAO,MAAM,IAAI,CAAC,cAAc,CAAC,QAAQ,EAAE,GAAG,EAAE,KAAK,EAAE,aAAa,CAAC,CAAC;oBACxE,CAAC;oBAAC,OAAO,KAAU,EAAE,CAAC;wBACpB,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,8BAA8B,QAAQ,KAAK,KAAK,CAAC,OAAO,uCAAuC,CAChG,CAAC;oBACJ,CAAC;gBACH,CAAC;gBAED,+CAA+C;gBAC/C,MAAM,kBAAkB,GAAG,MAAM,IAAI,CAAC,eAAe,CAAC,QAAQ,CAAC,CAAC;gBAEhE,mCAAmC;gBACnC,MAAM,mBAAmB,GAAG,MAAM,IAAI,CAAC,eAAe,CACpD,kBAAkB,EAClB,aAAa,CACd,CAAC;gBAEF,mCAAmC;gBACnC,MAAM,IAAI,CAAC,OAAO,CAAC,kBAAkB,EAAE,QAAQ,CAAC,CAAC;gBAEjD,+BAA+B;gBAC/B,MAAM,QAAQ,+CACZ,IAAI,EAAE,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAC5B,WAAW,EAAE,IAAI,CAAC,mBAAmB,CAAC,GAAG,CAAC,EAC1C,QAAQ,EAAE,cAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EACjC,QAAQ,EAAE,QAAQ,EAClB,QAAQ,EAAE,KAAK,CAAC,IAAI,EACpB,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,WAAW,EAAE,EACxC,UAAU,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,EACrC,SAAS,EAAE,GAAG,IAGX,aAAa,GAGb,mBAAmB,CAAC,QAAQ,KAE/B,MAAM,EAAE,SAAS,EACjB,cAAc,EAAE,mBAAmB,CAAC,cAAc,EAClD,mBAAmB,EAAE,mBAAmB,CAAC,OAAO,CAAC,MAAM,EACvD,gBAAgB,EAAE,mBAAmB,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,GAChE,CAAC;gBAEF,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,4BAA4B,QAAQ,KAAK,mBAAmB,CAAC,OAAO,CAAC,MAAM,aAAa,CACzF,CAAC;gBAEF,OAAO;oBACL,MAAM,EAAE,MAAM,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,mBAAmB,CAAC,OAAO,CAAC;oBAC7D,QAAQ,EAAE,QAAQ;iBACnB,CAAC;YACJ,CAAC;YAAC,OAAO,KAAU,EAAE,CAAC;gBACpB,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,6BAA6B,QAAQ,KAAK,KAAK,CAAC,OAAO,EAAE,CAC1D,CAAC;gBAEF,OAAO;oBACL,MAAM,EAAE,EAAE;oBACV,QAAQ,EAAE;wBACR,IAAI,EAAE,IAAI,CAAC,YAAY,CAAC,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;wBAC7D,WAAW,EAAE,IAAI,CAAC,mBAAmB,CAAC,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;wBAC7D,QAAQ,EAAE,cAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;wBACjC,QAAQ,EAAE,QAAQ;wBAClB,MAAM,EAAE,OAAO;wBACf,KAAK,EAAE,KAAK,CAAC,OAAO;wBACpB,SAAS,EAAE,KAAK,CAAC,IAAI;wBACrB,cAAc,EAAE,KAAK,CAAC,IAAI,IAAI,SAAS;qBACxC;iBACF,CAAC;YACJ,CAAC;QACH,CAAC;KAAA;IAED;;OAEG;IACW,oBAAoB,CAAC,QAAgB;;;YACjD,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,IAAA,gBAAS,EAAC,uBAAM,CAAC,OAAO,CAAC,CAAC;gBAC1C,MAAM,QAAQ,GAAG,CAAC,MAAM,OAAO,CAAC,QAAQ,CAAC,CAAQ,CAAC;gBAElD,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI;gBACvC,mBAAmB;gBACnB,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,KAAK,OAAO,CAC1C,CAAC;gBACF,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI;gBACvC,mBAAmB;gBACnB,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,KAAK,OAAO,CAC1C,CAAC;gBAEF,OAAO;oBACL,wBAAwB;oBACxB,QAAQ,EAAE,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,QAAQ,IAAI,GAAG,CAAC;oBACrD,OAAO,EAAE,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,QAAQ,IAAI,GAAG,CAAC;oBAClD,UAAU,EAAE,QAAQ,CAAC,MAAM,CAAC,WAAW;oBACvC,cAAc,EAAE,QAAQ,CAAC,MAAM,CAAC,gBAAgB;oBAEhD,wBAAwB;oBACxB,UAAU,EAAE,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,UAAU,KAAI,SAAS;oBAChD,eAAe,EAAE,QAAQ,CAAC,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,WAAW,KAAI,GAAG,CAAC;oBAC1D,aAAa,EAAE,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,QAAQ,KAAI,CAAC;oBACzC,kBAAkB,EAAE,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,cAAc,KAAI,SAAS;oBAC5D,YAAY,EAAE,QAAQ,CAAC,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,QAAQ,KAAI,GAAG,CAAC;oBAEpD,8BAA8B;oBAC9B,QAAQ,EAAE,CAAC,CAAC,WAAW;oBACvB,UAAU,EAAE,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,UAAU,KAAI,IAAI;oBAC3C,eAAe,EAAE,WAAW;wBAC1B,CAAC,CAAC,GAAG,WAAW,CAAC,KAAK,IAAI,WAAW,CAAC,MAAM,EAAE;wBAC9C,CAAC,CAAC,IAAI;oBACR,cAAc,EAAE,CAAA,WAAW,aAAX,WAAW,uBAAX,WAAW,CAAE,YAAY,KAAI,IAAI;oBAEjD,gBAAgB;oBAChB,KAAK,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,KAAK,KAAI,EAAE;oBACxC,MAAM,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,MAAM,KAAI,EAAE;oBAC1C,KAAK,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,KAAK,KAAI,EAAE;oBACxC,KAAK,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,KAAK,KAAI,EAAE;oBACxC,IAAI,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,IAAI,KAAI,EAAE;oBACtC,OAAO,EAAE,CAAA,MAAA,QAAQ,CAAC,MAAM,CAAC,IAAI,0CAAE,OAAO,KAAI,EAAE;iBAC7C,CAAC;YACJ,CAAC;YAAC,OAAO,KAAU,EAAE,CAAC;gBACpB,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,qCAAqC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;gBACvE,OAAO;oBACL,QAAQ,EAAE,CAAC;oBACX,OAAO,EAAE,CAAC;oBACV,UAAU,EAAE,SAAS;oBACrB,UAAU,EAAE,SAAS;oBACrB,eAAe,EAAE,CAAC;oBAClB,aAAa,EAAE,CAAC;iBACjB,CAAC;YACJ,CAAC;QACH,CAAC;KAAA;IAED;;OAEG;IACW,eAAe,CAAC,QAAgB;;YAC5C,MAAM,GAAG,GAAG,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;YAEjD,2DAA2D;YAC3D,IAAI,GAAG,KAAK,MAAM,EAAE,CAAC;gBACnB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,oBAAoB,CAAC,QAAQ,CAAC,CAAC;gBAC3D,IAAI,QAAQ,CAAC,eAAe,KAAK,KAAK,IAAI,QAAQ,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC;oBACtE,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,4DAA4D,CAC7D,CAAC;oBACF,OAAO,QAAQ,CAAC;gBAClB,CAAC;YACH,CAAC;YAED,MAAM,UAAU,GAAG,cAAI,CAAC,OAAO,CAC7B,cAAI,CAAC,IAAI,CACP,IAAI,CAAC,OAAO,EACZ,GAAG,cAAI,CAAC,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC,cAAc,IAAI,CAAC,GAAG,EAAE,MAAM,CAC9D,CACF,CAAC;YAEF,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;gBACrC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,cAAc,QAAQ,4BAA4B,CAAC,CAAC;gBAEtE,IAAA,uBAAM,EAAC,QAAQ,CAAC;qBACb,cAAc,CAAC,KAAK,CAAC,CAAC,sBAAsB;qBAC5C,aAAa,CAAC,CAAC,CAAC,CAAC,yCAAyC;qBAC1D,UAAU,CAAC,WAAW,CAAC,CAAC,mBAAmB;qBAC3C,MAAM,CAAC,KAAK,CAAC;qBACb,EAAE,CAAC,OAAO,EAAE,CAAC,WAAW,EAAE,EAAE;oBAC3B,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,2BAA2B,WAAW,EAAE,CAAC,CAAC;gBAC9D,CAAC,CAAC;qBACD,EAAE,CAAC,UAAU,EAAE,CAAC,QAAQ,EAAE,EAAE;oBAC3B,IAAI,QAAQ,CAAC,OAAO,EAAE,CAAC;wBACrB,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,8BAA8B,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAC9D,CAAC;oBACJ,CAAC;gBACH,CAAC,CAAC;qBACD,EAAE,CAAC,OAAO,EAAE,CAAC,KAAK,EAAE,EAAE;oBACrB,mBAAmB;oBACnB,KAAK,CAAC,IAAI,GAAG,qBAAqB,CAAC;oBACnC,MAAM,CAAC,KAAK,CAAC,CAAC;gBAChB,CAAC,CAAC;qBACD,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE;oBACd,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,+BAA+B,UAAU,EAAE,CAAC,CAAC;oBAC/D,OAAO,CAAC,UAAU,CAAC,CAAC;gBACtB,CAAC,CAAC;qBACD,IAAI,CAAC,UAAU,CAAC,CAAC;YACtB,CAAC,CAAC,CAAC;QACL,CAAC;KAAA;IAED;;OAEG;IACW,eAAe,CAC3B,SAAiB,EACjB,aAAkB;;YAElB,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YAE7B,IAAI,CAAC;gBACH,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,oCAAoC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,CAC7D,CAAC;gBAEF,MAAM,OAAO,GAAG;oBACd,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS;oBACjC,+BAA+B,EAAE,KAAK,EAAE,iCAAiC;oBACzE,qBAAqB,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS;oBAC7C,YAAY;oBACZ,wDAAwD;oBACxD,wDAAwD;oBACxD,qDAAqD;oBACrD,KAAK;oBACL,cAAc,EAAE;wBACd,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,uBAAuB;wBACxD,YAAY,EAAE,IAAI;wBAClB,YAAY,EAAE,KAAK,EAAE,uBAAuB;wBAC5C,WAAW,EAAE,KAAK,EAAE,iBAAiB;wBACrC,cAAc,EAAE,KAAK,EAAE,wBAAwB;wBAC/C,kBAAkB,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS;wBAC1C,WAAW,EAAE,KAAK;qBACnB;iBACF,CAAC;gBAEF,4BAA4B;gBAC5B,MAAM,OAAO,GAAG,MAAM,IAAA,4BAAW,EAAC,SAAS,EAAE,OAAO,CAAC,CAAC;gBAEtD,MAAM,cAAc,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC;gBAE9C,+BAA+B;gBAC/B,MAAM,qBAAqB,GAAG;oBAC5B,YAAY,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS;oBACpC,gBAAgB,EAAG,OAAe,CAAC,QAAQ,IAAI,SAAS;oBACxD,uBAAuB,EAAG,OAAe,CAAC,UAAU,IAAI,IAAI;oBAC5D,SAAS,EAAE,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM;oBACxE,aAAa,EAAE,OAAO,CAAC,cAAc,CAAC,cAAc;oBACpD,gBAAgB,EAAE,cAAc;iBACjC,CAAC;gBAEF,OAAO;oBACL,OAAO,EAAE,OAAO,CAAC,IAAI,EAAE;oBACvB,QAAQ,EAAE,qBAAqB;oBAC/B,cAAc,EAAE,cAAc;iBAC/B,CAAC;YACJ,CAAC;YAAC,OAAO,KAAU,EAAE,CAAC;gBACpB,KAAK,CAAC,IAAI,GAAG,eAAe,CAAC;gBAC7B,MAAM,KAAK,CAAC;YACd,CAAC;QACH,CAAC;KAAA;IAED;;;;;;;;;OASG;IACW,cAAc,CAC1B,QAAgB,EAChB,GAAW,EACX,KAAqD,EACrD,aAAkB;;;YAElB,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YAC7B,MAAM,OAAO,GAAG,GAAG,QAAQ,kBAAkB,CAAC;YAE9C,IAAI,MAAM,GAAG,KAAK,CAAC;YACnB,IAAI,MAAM,IAAI,CAAC,cAAc,CAAC,OAAO,EAAE,QAAQ,CAAC,EAAE,CAAC;gBACjD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,6BAA6B,OAAO,EAAE,CAAC,CAAC;gBAC1D,MAAM,GAAG,IAAI,CAAC;YAChB,CAAC;iBAAM,CAAC;gBACN,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;YACjD,CAAC;YAED,MAAM,GAAG,GAAG,MAAM,kBAAE,CAAC,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAChD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAgB,CAAC;YAC9C,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,KAAK,CAAC,gDAAgD,OAAO,EAAE,CAAC,CAAC;YAC7E,CAAC;YACD,MAAM,KAAK,GAAW,MAAM;iBACzB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,WAAC,OAAA,CAAC,EAAE,OAAO,EAAE,MAAM,CAAC,MAAA,CAAC,CAAC,OAAO,mCAAI,SAAS,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA,EAAA,CAAC;iBACrF,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACzB,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACvB,MAAM,IAAI,KAAK,CAAC,4CAA4C,OAAO,EAAE,CAAC,CAAC;YACzE,CAAC;YAED,MAAM,YAAY,GAAG,MAAA,IAAI,CAAC,OAAO,CAAC,YAAY,mCAAI,IAAI,CAAC;YACvD,MAAM,MAAM,GAAG,MAAM,IAAA,uBAAS,EAAC,KAAK,EAAE,QAAQ,EAAE,YAAY,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;YAC5E,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC;YAE3D,MAAM,QAAQ,iCACZ,IAAI,EAAE,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,EAC5B,WAAW,EAAE,IAAI,CAAC,mBAAmB,CAAC,GAAG,CAAC,EAC1C,QAAQ,EAAE,cAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EACjC,QAAQ,EACR,QAAQ,EAAE,KAAK,CAAC,IAAI,EACpB,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,WAAW,EAAE,EACxC,UAAU,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,EACrC,SAAS,EAAE,GAAG,IACX,aAAa,KAChB,SAAS,EAAE,MAAM,EACjB,SAAS,EAAE,MAAA,MAAA,IAAI,CAAC,OAAO,CAAC,IAAI,0CAAE,GAAG,mCAAI,MAAM,EAC3C,QAAQ,EAAE,MAAA,MAAA,IAAI,CAAC,OAAO,CAAC,IAAI,0CAAE,OAAO,mCAAI,IAAI,EAC5C,YAAY,EAAE,QAAQ,EACtB,KAAK,EAAE,KAAK,CAAC,MAAM,EACnB,OAAO,EACP,aAAa,EAAE,MAAM,EACrB,MAAM,EAAE,SAAS,EACjB,cAAc,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,GACvC,CAAC;YAEF,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,wBAAwB,QAAQ,KAAK,KAAK,CAAC,MAAM,WAAW,QAAQ,gBAAgB,MAAM,CAAC,MAAM,WAAW,CAC7G,CAAC;YACF,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;QAC9B,CAAC;KAAA;IAED,sEAAsE;IAC9D,YAAY,CAAC,CAAY;QAC/B,KAAK,MAAM,CAAC,IAAI,CAAC,WAAW,EAAE,UAAU,EAAE,SAAS,CAAU,EAAE,CAAC;YAC9D,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YACf,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,IAAI,EAAE;gBAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;QACzD,CAAC;QACD,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACvC,IAAI,CAAC,KAAK,OAAO,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,SAAS;gBAAE,SAAS;YAC9D,IAAI,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,IAAI,EAAE;gBAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;QACzD,CAAC;QACD,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,6EAA6E;IAC/D,cAAc,CAAC,OAAe,EAAE,QAAgB;;YAC5D,IAAI,CAAC;gBACH,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC,IAAA,eAAI,EAAC,OAAO,CAAC,EAAE,IAAA,eAAI,EAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAClE,OAAO,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,OAAO,CAAC;YAChC,CAAC;YAAC,WAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC;KAAA;IAED;;;;;OAKG;IACW,gBAAgB,CAAC,QAAgB,EAAE,OAAe;;YAC9D,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC;YAC/B,IAAI,CAAC,IAAI;gBAAE,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;YAE1D,MAAM,QAAQ,GAAG,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;YACxC,MAAM,MAAM,GAAG,cAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;YACrC,MAAM,OAAO,GAAG,CAAC,YAAY,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC;YAC7E,IAAI,CAAC,IAAI,CAAC,OAAO;gBAAE,OAAO,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;YAChD,IAAI,IAAI,CAAC,WAAW;gBAAE,OAAO,CAAC,IAAI,CAAC,gBAAgB,EAAE,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC;YAC/E,IAAI,IAAI,CAAC,MAAM;gBAAE,OAAO,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;YAEvD,IAAI,OAAe,CAAC;YACpB,IAAI,IAAc,CAAC;YACnB,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;gBACpB,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC;gBAC1B,IAAI,GAAG,CAAC,IAAI,EAAE,gBAAgB,EAAE,GAAG,OAAO,CAAC,CAAC;YAC9C,CAAC;iBAAM,CAAC;gBACN,OAAO,GAAG,IAAI,CAAC;gBACf,IAAI,GAAG,CAAC,KAAK,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,EAAE,QAAQ,EAAE,IAAI,EAAE,gBAAgB,EAAE,GAAG,OAAO,CAAC,CAAC;YAC7F,CAAC;YAED,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,aAAa,OAAO,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YAE3D,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;;gBAC1C,MAAM,KAAK,GAAG,IAAA,qBAAK,EAAC,OAAO,EAAE,IAAI,EAAE;oBACjC,GAAG,EAAE,cAAI,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC;oBAClC,KAAK,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC;oBACjC,GAAG,kCAAO,OAAO,CAAC,GAAG,KAAE,gBAAgB,EAAE,GAAG,GAAE;iBAC/C,CAAC,CAAC;gBACH,IAAI,MAAM,GAAG,EAAE,CAAC;gBAChB,MAAA,KAAK,CAAC,MAAM,0CAAE,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;gBAC9F,MAAA,KAAK,CAAC,MAAM,0CAAE,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE;oBAC7B,MAAM,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;oBACvB,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBAC/D,CAAC,CAAC,CAAC;gBACH,MAAM,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE;oBAC5B,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;oBACtB,MAAM,CAAC,IAAI,KAAK,CAAC,kCAAkC,IAAI,CAAC,SAAS,IAAI,CAAC,CAAC,CAAC;gBAC1E,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;gBACnB,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE;oBACxB,YAAY,CAAC,KAAK,CAAC,CAAC;oBACpB,MAAM,CAAC,IAAI,KAAK,CAAC,oCAAoC,OAAO,MAAM,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;gBACpF,CAAC,CAAC,CAAC;gBACH,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,IAAI,EAAE,EAAE;oBACzB,YAAY,CAAC,KAAK,CAAC,CAAC;oBACpB,IAAI,IAAI,KAAK,CAAC;wBAAE,OAAO,EAAE,CAAC;;wBACrB,MAAM,CAAC,IAAI,KAAK,CAAC,yBAAyB,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,KAAK,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;gBAC3G,CAAC,CAAC,CAAC;YACL,CAAC,CAAC,CAAC;QACL,CAAC;KAAA;IAED;;OAEG;IACW,OAAO,CACnB,aAAqB,EACrB,YAAoB;;YAEpB,IAAI,CAAC;gBACH,kDAAkD;gBAClD,IACE,aAAa,KAAK,YAAY;oBAC9B,aAAa,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,EACpC,CAAC;oBACD,MAAM,kBAAE,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;oBAC/B,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,8BAA8B,aAAa,EAAE,CAAC,CAAC;gBACnE,CAAC;YACH,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,qCAAqC,aAAa,KAAK,KAAK,EAAE,CAC/D,CAAC;YACJ,CAAC;QACH,CAAC;KAAA;IAED;;OAEG;IACW,aAAa;;YACzB,IAAI,CAAC;gBACH,MAAM,kBAAE,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;YACpD,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,mCAAmC,IAAI,CAAC,OAAO,KAAK,KAAK,EAAE,CAC5D,CAAC;YACJ,CAAC;QACH,CAAC;KAAA;IAED;;OAEG;IACK,YAAY,CAAC,GAAW;QAC9B,MAAM,SAAS,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;QACpE,MAAM,SAAS,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;QAEpD,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC;YAAE,OAAO,OAAO,CAAC;QAC5C,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC;YAAE,OAAO,OAAO,CAAC;QAC5C,OAAO,OAAO,CAAC;IACjB,CAAC;IAED;;OAEG;IACK,mBAAmB,CAAC,GAAW;QACrC,MAAM,YAAY,GAA8B;YAC9C,MAAM,EAAE,gBAAgB;YACxB,MAAM,EAAE,gBAAgB;YACxB,MAAM,EAAE,gBAAgB;YACxB,MAAM,EAAE,gBAAgB;YACxB,MAAM,EAAE,gBAAgB;YACxB,OAAO,EAAE,iBAAiB;YAC1B,MAAM,EAAE,gBAAgB;YACxB,OAAO,EAAE,iBAAiB;YAC1B,MAAM,EAAE,qBAAqB;YAC7B,MAAM,EAAE,gBAAgB;SACzB,CAAC;QACF,OAAO,YAAY,CAAC,GAAG,CAAC,IAAI,YAAY,CAAC;IAC3C,CAAC;CACF;AAvgBD,kCAugBC"}

package/dist/core/processor/readers/BinaryReader.js

@@ -0,0 +1,90 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.BinaryReader = void 0;
+const path = __importStar(require("path"));
+const FileReader_1 = require("./FileReader");
+/**
+ * Final-fallback reader: the catch-all for binary / unknown files.
+ *
+ * Registered LAST in the factory, it claims anything no specific reader
+ * recognized (overriding `canRead` to always match) and produces *no* chunks —
+ * so binaries are skipped gracefully instead of being read as UTF-8 mojibake and
+ * shipped to the LLM. The carried extension list is documentation of the obvious
+ * binary types; routing relies on the catch-all `canRead`, not the list.
+ */
+class BinaryReader extends FileReader_1.FileReader {
+    constructor(chunker, logger) {
+        super([
+            ".bin", ".dat", ".exe", ".dll", ".so", ".dylib",
+            ".o", ".a", ".lib", ".class", ".jar", ".wasm",
+        ], chunker, logger);
+    }
+    getName() {
+        return "BinaryReader";
+    }
+    /** Catch-all: claims any file that fell through every specific reader. */
+    canRead(_filePath) {
+        return true;
+    }
+    read(filePath) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const fileName = path.basename(filePath);
+            // No bytes read, no LLM call — emit zero chunks and flag a graceful skip the
+            // pipeline honors before its "no content extracted" guard fires.
+            this.logger.info(`Skipping binary / unsupported file: ${fileName}`);
+            return {
+                chunks: [],
+                metadata: {
+                    type: "binary",
+                    skip: true,
+                    fileName,
+                    extension: path.extname(filePath).toLowerCase(),
+                },
+            };
+        });
+    }
+}
+exports.BinaryReader = BinaryReader;
+//# sourceMappingURL=BinaryReader.js.map

package/dist/core/processor/readers/BinaryReader.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"BinaryReader.js","sourceRoot":"","sources":["../../../../src/core/processor/readers/BinaryReader.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,2CAA6B;AAC7B,6CAA0D;AAI1D;;;;;;;;GAQG;AACH,MAAa,YAAa,SAAQ,uBAAU;IAC1C,YAAY,OAAoB,EAAE,MAAc;QAC9C,KAAK,CACH;YACE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ;YAC/C,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,OAAO;SAC9C,EACD,OAAO,EACP,MAAM,CACP,CAAC;IACJ,CAAC;IAED,OAAO;QACL,OAAO,cAAc,CAAC;IACxB,CAAC;IAED,0EAA0E;IAC1E,OAAO,CAAC,SAAiB;QACvB,OAAO,IAAI,CAAC;IACd,CAAC;IAEK,IAAI,CAAC,QAAgB;;YACzB,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACzC,6EAA6E;YAC7E,iEAAiE;YACjE,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,uCAAuC,QAAQ,EAAE,CAAC,CAAC;YACpE,OAAO;gBACL,MAAM,EAAE,EAAE;gBACV,QAAQ,EAAE;oBACR,IAAI,EAAE,QAAQ;oBACd,IAAI,EAAE,IAAI;oBACV,QAAQ;oBACR,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;iBAChD;aACF,CAAC;QACJ,CAAC;KAAA;CACF;AApCD,oCAoCC"}