get_file_content
Extract text content from Google Drive files including Docs, Sheets, and text documents using file ID to access document data.
Instructions
Get content from a Google Drive file (Docs, Sheets, TXT, MD)
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| driveId | No | ||
| fileId | Yes |
Implementation Reference
- src/mcp/tools/get-file-content.ts:25-51 (handler)The handler function of the get_file_content tool. It validates the fileId parameter, fetches the file content using googleDriveService, and returns both text and structured content.handler: async (params: { fileId: string; driveId?: string }) => { const { fileId, driveId } = params; if (!fileId) { throw new Error("fileId is required"); } const fileContent = await googleDriveService.getFileContent( fileId, driveId ); return { content: [ { type: "text" as const, text: JSON.stringify(fileContent, null, 2), }, ], structuredContent: { fileId: fileContent.fileId, name: fileContent.fileName, mimeType: fileContent.mimeType, content: fileContent.content, }, }; },
- Input and output schemas defined using Zod for the get_file_content tool.config: { title: "Get File Content", description: "Get content from a Google Drive file (Docs, Sheets, TXT, MD)", inputSchema: { fileId: z.string().describe("ID of the file to read"), driveId: z.string().optional().describe("Drive ID (if specified)"), }, outputSchema: { fileId: z.string(), name: z.string(), mimeType: z.string(), content: z.string(), }, },
- src/mcp/server.ts:26-32 (registration)Dynamic registration of all exported tools (including get_file_content) to the MCP server using registerTool.// Registro automático de todas las tools const toolList = Object.values(tools); toolList.forEach((tool) => { server.registerTool(tool.name, tool.config as any, tool.handler as any); }); logger.info(`MCP Server initialized with ${toolList.length} tools`);
- Helper service method that performs the actual Google Drive API calls to retrieve and export file content based on file type.async getFileContent(fileId: string, driveId?: string): Promise<FileContent> { const targetDriveId = driveId || Object.keys(drivesConfigLoader.getConfig().drives)[0]; const drive = await this.getDriveClient(targetDriveId); try { // Obtener metadatos del archivo const metadata = await drive.files.get({ fileId, fields: "id, name, mimeType", }); const file = metadata.data; const mimeType = file.mimeType!; let content = ""; // Google Docs: exportar como texto plano if (mimeType === SUPPORTED_MIME_TYPES.GOOGLE_DOC) { const response = await drive.files.export({ fileId, mimeType: SUPPORTED_MIME_TYPES.EXPORT_TEXT, }); content = response.data as string; } // Google Sheets: exportar como CSV else if (mimeType === SUPPORTED_MIME_TYPES.GOOGLE_SHEET) { const response = await drive.files.export({ fileId, mimeType: SUPPORTED_MIME_TYPES.EXPORT_CSV, }); content = response.data as string; } // Archivos de texto plano: descarga directa else if ( mimeType === SUPPORTED_MIME_TYPES.TEXT_PLAIN || mimeType === SUPPORTED_MIME_TYPES.TEXT_MARKDOWN || mimeType === "text/markdown" || file.name?.endsWith(".txt") || file.name?.endsWith(".md") ) { const response = await drive.files.get( { fileId, alt: "media" }, { responseType: "text" } ); content = response.data as string; } // Tipo de archivo no soportado else { throw new Error(`Unsupported file type: ${mimeType}`); } logger.info(`Retrieved content for file: ${file.name} (${mimeType})`); return { fileId: file.id!, fileName: file.name!, mimeType, content, extractedAt: new Date().toISOString(), }; } catch (error) { logger.error("Error getting file content", { fileId, driveId: targetDriveId, error, }); throw error; } }
- src/config/types.ts:121-124 (schema)TypeScript interface defining the parameters for getFileContent.export interface GetFileContentParams { fileId: string; driveId?: string; }
- src/config/types.ts:92-98 (schema)TypeScript interface for the file content returned by the service.export interface FileContent { fileId: string; fileName: string; mimeType: string; content: string; extractedAt: string; }