@zvec/zvec-grep 0.1.4 → 0.1.6

package/dist/engine/extraction/index.js

@@ -1,281 +1,2 @@
-import { EngineError } from "../errors/index.js";
-import { makeEntityId } from "./ids.js";
-import { CodeExtractor } from "./code/extractor.js";
-import { MarkdownExtractor } from "./markdown/extractor.js";
-const DEFAULT_TEXT_CHUNK_CHARS = 3600;
-const DEFAULT_TEXT_CHUNK_OVERLAP_CHARS = 540;
-export class TextExtractor {
-    maxChunkChars;
-    chunkOverlapChars;
-    constructor(options = {}) {
-        const maxChunkChars = options.maxChunkChars ?? DEFAULT_TEXT_CHUNK_CHARS;
-        const chunkOverlapChars = options.chunkOverlapChars ?? DEFAULT_TEXT_CHUNK_OVERLAP_CHARS;
-        if (!Number.isInteger(maxChunkChars) || maxChunkChars <= 0) {
-            throw new EngineError("Text extractor requires a positive integer chunk size", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.TEXT_INVALID_CHUNK_SIZE",
-                context: `maxChunkChars=${maxChunkChars}`,
-            });
-        }
-        if (!Number.isInteger(chunkOverlapChars)
-            || chunkOverlapChars < 0
-            || chunkOverlapChars >= maxChunkChars) {
-            throw new EngineError("Text extractor requires overlap to be smaller than chunk size", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.TEXT_INVALID_CHUNK_OVERLAP",
-                context: `maxChunkChars=${maxChunkChars} chunkOverlapChars=${chunkOverlapChars}`,
-            });
-        }
-        this.maxChunkChars = maxChunkChars;
-        this.chunkOverlapChars = chunkOverlapChars;
-    }
-    supports(source) {
-        return source.kind === "text";
-    }
-    async extract(source) {
-        if (source.kind !== "text") {
-            throw new EngineError("Text extractor received a non-text source", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.TEXT_UNSUPPORTED_SOURCE",
-                context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-            });
-        }
-        validateSourceFile(source);
-        const chunks = chunkText(source.text, this.maxChunkChars, this.chunkOverlapChars);
-        return chunks.map((chunk, index) => ({
-            id: makeEntityId(source.file.id, index),
-            fileId: source.file.id,
-            range: chunk.range,
-            content: {
-                kind: "text",
-                text: chunk.text,
-            },
-        }));
-    }
-}
-export class ImageExtractor {
-    supports(source) {
-        return source.kind === "image";
-    }
-    async extract(source) {
-        if (source.kind !== "image") {
-            throw new EngineError("Image extractor received a non-image source", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.IMAGE_UNSUPPORTED_SOURCE",
-                context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-            });
-        }
-        validateSourceFile(source);
-        if (source.data.byteLength === 0) {
-            throw new EngineError("Image extractor requires non-empty image data", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.IMAGE_EMPTY_DATA",
-                context: `fileId=${source.file.id} format=${source.format}`,
-            });
-        }
-        return [
-            {
-                id: makeEntityId(source.file.id, 0),
-                fileId: source.file.id,
-                range: {
-                    kind: "file",
-                },
-                content: {
-                    kind: "image",
-                    data: source.data,
-                    format: source.format,
-                },
-            },
-        ];
-    }
-}
-const DEFAULT_EXTRACTORS = [
-    new CodeExtractor(),
-    new MarkdownExtractor(),
-    new TextExtractor(),
-    new ImageExtractor(),
-];
-export class ExtractorRegistry {
-    extractors;
-    constructor(extractors = DEFAULT_EXTRACTORS) {
-        if (extractors.length === 0) {
-            throw new EngineError("Extractor registry requires at least one extractor", {
-                code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_REGISTRY",
-            });
-        }
-        this.extractors = extractors;
-    }
-    resolve(source) {
-        for (const extractor of this.extractors) {
-            if (extractor.supports(source)) {
-                return extractor;
-            }
-        }
-        throw new EngineError("No extractor supports source", {
-            code: "ZVEC_GREP.ENGINE.EXTRACTORS.UNSUPPORTED_SOURCE",
-            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-        });
-    }
-    extract(source) {
-        return this.extractFirst(source);
-    }
-    async extractFirst(source) {
-        let matched = false;
-        let lastResult = [];
-        for (const extractor of this.extractors) {
-            if (!extractor.supports(source)) {
-                continue;
-            }
-            matched = true;
-            lastResult = await extractor.extract(source);
-            if (lastResult.length > 0) {
-                return lastResult;
-            }
-        }
-        if (matched) {
-            return lastResult;
-        }
-        throw new EngineError("No extractor supports source", {
-            code: "ZVEC_GREP.ENGINE.EXTRACTORS.UNSUPPORTED_SOURCE",
-            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-        });
-    }
-}
-export function extractFragments(source) {
-    return new ExtractorRegistry().extract(source);
-}
-function chunkText(text, maxChars, overlapChars) {
-    if (text.trim().length === 0) {
-        return [];
-    }
-    const lines = text.split("\n");
-    const lineOffsets = computeLineOffsets(lines);
-    const chunks = [];
-    let startIndex = 0;
-    while (startIndex < lines.length) {
-        if (lines[startIndex].length + 1 > maxChars) {
-            const line = lines[startIndex];
-            const lineNumber = startIndex + 1;
-            let offset = 0;
-            while (offset < line.length) {
-                const remaining = line.length - offset;
-                const sliceLength = remaining <= maxChars
-                    ? remaining
-                    : findLineCut(line.slice(offset), maxChars);
-                const slice = line.slice(offset, offset + sliceLength);
-                if (slice.trim().length > 0) {
-                    chunks.push({
-                        text: slice,
-                        range: {
-                            kind: "text",
-                            startLine: lineNumber,
-                            endLine: lineNumber,
-                            startOffset: lineOffsets[startIndex] + offset,
-                            endOffset: lineOffsets[startIndex] + offset + slice.length,
-                        },
-                    });
-                }
-                offset += sliceLength;
-            }
-            startIndex++;
-            continue;
-        }
-        let usedChars = 0;
-        let endIndex = startIndex;
-        while (endIndex < lines.length) {
-            const lineLength = lines[endIndex].length + 1;
-            if (usedChars + lineLength > maxChars && endIndex > startIndex) {
-                break;
-            }
-            usedChars += lineLength;
-            endIndex++;
-        }
-        const chunkLines = lines.slice(startIndex, endIndex);
-        const chunk = chunkLines.join("\n");
-        if (chunk.trim().length > 0) {
-            const endLineIndex = endIndex - 1;
-            chunks.push({
-                text: chunk,
-                range: {
-                    kind: "text",
-                    startLine: startIndex + 1,
-                    endLine: endIndex,
-                    startOffset: lineOffsets[startIndex],
-                    endOffset: lineOffsets[endLineIndex] + lines[endLineIndex].length,
-                },
-            });
-        }
-        if (endIndex >= lines.length) {
-            break;
-        }
-        startIndex = computeNextStartLine(lines, startIndex, endIndex, overlapChars);
-    }
-    return chunks;
-}
-function computeLineOffsets(lines) {
-    const offsets = [];
-    let offset = 0;
-    for (const line of lines) {
-        offsets.push(offset);
-        offset += line.length + 1;
-    }
-    return offsets;
-}
-function computeNextStartLine(lines, startIndex, endIndex, overlapChars) {
-    if (overlapChars === 0) {
-        return endIndex;
-    }
-    let overlapLines = 0;
-    let overlapCount = 0;
-    for (let index = endIndex - 1; index >= startIndex && overlapCount < overlapChars; index--) {
-        overlapCount += lines[index].length + 1;
-        overlapLines++;
-    }
-    const nextStart = endIndex - overlapLines;
-    return nextStart > startIndex ? nextStart : endIndex;
-}
-function findLineCut(line, maxChars) {
-    if (line.length <= maxChars) {
-        return line.length;
-    }
-    const minPosition = Math.floor(maxChars * 0.7);
-    let bestPosition = -1;
-    let bestScore = 0;
-    for (let index = minPosition; index < maxChars; index++) {
-        const character = line[index];
-        let score = 0;
-        if (character === "." || character === "!" || character === "?") {
-            score = 4;
-        }
-        else if (character === "," || character === ";" || character === ":") {
-            score = 3;
-        }
-        else if (character === " " || character === "\t") {
-            score = 2;
-        }
-        else if (character === "-" || character === "/" || character === "\\") {
-            score = 1;
-        }
-        if (score > 0 && score >= bestScore) {
-            bestScore = score;
-            bestPosition = index + 1;
-        }
-    }
-    return bestPosition > 0 ? bestPosition : maxChars;
-}
-function validateSourceFile(source) {
-    if (source.file.id.trim().length === 0) {
-        throw new EngineError("Extractor source requires a non-empty file id", {
-            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_FILE_ID",
-            context: `sourceKind=${source.kind}`,
-        });
-    }
-    if (source.file.absolutePath.trim().length === 0) {
-        throw new EngineError("Extractor source requires a non-empty absolute file path", {
-            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_ABSOLUTE_FILE_PATH",
-            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-        });
-    }
-    if (source.file.relativePath.trim().length === 0) {
-        throw new EngineError("Extractor source requires a non-empty relative file path", {
-            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_RELATIVE_FILE_PATH",
-            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
-        });
-    }
-}
+export { extract } from "./runtime.js";
 //# sourceMappingURL=index.js.map

package/dist/engine/extraction/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/engine/extraction/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AAOjD,OAAO,EAAE,YAAY,EAAE,MAAM,UAAU,CAAC;AACxC,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AA4C5D,MAAM,wBAAwB,GAAG,IAAI,CAAC;AAGtC,MAAM,gCAAgC,GAAG,GAAG,CAAC;AAG7C,MAAM,OAAO,aAAa;IACP,aAAa,CAAS;IACtB,iBAAiB,CAAS;IAG3C,YAAY,UAAgC,EAAE;QAC5C,MAAM,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,wBAAwB,CAAC;QACxE,MAAM,iBAAiB,GAAG,OAAO,CAAC,iBAAiB,IAAI,gCAAgC,CAAC;QAExF,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,aAAa,CAAC,IAAI,aAAa,IAAI,CAAC,EAAE,CAAC;YAC3D,MAAM,IAAI,WAAW,CAAC,uDAAuD,EAAE;gBAC7E,IAAI,EAAE,qDAAqD;gBAC3D,OAAO,EAAE,iBAAiB,aAAa,EAAE;aAC1C,CAAC,CAAC;QACL,CAAC;QAED,IACE,CAAC,MAAM,CAAC,SAAS,CAAC,iBAAiB,CAAC;eACjC,iBAAiB,GAAG,CAAC;eACrB,iBAAiB,IAAI,aAAa,EACrC,CAAC;YACD,MAAM,IAAI,WAAW,CAAC,+DAA+D,EAAE;gBACrF,IAAI,EAAE,wDAAwD;gBAC9D,OAAO,EAAE,iBAAiB,aAAa,sBAAsB,iBAAiB,EAAE;aACjF,CAAC,CAAC;QACL,CAAC;QAED,IAAI,CAAC,aAAa,GAAG,aAAa,CAAC;QACnC,IAAI,CAAC,iBAAiB,GAAG,iBAAiB,CAAC;IAC7C,CAAC;IAGD,QAAQ,CAAC,MAAc;QACrB,OAAO,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC;IAChC,CAAC;IAGD,KAAK,CAAC,OAAO,CAAC,MAAc;QAC1B,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,MAAM,IAAI,WAAW,CAAC,2CAA2C,EAAE;gBACjE,IAAI,EAAE,qDAAqD;gBAC3D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;aAC9D,CAAC,CAAC;QACL,CAAC;QAED,kBAAkB,CAAC,MAAM,CAAC,CAAC;QAE3B,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,aAAa,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC;QAElF,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAkB,EAAE,CAAC,CAAC;YACnD,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC;YACvC,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;YACtB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,OAAO,EAAE;gBACP,IAAI,EAAE,MAAM;gBACZ,IAAI,EAAE,KAAK,CAAC,IAAI;aACjB;SACF,CAAC,CAAC,CAAC;IACN,CAAC;CACF;AAGD,MAAM,OAAO,cAAc;IACzB,QAAQ,CAAC,MAAc;QACrB,OAAO,MAAM,CAAC,IAAI,KAAK,OAAO,CAAC;IACjC,CAAC;IAGD,KAAK,CAAC,OAAO,CAAC,MAAc;QAC1B,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,MAAM,IAAI,WAAW,CAAC,6CAA6C,EAAE;gBACnE,IAAI,EAAE,sDAAsD;gBAC5D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;aAC9D,CAAC,CAAC;QACL,CAAC;QAED,kBAAkB,CAAC,MAAM,CAAC,CAAC;QAE3B,IAAI,MAAM,CAAC,IAAI,CAAC,UAAU,KAAK,CAAC,EAAE,CAAC;YACjC,MAAM,IAAI,WAAW,CAAC,+CAA+C,EAAE;gBACrE,IAAI,EAAE,8CAA8C;gBACpD,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,WAAW,MAAM,CAAC,MAAM,EAAE;aAC5D,CAAC,CAAC;QACL,CAAC;QAED,OAAO;YACL;gBACE,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC;gBACnC,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;gBACtB,KAAK,EAAE;oBACL,IAAI,EAAE,MAAM;iBACb;gBACD,OAAO,EAAE;oBACP,IAAI,EAAE,OAAO;oBACb,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,MAAM,EAAE,MAAM,CAAC,MAAM;iBACtB;aACF;SACF,CAAC;IACJ,CAAC;CACF;AAGD,MAAM,kBAAkB,GAAyB;IAC/C,IAAI,aAAa,EAAE;IACnB,IAAI,iBAAiB,EAAE;IACvB,IAAI,aAAa,EAAE;IACnB,IAAI,cAAc,EAAE;CACrB,CAAC;AAGF,MAAM,OAAO,iBAAiB;IACX,UAAU,CAAuB;IAGlD,YAAY,aAAmC,kBAAkB;QAC/D,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC5B,MAAM,IAAI,WAAW,CAAC,oDAAoD,EAAE;gBAC1E,IAAI,EAAE,4CAA4C;aACnD,CAAC,CAAC;QACL,CAAC;QAED,IAAI,CAAC,UAAU,GAAG,UAAU,CAAC;IAC/B,CAAC;IAGD,OAAO,CAAC,MAAc;QACpB,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACxC,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC/B,OAAO,SAAS,CAAC;YACnB,CAAC;QACH,CAAC;QAED,MAAM,IAAI,WAAW,CAAC,8BAA8B,EAAE;YACpD,IAAI,EAAE,gDAAgD;YACtD,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CAAC,CAAC;IACL,CAAC;IAGD,OAAO,CAAC,MAAc;QACpB,OAAO,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;IACnC,CAAC;IAGO,KAAK,CAAC,YAAY,CAAC,MAAc;QACvC,IAAI,OAAO,GAAG,KAAK,CAAC;QACpB,IAAI,UAAU,GAAqB,EAAE,CAAC;QAEtC,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACxC,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gBAChC,SAAS;YACX,CAAC;YAED,OAAO,GAAG,IAAI,CAAC;YACf,UAAU,GAAG,MAAM,SAAS,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;YAE7C,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC1B,OAAO,UAAU,CAAC;YACpB,CAAC;QACH,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,UAAU,CAAC;QACpB,CAAC;QAED,MAAM,IAAI,WAAW,CAAC,8BAA8B,EAAE;YACpD,IAAI,EAAE,gDAAgD;YACtD,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CAAC,CAAC;IACL,CAAC;CACF;AAGD,MAAM,UAAU,gBAAgB,CAAC,MAAc;IAC7C,OAAO,IAAI,iBAAiB,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;AACjD,CAAC;AASD,SAAS,SAAS,CAAC,IAAY,EAAE,QAAgB,EAAE,YAAoB;IACrE,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC7B,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC/B,MAAM,WAAW,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAC9C,MAAM,MAAM,GAAgB,EAAE,CAAC;IAE/B,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,OAAO,UAAU,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QACjC,IAAI,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,GAAG,CAAC,GAAG,QAAQ,EAAE,CAAC;YAC5C,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC;YAC/B,MAAM,UAAU,GAAG,UAAU,GAAG,CAAC,CAAC;YAClC,IAAI,MAAM,GAAG,CAAC,CAAC;YAEf,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC5B,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;gBACvC,MAAM,WAAW,GAAG,SAAS,IAAI,QAAQ;oBACvC,CAAC,CAAC,SAAS;oBACX,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,QAAQ,CAAC,CAAC;gBAC9C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,WAAW,CAAC,CAAC;gBAEvD,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC5B,MAAM,CAAC,IAAI,CAAC;wBACV,IAAI,EAAE,KAAK;wBACX,KAAK,EAAE;4BACL,IAAI,EAAE,MAAM;4BACZ,SAAS,EAAE,UAAU;4BACrB,OAAO,EAAE,UAAU;4BACnB,WAAW,EAAE,WAAW,CAAC,UAAU,CAAC,GAAG,MAAM;4BAC7C,SAAS,EAAE,WAAW,CAAC,UAAU,CAAC,GAAG,MAAM,GAAG,KAAK,CAAC,MAAM;yBAC3D;qBACF,CAAC,CAAC;gBACL,CAAC;gBAED,MAAM,IAAI,WAAW,CAAC;YACxB,CAAC;YAED,UAAU,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QAED,IAAI,SAAS,GAAG,CAAC,CAAC;QAClB,IAAI,QAAQ,GAAG,UAAU,CAAC;QAE1B,OAAO,QAAQ,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YAC/B,MAAM,UAAU,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;YAE9C,IAAI,SAAS,GAAG,UAAU,GAAG,QAAQ,IAAI,QAAQ,GAAG,UAAU,EAAE,CAAC;gBAC/D,MAAM;YACR,CAAC;YAED,SAAS,IAAI,UAAU,CAAC;YACxB,QAAQ,EAAE,CAAC;QACb,CAAC;QAED,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;QACrD,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAEpC,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5B,MAAM,YAAY,GAAG,QAAQ,GAAG,CAAC,CAAC;YAElC,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,KAAK;gBACX,KAAK,EAAE;oBACL,IAAI,EAAE,MAAM;oBACZ,SAAS,EAAE,UAAU,GAAG,CAAC;oBACzB,OAAO,EAAE,QAAQ;oBACjB,WAAW,EAAE,WAAW,CAAC,UAAU,CAAC;oBACpC,SAAS,EAAE,WAAW,CAAC,YAAY,CAAC,GAAG,KAAK,CAAC,YAAY,CAAC,CAAC,MAAM;iBAClE;aACF,CAAC,CAAC;QACL,CAAC;QAED,IAAI,QAAQ,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YAC7B,MAAM;QACR,CAAC;QAED,UAAU,GAAG,oBAAoB,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,EAAE,YAAY,CAAC,CAAC;IAC/E,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAGD,SAAS,kBAAkB,CAAC,KAAwB;IAClD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrB,MAAM,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAGD,SAAS,oBAAoB,CAC3B,KAAwB,EACxB,UAAkB,EAClB,QAAgB,EAChB,YAAoB;IAEpB,IAAI,YAAY,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,IAAI,YAAY,GAAG,CAAC,CAAC;IAErB,KAAK,IAAI,KAAK,GAAG,QAAQ,GAAG,CAAC,EAAE,KAAK,IAAI,UAAU,IAAI,YAAY,GAAG,YAAY,EAAE,KAAK,EAAE,EAAE,CAAC;QAC3F,YAAY,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;QACxC,YAAY,EAAE,CAAC;IACjB,CAAC;IAED,MAAM,SAAS,GAAG,QAAQ,GAAG,YAAY,CAAC;IAE1C,OAAO,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;AACvD,CAAC;AAGD,SAAS,WAAW,CAAC,IAAY,EAAE,QAAgB;IACjD,IAAI,IAAI,CAAC,MAAM,IAAI,QAAQ,EAAE,CAAC;QAC5B,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAED,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC;IAC/C,IAAI,YAAY,GAAG,CAAC,CAAC,CAAC;IACtB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,IAAI,KAAK,GAAG,WAAW,EAAE,KAAK,GAAG,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC;QACxD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YAChE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YACvE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACnD,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACxE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;QAED,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,IAAI,SAAS,EAAE,CAAC;YACpC,SAAS,GAAG,KAAK,CAAC;YAClB,YAAY,GAAG,KAAK,GAAG,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,OAAO,YAAY,GAAG,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC;AACpD,CAAC;AAGD,SAAS,kBAAkB,CAAC,MAAc;IACxC,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvC,MAAM,IAAI,WAAW,CAAC,+CAA+C,EAAE;YACrE,IAAI,EAAE,2CAA2C;YACjD,OAAO,EAAE,cAAc,MAAM,CAAC,IAAI,EAAE;SACrC,CAAC,CAAC;IACL,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,WAAW,CAAC,0DAA0D,EAAE;YAChF,IAAI,EAAE,sDAAsD;YAC5D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CAAC,CAAC;IACL,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,WAAW,CAAC,0DAA0D,EAAE;YAChF,IAAI,EAAE,sDAAsD;YAC5D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CAAC,CAAC;IACL,CAAC;AACH,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/engine/extraction/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC"}

package/dist/engine/extraction/markdown/extractor.d.ts

@@ -1,13 +1,7 @@
 import type { EntityFragment } from "../../types.js";
-import type { Extractor, Source } from "../index.js";
-export type MarkdownExtractorOptions = {
-    maxChunkChars?: number;
-    chunkOverlapChars?: number;
-};
-export declare class MarkdownExtractor implements Extractor {
-    private readonly maxChunkChars;
-    private readonly chunkOverlapChars;
-    constructor(options?: MarkdownExtractorOptions);
-    supports(source: Source): boolean;
-    extract(source: Source): Promise<EntityFragment[]>;
+import { type Source } from "../source.js";
+import type { ChunkOptions } from "../types.js";
+export declare class MarkdownExtractor {
+    extract(source: Source, options?: ChunkOptions): Promise<EntityFragment[]>;
+    private fallback;
 }

package/dist/engine/extraction/markdown/extractor.js

@@ -1,24 +1,20 @@
+import { EngineError } from "../../errors.js";
 import { makeEntityId } from "../ids.js";
+import { validateSourceFile } from "../source.js";
+import { extractPlainTextFragments } from "../text/extractor.js";
 const DEFAULT_MARKDOWN_CHUNK_CHARS = 3600;
 const DEFAULT_MARKDOWN_CHUNK_OVERLAP_CHARS = 540;
 export class MarkdownExtractor {
-    maxChunkChars;
-    chunkOverlapChars;
-    constructor(options = {}) {
-        this.maxChunkChars = options.maxChunkChars ?? DEFAULT_MARKDOWN_CHUNK_CHARS;
-        this.chunkOverlapChars = options.chunkOverlapChars ?? DEFAULT_MARKDOWN_CHUNK_OVERLAP_CHARS;
-    }
-    supports(source) {
-        return source.kind === "text" && source.file.format === "markdown";
-    }
-    async extract(source) {
-        if (!this.supports(source) || source.kind !== "text") {
+    async extract(source, options = {}) {
+        if (source.kind !== "text" || source.file.format !== "markdown") {
             return [];
         }
+        validateSourceFile(source);
+        const chunkOptions = resolveMarkdownChunkOptions(options);
         const lines = source.text.split("\n");
         const headings = scanHeadings(lines);
         if (headings.length === 0) {
-            return [];
+            return this.fallback(source, chunkOptions);
         }
         const lineOffsets = computeLineOffsets(lines);
         const fenceLines = computeFenceLines(lines);
@@ -31,8 +27,8 @@ export class MarkdownExtractor {
                 lineOffsets,
                 fenceLines,
                 section,
-                maxChars: this.maxChunkChars,
-                overlapChars: this.chunkOverlapChars,
+                maxChars: chunkOptions.maxChunkChars,
+                overlapChars: chunkOptions.chunkOverlapChars,
             });
             if (windows.length > 1) {
                 const id = makeEntityId(source.file.id, fragments.length);
@@ -65,8 +61,32 @@ export class MarkdownExtractor {
                 });
             }
         }
-        return fragments;
+        return fragments.length > 0
+            ? fragments
+            : this.fallback(source, chunkOptions);
+    }
+    fallback(source, options) {
+        return extractPlainTextFragments(source, options.maxChunkChars, options.chunkOverlapChars);
+    }
+}
+function resolveMarkdownChunkOptions(options) {
+    const maxChunkChars = options.maxChunkChars ?? DEFAULT_MARKDOWN_CHUNK_CHARS;
+    const chunkOverlapChars = options.chunkOverlapChars ?? DEFAULT_MARKDOWN_CHUNK_OVERLAP_CHARS;
+    if (!Number.isInteger(maxChunkChars) || maxChunkChars <= 0) {
+        throw new EngineError("Markdown extractor requires a positive integer chunk size", {
+            code: "ZVEC_GREP.ENGINE.EXTRACTORS.MARKDOWN_INVALID_CHUNK_SIZE",
+            context: `maxChunkChars=${maxChunkChars}`,
+        });
+    }
+    if (!Number.isInteger(chunkOverlapChars) ||
+        chunkOverlapChars < 0 ||
+        chunkOverlapChars >= maxChunkChars) {
+        throw new EngineError("Markdown extractor requires overlap to be smaller than chunk size", {
+            code: "ZVEC_GREP.ENGINE.EXTRACTORS.MARKDOWN_INVALID_CHUNK_OVERLAP",
+            context: `maxChunkChars=${maxChunkChars} chunkOverlapChars=${chunkOverlapChars}`,
+        });
     }
+    return { maxChunkChars, chunkOverlapChars };
 }
 function markdownWindowToFragment(source, metadata, window, index, group) {
     return {
@@ -129,7 +149,8 @@ function buildSections(headings, lines) {
     const stack = [];
     const sections = [];
     const firstHeading = headings[0];
-    if (firstHeading.lineIndex > 0 && lines.slice(0, firstHeading.lineIndex).join("\n").trim().length > 0) {
+    if (firstHeading.lineIndex > 0 &&
+        lines.slice(0, firstHeading.lineIndex).join("\n").trim().length > 0) {
         sections.push({
             heading: null,
             startIndex: 0,

package/dist/engine/extraction/markdown/extractor.js.map

@@ -1 +1 @@
-{"version":3,"file":"extractor.js","sourceRoot":"","sources":["../../../../src/engine/extraction/markdown/extractor.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AAUzC,MAAM,4BAA4B,GAAG,IAAI,CAAC;AAC1C,MAAM,oCAAoC,GAAG,GAAG,CAAC;AAwBjD,MAAM,OAAO,iBAAiB;IACX,aAAa,CAAS;IACtB,iBAAiB,CAAS;IAG3C,YAAY,UAAoC,EAAE;QAChD,IAAI,CAAC,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,4BAA4B,CAAC;QAC3E,IAAI,CAAC,iBAAiB,GAAG,OAAO,CAAC,iBAAiB,IAAI,oCAAoC,CAAC;IAC7F,CAAC;IAGD,QAAQ,CAAC,MAAc;QACrB,OAAO,MAAM,CAAC,IAAI,KAAK,MAAM,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,KAAK,UAAU,CAAC;IACrE,CAAC;IAGD,KAAK,CAAC,OAAO,CAAC,MAAc;QAC1B,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YACrD,OAAO,EAAE,CAAC;QACZ,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;QACrC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC1B,OAAO,EAAE,CAAC;QACZ,CAAC;QAED,MAAM,WAAW,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;QAC9C,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAC;QAC5C,MAAM,QAAQ,GAAG,aAAa,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;QAChD,MAAM,SAAS,GAAqB,EAAE,CAAC;QAEvC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,MAAM,QAAQ,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;YAC3C,MAAM,OAAO,GAAG,oBAAoB,CAAC;gBACnC,KAAK;gBACL,WAAW;gBACX,UAAU;gBACV,OAAO;gBACP,QAAQ,EAAE,IAAI,CAAC,aAAa;gBAC5B,YAAY,EAAE,IAAI,CAAC,iBAAiB;aACrC,CAAC,CAAC;YAEH,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACvB,MAAM,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC;gBAC1D,SAAS,CAAC,IAAI,CAAC;oBACb,EAAE;oBACF,KAAK,EAAE,EAAE;oBACT,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;oBACtB,KAAK,EAAE,aAAa,CAAC,KAAK,EAAE,WAAW,EAAE,OAAO,CAAC,UAAU,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC,KAAK;oBACpF,OAAO,EAAE;wBACP,IAAI,EAAE,MAAM;wBACZ,IAAI,EAAE,eAAe,CAAC,QAAQ,CAAC;qBAChC;oBACD,QAAQ;iBACT,CAAC,CAAC;gBAEH,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;oBAC7B,SAAS,CAAC,IAAI,CAAC,wBAAwB,CACrC,MAAM,EACN,QAAQ,EACR,MAAM,EACN,SAAS,CAAC,MAAM,EAChB,EAAE,CACH,CAAC,CAAC;gBACL,CAAC;gBAED,SAAS;YACX,CAAC;YAED,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;gBAC7B,SAAS,CAAC,IAAI,CAAC;oBACb,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,CAAC,MAAM,CAAC;oBAClD,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;oBACtB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,OAAO,EAAE;wBACP,IAAI,EAAE,MAAM;wBACZ,IAAI,EAAE,MAAM,CAAC,IAAI;qBAClB;oBACD,QAAQ;iBACT,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO,SAAS,CAAC;IACnB,CAAC;CACF;AAGD,SAAS,wBAAwB,CAC/B,MAAkB,EAClB,QAAgC,EAChC,MAAsB,EACtB,KAAa,EACb,KAAa;IAEb,OAAO;QACL,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC;QACvC,KAAK;QACL,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;QACtB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,OAAO,EAAE;YACP,IAAI,EAAE,MAAM;YACZ,IAAI,EAAE,MAAM,CAAC,IAAI;SAClB;QACD,QAAQ;KACT,CAAC;AACJ,CAAC;AAGD,SAAS,eAAe,CAAC,QAAgC;IACvD,OAAO,QAAQ,CAAC,OAAO,IAAI,kBAAkB,CAAC;AAChD,CAAC;AAGD,SAAS,YAAY,CAAC,KAAwB;IAC5C,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,KAAK,GAAyB,IAAI,CAAC;IAEvC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QAClD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;QAC1B,MAAM,OAAO,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC;QAEjC,IAAI,KAAK,EAAE,CAAC;YACV,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC9B,KAAK,GAAG,IAAI,CAAC;YACf,CAAC;YACD,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;YACd,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;YACd,SAAS;QACX,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,4BAA4B,CAAC,CAAC;QACrD,IAAI,GAAG,EAAE,CAAC;YACR,QAAQ,CAAC,IAAI,CAAC;gBACZ,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM;gBACpB,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE;gBACnB,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAChE,QAAQ,CAAC,IAAI,CAAC;gBACZ,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnC,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;gBACjB,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAGD,SAAS,aAAa,CAAC,QAA4B,EAAE,KAAwB;IAC3E,MAAM,KAAK,GAAc,EAAE,CAAC;IAC5B,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAEjC,IAAI,YAAY,CAAC,SAAS,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtG,QAAQ,CAAC,IAAI,CAAC;YACZ,OAAO,EAAE,IAAI;YACb,UAAU,EAAE,CAAC;YACb,QAAQ,EAAE,YAAY,CAAC,SAAS,GAAG,CAAC;YACpC,UAAU,EAAE,EAAE;SACf,CAAC,CAAC;IACL,CAAC;IAED,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QACrD,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAEhC,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;YAC1E,KAAK,CAAC,GAAG,EAAE,CAAC;QACd,CAAC;QAED,QAAQ,CAAC,IAAI,CAAC;YACZ,OAAO;YACP,UAAU,EAAE,OAAO,CAAC,SAAS;YAC7B,QAAQ,EAAE,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC,MAAM;gBACnC,CAAC,CAAC,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC;gBACnC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC;YACpB,UAAU,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SAC3C,CAAC,CAAC;QACH,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACtB,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAGD,SAAS,oBAAoB,CAAC,KAO7B;IACC,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,UAAU,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC;IAE1C,OAAO,UAAU,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;QAC5C,IAAI,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,GAAG,CAAC,GAAG,KAAK,CAAC,QAAQ,EAAE,CAAC;YACxD,OAAO,CAAC,IAAI,CAAC,GAAG,aAAa,CAC3B,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,EACvB,UAAU,EACV,KAAK,CAAC,WAAW,CAAC,UAAU,CAAC,EAC7B,KAAK,CAAC,QAAQ,CACf,CAAC,CAAC;YACH,UAAU,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QAED,IAAI,QAAQ,GAAG,UAAU,CAAC;QAC1B,IAAI,SAAS,GAAG,CAAC,CAAC;QAElB,OAAO,QAAQ,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YAC1C,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;YACpD,IAAI,SAAS,GAAG,UAAU,GAAG,KAAK,CAAC,QAAQ,IAAI,QAAQ,GAAG,UAAU,EAAE,CAAC;gBACrE,MAAM;YACR,CAAC;YACD,SAAS,IAAI,UAAU,CAAC;YACxB,QAAQ,EAAE,CAAC;QACb,CAAC;QAED,IAAI,QAAQ,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,IAAI,QAAQ,GAAG,UAAU,GAAG,CAAC,EAAE,CAAC;YACpE,QAAQ,GAAG,mBAAmB,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,UAAU,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;QACtF,CAAC;QAED,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,WAAW,EAAE,UAAU,EAAE,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC;QAEtF,IAAI,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YACtC,MAAM;QACR,CAAC;QAED,MAAM,YAAY,GAAG,2BAA2B,CAC9C,KAAK,CAAC,KAAK,EACX,UAAU,EACV,QAAQ,EACR,KAAK,CAAC,YAAY,CACnB,CAAC;QACF,MAAM,SAAS,GAAG,QAAQ,GAAG,YAAY,CAAC;QAC1C,UAAU,GAAG,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;IAC7D,CAAC;IAED,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AACnE,CAAC;AAGD,SAAS,mBAAmB,CAC1B,KAAwB,EACxB,UAA8B,EAC9B,UAAkB,EAClB,QAAgB;IAEhB,MAAM,QAAQ,GAAG,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,UAAU,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;IACrF,IAAI,SAAS,GAAG,QAAQ,CAAC;IACzB,IAAI,SAAS,GAAG,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;IAEhE,KAAK,IAAI,KAAK,GAAG,QAAQ,EAAE,KAAK,IAAI,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC;QACtD,MAAM,KAAK,GAAG,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC;QAC3D,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YACtB,SAAS,GAAG,KAAK,CAAC;YAClB,SAAS,GAAG,KAAK,CAAC;QACpB,CAAC;IACH,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAGD,SAAS,kBAAkB,CACzB,KAAwB,EACxB,UAA8B,EAC9B,UAAkB;IAElB,IAAI,UAAU,IAAI,CAAC,IAAI,UAAU,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;QAC5E,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,OAAO,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,CAAC;IACzC,MAAM,QAAQ,GAAG,KAAK,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAE9C,IAAI,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/B,OAAO,GAAG,CAAC;IACb,CAAC;IAED,IAAI,QAAQ,KAAK,EAAE,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;QACtC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,QAAQ,KAAK,EAAE,EAAE,CAAC;QACpB,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,mBAAmB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QACtC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1B,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,OAAO,EAAE,CAAC;AACZ,CAAC;AAGD,SAAS,aAAa,CACpB,IAAY,EACZ,SAAiB,EACjB,UAAkB,EAClB,QAAgB;IAEhB,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACvC,MAAM,WAAW,GAAG,SAAS,IAAI,QAAQ;YACvC,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,QAAQ,CAAC,CAAC;QAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,WAAW,CAAC,CAAC;QAEtD,OAAO,CAAC,IAAI,CAAC;YACX,IAAI;YACJ,KAAK,EAAE;gBACL,IAAI,EAAE,MAAM;gBACZ,SAAS,EAAE,SAAS,GAAG,CAAC;gBACxB,OAAO,EAAE,SAAS,GAAG,CAAC;gBACtB,WAAW,EAAE,UAAU,GAAG,MAAM;gBAChC,SAAS,EAAE,UAAU,GAAG,MAAM,GAAG,IAAI,CAAC,MAAM;aAC7C;SACF,CAAC,CAAC;QACH,MAAM,IAAI,WAAW,CAAC;IACxB,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAGD,SAAS,aAAa,CACpB,KAAwB,EACxB,WAA8B,EAC9B,UAAkB,EAClB,QAAgB;IAEhB,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;QACtD,KAAK,EAAE;YACL,IAAI,EAAE,MAAM;YACZ,SAAS,EAAE,UAAU,GAAG,CAAC;YACzB,OAAO,EAAE,QAAQ,GAAG,CAAC;YACrB,WAAW,EAAE,WAAW,CAAC,UAAU,CAAC;YACpC,SAAS,EAAE,WAAW,CAAC,QAAQ,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM;SAC1D;KACF,CAAC;AACJ,CAAC;AAGD,SAAS,kBAAkB,CAAC,KAAwB;IAClD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrB,MAAM,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAGD,SAAS,iBAAiB,CAAC,KAAwB;IACjD,MAAM,OAAO,GAAG,IAAI,KAAK,CAAU,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC7D,IAAI,KAAK,GAAyB,IAAI,CAAC;IAEvC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QAClD,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,SAAS,EAAE,CAAC;QAEzC,IAAI,KAAK,EAAE,CAAC;YACV,OAAO,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACtB,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC9B,KAAK,GAAG,IAAI,CAAC;YACf,CAAC;YACD,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;QAChB,CAAC;aAAM,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACrC,KAAK,GAAG,KAAK,CAAC;QAChB,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAGD,SAAS,2BAA2B,CAClC,KAAwB,EACxB,UAAkB,EAClB,QAAgB,EAChB,YAAoB;IAEpB,IAAI,YAAY,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,CAAC;IACX,CAAC;IAED,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,IAAI,KAAK,GAAG,QAAQ,GAAG,CAAC,EAAE,KAAK,GAAG,UAAU,EAAE,KAAK,EAAE,EAAE,CAAC;QAC3D,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;QACjC,IAAI,KAAK,GAAG,YAAY,EAAE,CAAC;YACzB,MAAM;QACR,CAAC;QACD,KAAK,EAAE,CAAC;IACV,CAAC;IAED,OAAO,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AAClE,CAAC;AAGD,SAAS,WAAW,CAAC,IAAY,EAAE,QAAgB;IACjD,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC;IAC/C,IAAI,YAAY,GAAG,CAAC,CAAC,CAAC;IACtB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,IAAI,KAAK,GAAG,WAAW,EAAE,KAAK,GAAG,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC;QACxD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YAChE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YACvE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACnD,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACxE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;QAED,IAAI,KAAK,IAAI,SAAS,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACpC,SAAS,GAAG,KAAK,CAAC;YAClB,YAAY,GAAG,KAAK,GAAG,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,OAAO,YAAY,GAAG,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC;AACpD,CAAC;AAGD,SAAS,gBAAgB,CAAC,OAAgB;IACxC,OAAO;QACL,IAAI,EAAE,UAAU;QAChB,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,IAAI,IAAI,IAAI;QACtC,KAAK,EAAE,OAAO,CAAC,OAAO,EAAE,KAAK,IAAI,IAAI;QACrC,KAAK,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI;KAC5E,CAAC;AACJ,CAAC"}
+{"version":3,"file":"extractor.js","sourceRoot":"","sources":["../../../../src/engine/extraction/markdown/extractor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAC;AAM9C,OAAO,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AACzC,OAAO,EAAE,kBAAkB,EAAgC,MAAM,cAAc,CAAC;AAChF,OAAO,EAAE,yBAAyB,EAAE,MAAM,sBAAsB,CAAC;AAGjE,MAAM,4BAA4B,GAAG,IAAI,CAAC;AAC1C,MAAM,oCAAoC,GAAG,GAAG,CAAC;AAoBjD,MAAM,OAAO,iBAAiB;IAC5B,KAAK,CAAC,OAAO,CACX,MAAc,EACd,UAAwB,EAAE;QAE1B,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,KAAK,UAAU,EAAE,CAAC;YAChE,OAAO,EAAE,CAAC;QACZ,CAAC;QAED,kBAAkB,CAAC,MAAM,CAAC,CAAC;QAC3B,MAAM,YAAY,GAAG,2BAA2B,CAAC,OAAO,CAAC,CAAC;QAE1D,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;QACrC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC1B,OAAO,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;QAC7C,CAAC;QAED,MAAM,WAAW,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;QAC9C,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAC;QAC5C,MAAM,QAAQ,GAAG,aAAa,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;QAChD,MAAM,SAAS,GAAqB,EAAE,CAAC;QAEvC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,MAAM,QAAQ,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;YAC3C,MAAM,OAAO,GAAG,oBAAoB,CAAC;gBACnC,KAAK;gBACL,WAAW;gBACX,UAAU;gBACV,OAAO;gBACP,QAAQ,EAAE,YAAY,CAAC,aAAa;gBACpC,YAAY,EAAE,YAAY,CAAC,iBAAiB;aAC7C,CAAC,CAAC;YAEH,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACvB,MAAM,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC;gBAC1D,SAAS,CAAC,IAAI,CAAC;oBACb,EAAE;oBACF,KAAK,EAAE,EAAE;oBACT,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;oBACtB,KAAK,EAAE,aAAa,CAClB,KAAK,EACL,WAAW,EACX,OAAO,CAAC,UAAU,EAClB,OAAO,CAAC,QAAQ,CACjB,CAAC,KAAK;oBACP,OAAO,EAAE;wBACP,IAAI,EAAE,MAAM;wBACZ,IAAI,EAAE,eAAe,CAAC,QAAQ,CAAC;qBAChC;oBACD,QAAQ;iBACT,CAAC,CAAC;gBAEH,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;oBAC7B,SAAS,CAAC,IAAI,CACZ,wBAAwB,CACtB,MAAM,EACN,QAAQ,EACR,MAAM,EACN,SAAS,CAAC,MAAM,EAChB,EAAE,CACH,CACF,CAAC;gBACJ,CAAC;gBAED,SAAS;YACX,CAAC;YAED,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;gBAC7B,SAAS,CAAC,IAAI,CAAC;oBACb,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,CAAC,MAAM,CAAC;oBAClD,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;oBACtB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,OAAO,EAAE;wBACP,IAAI,EAAE,MAAM;wBACZ,IAAI,EAAE,MAAM,CAAC,IAAI;qBAClB;oBACD,QAAQ;iBACT,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO,SAAS,CAAC,MAAM,GAAG,CAAC;YACzB,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;IAC1C,CAAC;IAEO,QAAQ,CACd,MAAkB,EAClB,OAA+B;QAE/B,OAAO,yBAAyB,CAC9B,MAAM,EACN,OAAO,CAAC,aAAa,EACrB,OAAO,CAAC,iBAAiB,CAC1B,CAAC;IACJ,CAAC;CACF;AAED,SAAS,2BAA2B,CAClC,OAAqB;IAErB,MAAM,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,4BAA4B,CAAC;IAC5E,MAAM,iBAAiB,GACrB,OAAO,CAAC,iBAAiB,IAAI,oCAAoC,CAAC;IAEpE,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,aAAa,CAAC,IAAI,aAAa,IAAI,CAAC,EAAE,CAAC;QAC3D,MAAM,IAAI,WAAW,CACnB,2DAA2D,EAC3D;YACE,IAAI,EAAE,yDAAyD;YAC/D,OAAO,EAAE,iBAAiB,aAAa,EAAE;SAC1C,CACF,CAAC;IACJ,CAAC;IAED,IACE,CAAC,MAAM,CAAC,SAAS,CAAC,iBAAiB,CAAC;QACpC,iBAAiB,GAAG,CAAC;QACrB,iBAAiB,IAAI,aAAa,EAClC,CAAC;QACD,MAAM,IAAI,WAAW,CACnB,mEAAmE,EACnE;YACE,IAAI,EAAE,4DAA4D;YAClE,OAAO,EAAE,iBAAiB,aAAa,sBAAsB,iBAAiB,EAAE;SACjF,CACF,CAAC;IACJ,CAAC;IAED,OAAO,EAAE,aAAa,EAAE,iBAAiB,EAAE,CAAC;AAC9C,CAAC;AAED,SAAS,wBAAwB,CAC/B,MAAkB,EAClB,QAAgC,EAChC,MAAsB,EACtB,KAAa,EACb,KAAa;IAEb,OAAO;QACL,EAAE,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC;QACvC,KAAK;QACL,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE;QACtB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,OAAO,EAAE;YACP,IAAI,EAAE,MAAM;YACZ,IAAI,EAAE,MAAM,CAAC,IAAI;SAClB;QACD,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,SAAS,eAAe,CAAC,QAAgC;IACvD,OAAO,QAAQ,CAAC,OAAO,IAAI,kBAAkB,CAAC;AAChD,CAAC;AAED,SAAS,YAAY,CAAC,KAAwB;IAC5C,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,KAAK,GAAyB,IAAI,CAAC;IAEvC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QAClD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;QAC1B,MAAM,OAAO,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC;QAEjC,IAAI,KAAK,EAAE,CAAC;YACV,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC9B,KAAK,GAAG,IAAI,CAAC;YACf,CAAC;YACD,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;YACd,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;YACd,SAAS;QACX,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,4BAA4B,CAAC,CAAC;QACrD,IAAI,GAAG,EAAE,CAAC;YACR,QAAQ,CAAC,IAAI,CAAC;gBACZ,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM;gBACpB,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE;gBACnB,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAChE,QAAQ,CAAC,IAAI,CAAC;gBACZ,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnC,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;gBACjB,SAAS,EAAE,KAAK;aACjB,CAAC,CAAC;YACH,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,aAAa,CACpB,QAA4B,EAC5B,KAAwB;IAExB,MAAM,KAAK,GAAc,EAAE,CAAC;IAC5B,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAEjC,IACE,YAAY,CAAC,SAAS,GAAG,CAAC;QAC1B,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EACnE,CAAC;QACD,QAAQ,CAAC,IAAI,CAAC;YACZ,OAAO,EAAE,IAAI;YACb,UAAU,EAAE,CAAC;YACb,QAAQ,EAAE,YAAY,CAAC,SAAS,GAAG,CAAC;YACpC,UAAU,EAAE,EAAE;SACf,CAAC,CAAC;IACL,CAAC;IAED,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QACrD,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAEhC,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;YAC1E,KAAK,CAAC,GAAG,EAAE,CAAC;QACd,CAAC;QAED,QAAQ,CAAC,IAAI,CAAC;YACZ,OAAO;YACP,UAAU,EAAE,OAAO,CAAC,SAAS;YAC7B,QAAQ,EACN,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC,MAAM;gBACzB,CAAC,CAAC,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC;gBACnC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC;YACtB,UAAU,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SAC3C,CAAC,CAAC;QACH,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACtB,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,oBAAoB,CAAC,KAO7B;IACC,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,UAAU,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC;IAE1C,OAAO,UAAU,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;QAC5C,IAAI,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,GAAG,CAAC,GAAG,KAAK,CAAC,QAAQ,EAAE,CAAC;YACxD,OAAO,CAAC,IAAI,CACV,GAAG,aAAa,CACd,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,EACvB,UAAU,EACV,KAAK,CAAC,WAAW,CAAC,UAAU,CAAC,EAC7B,KAAK,CAAC,QAAQ,CACf,CACF,CAAC;YACF,UAAU,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QAED,IAAI,QAAQ,GAAG,UAAU,CAAC;QAC1B,IAAI,SAAS,GAAG,CAAC,CAAC;QAElB,OAAO,QAAQ,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YAC1C,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;YACpD,IAAI,SAAS,GAAG,UAAU,GAAG,KAAK,CAAC,QAAQ,IAAI,QAAQ,GAAG,UAAU,EAAE,CAAC;gBACrE,MAAM;YACR,CAAC;YACD,SAAS,IAAI,UAAU,CAAC;YACxB,QAAQ,EAAE,CAAC;QACb,CAAC;QAED,IAAI,QAAQ,IAAI,KAAK,CAAC,OAAO,CAAC,QAAQ,IAAI,QAAQ,GAAG,UAAU,GAAG,CAAC,EAAE,CAAC;YACpE,QAAQ,GAAG,mBAAmB,CAC5B,KAAK,CAAC,KAAK,EACX,KAAK,CAAC,UAAU,EAChB,UAAU,EACV,QAAQ,CACT,CAAC;QACJ,CAAC;QAED,OAAO,CAAC,IAAI,CACV,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,WAAW,EAAE,UAAU,EAAE,QAAQ,GAAG,CAAC,CAAC,CACxE,CAAC;QAEF,IAAI,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YACtC,MAAM;QACR,CAAC;QAED,MAAM,YAAY,GAAG,2BAA2B,CAC9C,KAAK,CAAC,KAAK,EACX,UAAU,EACV,QAAQ,EACR,KAAK,CAAC,YAAY,CACnB,CAAC;QACF,MAAM,SAAS,GAAG,QAAQ,GAAG,YAAY,CAAC;QAC1C,UAAU,GAAG,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;IAC7D,CAAC;IAED,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AACnE,CAAC;AAED,SAAS,mBAAmB,CAC1B,KAAwB,EACxB,UAA8B,EAC9B,UAAkB,EAClB,QAAgB;IAEhB,MAAM,QAAQ,GACZ,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,UAAU,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;IACtE,IAAI,SAAS,GAAG,QAAQ,CAAC;IACzB,IAAI,SAAS,GAAG,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;IAEhE,KAAK,IAAI,KAAK,GAAG,QAAQ,EAAE,KAAK,IAAI,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC;QACtD,MAAM,KAAK,GAAG,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC;QAC3D,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YACtB,SAAS,GAAG,KAAK,CAAC;YAClB,SAAS,GAAG,KAAK,CAAC;QACpB,CAAC;IACH,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,kBAAkB,CACzB,KAAwB,EACxB,UAA8B,EAC9B,UAAkB;IAElB,IAAI,UAAU,IAAI,CAAC,IAAI,UAAU,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;QAC5E,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,OAAO,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,CAAC;IACzC,MAAM,QAAQ,GAAG,KAAK,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAE9C,IAAI,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/B,OAAO,GAAG,CAAC;IACb,CAAC;IAED,IAAI,QAAQ,KAAK,EAAE,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;QACtC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,QAAQ,KAAK,EAAE,EAAE,CAAC;QACpB,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,mBAAmB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QACtC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1B,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,aAAa,CACpB,IAAY,EACZ,SAAiB,EACjB,UAAkB,EAClB,QAAgB;IAEhB,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACvC,MAAM,WAAW,GACf,SAAS,IAAI,QAAQ;YACnB,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,QAAQ,CAAC,CAAC;QAChD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,WAAW,CAAC,CAAC;QAEtD,OAAO,CAAC,IAAI,CAAC;YACX,IAAI;YACJ,KAAK,EAAE;gBACL,IAAI,EAAE,MAAM;gBACZ,SAAS,EAAE,SAAS,GAAG,CAAC;gBACxB,OAAO,EAAE,SAAS,GAAG,CAAC;gBACtB,WAAW,EAAE,UAAU,GAAG,MAAM;gBAChC,SAAS,EAAE,UAAU,GAAG,MAAM,GAAG,IAAI,CAAC,MAAM;aAC7C;SACF,CAAC,CAAC;QACH,MAAM,IAAI,WAAW,CAAC;IACxB,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,aAAa,CACpB,KAAwB,EACxB,WAA8B,EAC9B,UAAkB,EAClB,QAAgB;IAEhB,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;QACtD,KAAK,EAAE;YACL,IAAI,EAAE,MAAM;YACZ,SAAS,EAAE,UAAU,GAAG,CAAC;YACzB,OAAO,EAAE,QAAQ,GAAG,CAAC;YACrB,WAAW,EAAE,WAAW,CAAC,UAAU,CAAC;YACpC,SAAS,EAAE,WAAW,CAAC,QAAQ,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,MAAM;SAC1D;KACF,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAwB;IAClD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACrB,MAAM,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAwB;IACjD,MAAM,OAAO,GAAG,IAAI,KAAK,CAAU,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC7D,IAAI,KAAK,GAAyB,IAAI,CAAC;IAEvC,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QAClD,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,SAAS,EAAE,CAAC;QAEzC,IAAI,KAAK,EAAE,CAAC;YACV,OAAO,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC;YACtB,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC9B,KAAK,GAAG,IAAI,CAAC;YACf,CAAC;YACD,SAAS;QACX,CAAC;QAED,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9B,KAAK,GAAG,KAAK,CAAC;QAChB,CAAC;aAAM,IAAI,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACrC,KAAK,GAAG,KAAK,CAAC;QAChB,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,2BAA2B,CAClC,KAAwB,EACxB,UAAkB,EAClB,QAAgB,EAChB,YAAoB;IAEpB,IAAI,YAAY,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,CAAC;IACX,CAAC;IAED,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,IAAI,KAAK,GAAG,QAAQ,GAAG,CAAC,EAAE,KAAK,GAAG,UAAU,EAAE,KAAK,EAAE,EAAE,CAAC;QAC3D,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;QACjC,IAAI,KAAK,GAAG,YAAY,EAAE,CAAC;YACzB,MAAM;QACR,CAAC;QACD,KAAK,EAAE,CAAC;IACV,CAAC;IAED,OAAO,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AAClE,CAAC;AAED,SAAS,WAAW,CAAC,IAAY,EAAE,QAAgB;IACjD,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC;IAC/C,IAAI,YAAY,GAAG,CAAC,CAAC,CAAC;IACtB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,IAAI,KAAK,GAAG,WAAW,EAAE,KAAK,GAAG,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC;QACxD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YAChE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YACvE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACnD,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,GAAG,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACxE,KAAK,GAAG,CAAC,CAAC;QACZ,CAAC;QAED,IAAI,KAAK,IAAI,SAAS,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACpC,SAAS,GAAG,KAAK,CAAC;YAClB,YAAY,GAAG,KAAK,GAAG,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;IAED,OAAO,YAAY,GAAG,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC;AACpD,CAAC;AAED,SAAS,gBAAgB,CAAC,OAAgB;IACxC,OAAO;QACL,IAAI,EAAE,UAAU;QAChB,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,IAAI,IAAI,IAAI;QACtC,KAAK,EAAE,OAAO,CAAC,OAAO,EAAE,KAAK,IAAI,IAAI;QACrC,KAAK,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI;KAC5E,CAAC;AACJ,CAAC"}

package/dist/engine/extraction/runtime.d.ts

@@ -0,0 +1,4 @@
+import type { EntityFragment } from "../types.js";
+import type { Source } from "./source.js";
+import type { ChunkOptions } from "./types.js";
+export declare function extract(source: Source, options?: ChunkOptions): Promise<EntityFragment[]>;

package/dist/engine/extraction/runtime.js

@@ -0,0 +1,26 @@
+import { CodeExtractor } from "./code/extractor.js";
+import { ImageExtractor } from "./image/extractor.js";
+import { MarkdownExtractor } from "./markdown/extractor.js";
+import { TextExtractor } from "./text/extractor.js";
+const extractors = {
+    code: new CodeExtractor(),
+    image: new ImageExtractor(),
+    markdown: new MarkdownExtractor(),
+    text: new TextExtractor(),
+};
+export function extract(source, options = {}) {
+    return extractors[routeSource(source)].extract(source, options);
+}
+function routeSource(source) {
+    if (source.kind === "image") {
+        return "image";
+    }
+    if (source.file.kind === "code") {
+        return "code";
+    }
+    if (source.file.format === "markdown") {
+        return "markdown";
+    }
+    return "text";
+}
+//# sourceMappingURL=runtime.js.map

package/dist/engine/extraction/runtime.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"runtime.js","sourceRoot":"","sources":["../../../src/engine/extraction/runtime.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAE5D,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AASpD,MAAM,UAAU,GAAG;IACjB,IAAI,EAAE,IAAI,aAAa,EAAE;IACzB,KAAK,EAAE,IAAI,cAAc,EAAE;IAC3B,QAAQ,EAAE,IAAI,iBAAiB,EAAE;IACjC,IAAI,EAAE,IAAI,aAAa,EAAE;CACwB,CAAC;AAEpD,MAAM,UAAU,OAAO,CACrB,MAAc,EACd,UAAwB,EAAE;IAE1B,OAAO,UAAU,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AAClE,CAAC;AAED,SAAS,WAAW,CAAC,MAAc;IACjC,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;QAC5B,OAAO,OAAO,CAAC;IACjB,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;QAChC,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,KAAK,UAAU,EAAE,CAAC;QACtC,OAAO,UAAU,CAAC;IACpB,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}

package/dist/engine/extraction/source.d.ts

@@ -0,0 +1,18 @@
+import type { FileInfo, ImageFormat } from "../types.js";
+export type SourceKind = "text" | "image";
+type SourceBase = {
+    kind: SourceKind;
+    file: FileInfo;
+};
+export type TextSource = SourceBase & {
+    kind: "text";
+    text: string;
+};
+export type ImageSource = SourceBase & {
+    kind: "image";
+    data: Uint8Array;
+    format: ImageFormat;
+};
+export type Source = TextSource | ImageSource;
+export declare function validateSourceFile(source: Source): void;
+export {};

package/dist/engine/extraction/source.js

@@ -0,0 +1,22 @@
+import { EngineError } from "../errors.js";
+export function validateSourceFile(source) {
+    if (source.file.id.trim().length === 0) {
+        throw new EngineError("Extractor source requires a non-empty file id", {
+            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_FILE_ID",
+            context: `sourceKind=${source.kind}`,
+        });
+    }
+    if (source.file.absolutePath.trim().length === 0) {
+        throw new EngineError("Extractor source requires a non-empty absolute file path", {
+            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_ABSOLUTE_FILE_PATH",
+            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
+        });
+    }
+    if (source.file.relativePath.trim().length === 0) {
+        throw new EngineError("Extractor source requires a non-empty relative file path", {
+            code: "ZVEC_GREP.ENGINE.EXTRACTORS.EMPTY_RELATIVE_FILE_PATH",
+            context: `fileId=${source.file.id} sourceKind=${source.kind}`,
+        });
+    }
+}
+//# sourceMappingURL=source.js.map

package/dist/engine/extraction/source.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"source.js","sourceRoot":"","sources":["../../../src/engine/extraction/source.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAuB3C,MAAM,UAAU,kBAAkB,CAAC,MAAc;IAC/C,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvC,MAAM,IAAI,WAAW,CAAC,+CAA+C,EAAE;YACrE,IAAI,EAAE,2CAA2C;YACjD,OAAO,EAAE,cAAc,MAAM,CAAC,IAAI,EAAE;SACrC,CAAC,CAAC;IACL,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,WAAW,CACnB,0DAA0D,EAC1D;YACE,IAAI,EAAE,sDAAsD;YAC5D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CACF,CAAC;IACJ,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,WAAW,CACnB,0DAA0D,EAC1D;YACE,IAAI,EAAE,sDAAsD;YAC5D,OAAO,EAAE,UAAU,MAAM,CAAC,IAAI,CAAC,EAAE,eAAe,MAAM,CAAC,IAAI,EAAE;SAC9D,CACF,CAAC;IACJ,CAAC;AACH,CAAC"}

package/dist/engine/extraction/text/extractor.d.ts

@@ -0,0 +1,7 @@
+import type { EntityFragment } from "../../types.js";
+import { type Source, type TextSource } from "../source.js";
+import type { ChunkOptions } from "../types.js";
+export declare class TextExtractor {
+    extract(source: Source, options?: ChunkOptions): Promise<EntityFragment[]>;
+}
+export declare function extractPlainTextFragments(source: TextSource, maxChunkChars: number, chunkOverlapChars: number): EntityFragment[];