@customize-agent/knowledge 3.0.0 → 3.0.2

package/dist/core/knowledge-base-manager.d.ts

@@ -90,6 +90,13 @@ export declare class KnowledgeBaseManager {
     uploadFile(fileName: string, content: Buffer, targetRelativePath?: string, onProgress?: (progress: KnowledgeIndexProgress) => void, options?: {
         vectorMode?: 'sync' | 'defer';
     }): Promise<DiffResult>;
+    uploadFiles(files: Array<{
+        fileName: string;
+        content: Buffer;
+        targetRelativePath?: string;
+    }>, onProgress?: (progress: KnowledgeIndexProgress) => void, options?: {
+        vectorMode?: 'sync' | 'defer';
+    }): Promise<DiffResult>;
     listFailedFiles(): DiffResult['skippedFiles'];
     removeFile(relativePath: string): Promise<void>;
     tagFile(relativePath: string, tags: string[]): void;

package/dist/core/knowledge-base-manager.js

@@ -368,11 +368,16 @@ export class KnowledgeBaseManager {
         return targetRelativePath ?? this.defaultUploadRelativePath(fileName);
     }
     async uploadFile(fileName, content, targetRelativePath, onProgress, options = {}) {
+        return this.uploadFiles([{ fileName, content, targetRelativePath }], onProgress, options);
+    }
+    async uploadFiles(files, onProgress, options = {}) {
         this.initialize();
-        const relativePath = this.getUploadRelativePath(fileName, targetRelativePath);
-        const targetPath = this.resolveKbRelativePath(relativePath);
-        fs.mkdirSync(path.dirname(targetPath), { recursive: true });
-        fs.writeFileSync(targetPath, content);
+        for (const file of files) {
+            const relativePath = this.getUploadRelativePath(file.fileName, file.targetRelativePath);
+            const targetPath = this.resolveKbRelativePath(relativePath);
+            fs.mkdirSync(path.dirname(targetPath), { recursive: true });
+            fs.writeFileSync(targetPath, file.content);
+        }
         return this.incrementalIndex({ onProgress, vectorMode: options.vectorMode });
     }
     listFailedFiles() {

package/dist/extraction/content-extractor.d.ts

@@ -19,6 +19,8 @@ export declare class ContentExtractor {
     private parseDelimitedLine;
     private extractDelimitedText;
     private extractOfficeDocument;
+    private extractRtf;
+    private extractLegacyOfficeBinary;
     private extractSpreadsheet;
     private extractOfficeZip;
     private extractArchive;

package/dist/extraction/content-extractor.js

@@ -378,7 +378,12 @@ export class ContentExtractor {
         };
     }
     async extractOfficeDocument(file) {
-        if (path.extname(file.absolutePath).toLowerCase() === '.docx') {
+        const ext = path.extname(file.absolutePath).toLowerCase();
+        if (ext === '.rtf')
+            return this.extractRtf(file);
+        if (ext === '.doc' || ext === '.ppt')
+            return this.extractLegacyOfficeBinary(file);
+        if (ext === '.docx') {
             try {
                 const mammoth = await resolveAndImport('mammoth');
                 const result = await mammoth.extractRawText({ path: file.absolutePath });
@@ -397,7 +402,31 @@ export class ContentExtractor {
         }
         return this.extractOfficeZip(file);
     }
+    extractRtf(file) {
+        const raw = fs.readFileSync(file.absolutePath, 'utf8');
+        const text = raw
+            .replace(/\\'[0-9a-fA-F]{2}/gu, ' ')
+            .replace(/\\[a-zA-Z]+-?\d* ?/gu, ' ')
+            .replace(/[{}]/gu, ' ')
+            .replace(/\s+/gu, ' ')
+            .trim();
+        return {
+            text,
+            metadata: { extractionMode: 'builtin_rtf_text', vectorizable: true, contentCoverage: 'rtf_text' },
+            warnings: text ? [] : ['RTF 解析未提取到正文，未入库'],
+        };
+    }
+    extractLegacyOfficeBinary(file) {
+        const strings = this.extractBinaryStrings(file.absolutePath).slice(0, 1_000);
+        const text = strings.join('\n').trim();
+        return {
+            text,
+            metadata: { extractionMode: 'builtin_legacy_office_binary_strings', vectorizable: true, contentCoverage: 'legacy_office_binary_strings', stringCount: strings.length },
+            warnings: text ? [] : ['旧版 Office 二进制文件未提取到正文，未入库'],
+        };
+    }
     async extractSpreadsheet(file) {
+        const ext = path.extname(file.absolutePath).toLowerCase();
         try {
             const XLSX = await resolveAndImport('xlsx');
             const workbook = XLSX.readFile(file.absolutePath, { cellDates: true, cellFormula: true, cellNF: true, cellStyles: true });
@@ -442,8 +471,12 @@ export class ContentExtractor {
             }
         }
         catch {
+            if (ext === '.xls')
+                return this.extractLegacyOfficeBinary(file);
             // fallback below
         }
+        if (ext === '.xls')
+            return this.extractLegacyOfficeBinary(file);
         return this.extractOfficeZip(file);
     }
     async extractOfficeZip(file) {
@@ -461,13 +494,14 @@ export class ContentExtractor {
                     texts.push(`${entry.name}: ${stripped.slice(0, 8_000)}`);
             }
             metadata.entryCount = Object.keys(zip.files).length;
-            metadata.contentCoverage = texts.length > 0 ? 'office_zip_xml_text' : 'metadata_filename';
-            return { text: [this.metadataOnlyText(file), ...texts].join('\n'), metadata, warnings: texts.length ? [] : ['未从 Office 压缩结构中提取到正文，已跳过入库'] };
+            metadata.contentCoverage = texts.length > 0 ? 'office_zip_xml_text' : 'office_zip_empty_text';
+            return { text: texts.join('\n'), metadata, warnings: texts.length ? [] : ['Office/表格/演示文件未提取到正文，未入库'] };
         }
         catch (error) {
             metadata.extractionMode = 'office_zip_failed';
             metadata.parseError = error instanceof Error ? error.message : String(error);
-            return { text: '', metadata, warnings: ['Office/表格/演示文件解析失败，内置解析器未提取到正文，未入库'] };
+            metadata.contentCoverage = 'office_zip_failed';
+            return { text: '', metadata, warnings: [`Office/表格/演示文件解析失败: ${metadata.parseError}，未入库`] };
         }
     }
     async extractArchive(file) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@customize-agent/knowledge",
-  "version": "3.0.0",
+  "version": "3.0.2",
   "description": "Local knowledge base infrastructure for customize-agent",
   "type": "module",
   "main": "./dist/index.js",