read_pdf_text

Instructions

Implementation Reference

• src/server.js:165-277 (handler)

  The primary handler function for the 'read_pdf_text' tool. It destructures arguments, validates the PDF file and pdftotext availability, constructs and executes the pdftotext command with options for page, layout, and encoding, computes metadata and statistics, and returns a structured JSON response or error.

  async handleReadPdfText(args) {
    try {
      const { 
        path: filePath, 
        page, 
        layout = false, 
        encoding = 'UTF-8' 
      } = args;
  
      // Check if pdftotext is available
      if (!this.checkPdftotextAvailable()) {
        throw new Error(
          'pdftotext is not available. Please install poppler-utils:\n' +
          '  Ubuntu/Debian: sudo apt install poppler-utils\n' +
          '  macOS: brew install poppler\n' +
          '  Windows: choco install poppler'
        );
      }
  
      // Validate the PDF file
      this.validatePdfFile(filePath);
  
      // Build pdftotext command
      const args_array = ['pdftotext'];
      
      // Add encoding if specified
      if (encoding !== 'UTF-8') {
        args_array.push('-enc', encoding);
      }
  
      // Add layout preservation if requested
      if (layout) {
        args_array.push('-layout');
      }
  
      // Add page specification if provided
      if (page) {
        args_array.push('-f', page.toString(), '-l', page.toString());
      }
  
      // Add input file and output to stdout
      args_array.push(filePath, '-');
  
      // Execute pdftotext
      const text = execSync(args_array.join(' '), { 
        encoding: 'utf8',
        maxBuffer: 50 * 1024 * 1024, // 50MB buffer for very large PDFs
        timeout: 30000, // 30 second timeout
      });
  
      // Get file metadata
      const stats = fs.statSync(filePath);
      const fileName = path.basename(filePath);
      const fileDir = path.dirname(path.resolve(filePath));
  
      // Prepare response
      const response = {
        success: true,
        file: fileName,
        path: path.resolve(filePath),
        directory: fileDir,
        extractedText: text.trim(),
        pageSpecific: page || 'all',
        layoutPreserved: layout,
        encoding: encoding,
        fileSize: stats.size,
        lastModified: stats.mtime.toISOString(),
        extractedAt: new Date().toISOString(),
        textLength: text.trim().length,
        wordCount: text.trim().split(/\s+/).filter(word => word.length > 0).length,
      };
  
      return {
        content: [
          {
            type: 'text',
            text: JSON.stringify(response, null, 2),
          },
        ],
      };
  
    } catch (error) {
      // Prepare error response
      const errorResponse = {
        success: false,
        error: error.message,
        file: args.path || 'unknown',
        timestamp: new Date().toISOString(),
      };
  
      // Add specific error context if available
      if (error.code === 'ENOENT') {
        errorResponse.errorType = 'FILE_NOT_FOUND';
      } else if (error.code === 'EACCES') {
        errorResponse.errorType = 'PERMISSION_DENIED';
      } else if (error.message.includes('pdftotext')) {
        errorResponse.errorType = 'PDFTOTEXT_ERROR';
      } else if (error.message.includes('PDF')) {
        errorResponse.errorType = 'INVALID_PDF';
      } else {
        errorResponse.errorType = 'UNKNOWN_ERROR';
      }
  
      return {
        content: [
          {
            type: 'text',
            text: JSON.stringify(errorResponse, null, 2),
          },
        ],
      };
    }
  }

• src/server.js:76-101 (schema)

  The input schema for the 'read_pdf_text' tool, defining the expected parameters: path (required string), optional page (number >=1), layout (boolean, default false), encoding (string enum, default UTF-8).

  inputSchema: {
    type: 'object',
    properties: {
      path: {
        type: 'string',
        description: 'Path to the PDF file (relative to current working directory or absolute path)',
      },
      page: {
        type: 'number',
        description: 'Specific page number to extract (1-based indexing). If not specified, extracts all pages.',
        minimum: 1,
      },
      layout: {
        type: 'boolean',
        description: 'Preserve original text layout formatting (default: false)',
        default: false,
      },
      encoding: {
        type: 'string',
        description: 'Text encoding for output (default: UTF-8)',
        default: 'UTF-8',
        enum: ['UTF-8', 'Latin1', 'ASCII'],
      },
    },
    required: ['path'],
  },

• src/server.js:107-113 (registration)

  Registration of the CallToolRequest handler that checks if the tool name is 'read_pdf_text' and dispatches to the handleReadPdfText function.

  this.server.setRequestHandler(CallToolRequestSchema, async (request) => {
    if (request.params.name !== 'read_pdf_text') {
      throw new Error(`Unknown tool: ${request.params.name}`);
    }
  
    return await this.handleReadPdfText(request.params.arguments);
  });

• src/server.js:71-104 (registration)

  Registration of the ListToolsRequest handler that provides the tool metadata including name 'read_pdf_text', description, and input schema.

  this.server.setRequestHandler(ListToolsRequestSchema, async () => ({
    tools: [
      {
        name: 'read_pdf_text',
        description: 'Extract text content from a PDF file using pdftotext from poppler-utils',
        inputSchema: {
          type: 'object',
          properties: {
            path: {
              type: 'string',
              description: 'Path to the PDF file (relative to current working directory or absolute path)',
            },
            page: {
              type: 'number',
              description: 'Specific page number to extract (1-based indexing). If not specified, extracts all pages.',
              minimum: 1,
            },
            layout: {
              type: 'boolean',
              description: 'Preserve original text layout formatting (default: false)',
              default: false,
            },
            encoding: {
              type: 'string',
              description: 'Text encoding for output (default: UTF-8)',
              default: 'UTF-8',
              enum: ['UTF-8', 'Latin1', 'ASCII'],
            },
          },
          required: ['path'],
        },
      },
    ],
  }));

• src/server.js:131-159 (helper)

  Helper function to validate the PDF file: checks existence, readability, .pdf extension, and PDF magic bytes (%PDF header). Called by the handler.

  validatePdfFile(filePath) {
    // Check if file exists
    if (!fs.existsSync(filePath)) {
      throw new Error(`File not found: ${filePath}`);
    }
  
    // Check if file is readable
    try {
      fs.accessSync(filePath, fs.constants.R_OK);
    } catch (error) {
      throw new Error(`File is not readable: ${filePath}`);
    }
  
    // Basic PDF file validation (check extension and magic bytes)
    if (!filePath.toLowerCase().endsWith('.pdf')) {
      throw new Error(`File does not appear to be a PDF: ${filePath}`);
    }
  
    try {
      const buffer = fs.readFileSync(filePath, { start: 0, end: 4 });
      if (!buffer.toString().startsWith('%PDF')) {
        throw new Error(`File is not a valid PDF (missing PDF header): ${filePath}`);
      }
    } catch (error) {
      if (error.message.includes('PDF header')) {
        throw error;
      }
      throw new Error(`Unable to validate PDF file: ${filePath}`);
    }