Convert Data Between Formats in Jinba Flow

Use CONVERTER_ARRAY_TO_CSV, CONVERTER_HTML_TO_PDF, CONVERTER_HTML_TO_DOCX, and CONVERTER_IMAGE_TO_PDF to convert data between different formats and structures, such as CSV export, HTML to PDF or DOCX, and image to PDF, in a Jinba Flow workflow.

Data Conversion Tools provide essential functionality for transforming data between different formats. These tools handle array-to-CSV conversion, image-to-PDF conversion, HTML-to-PDF/DOCX conversion, and other data transformation needs in your workflows.

Tools

CONVERTER_ARRAY_TO_CSV

• Convert 2D arrays to CSV format with automatic escaping
• Handles special characters (quotes, commas, newlines)
• Text cells that start with =, +, -, @, tab or carriage return are prefixed with an apostrophe (') so spreadsheet applications show them as text instead of evaluating them as formulas; numeric values (e.g. -5, +1.5e3) are written as-is
• Input: input (2D array) — Output: result (CSV string)
• Compatible with spreadsheet applications

CONVERTER_IMAGE_TO_PDF

• Convert images to PDF format
• Support for TIFF, PNG, JPEG, GIF, and more
• Input: file_link (array of image URLs, max 10 files)
• Output: result (URL of the generated PDF)

CONVERTER_HTML_TO_PDF

• Convert HTML content to a PDF file
• Input: html_content (HTML string), filename (optional, without extension)
• Output: result (signed URL to the created PDF file)

CONVERTER_HTML_TO_DOCX

• Convert HTML content to a DOCX (Word) file
• Input: html_content (HTML string), filename (optional, without extension)
• Output: result (signed URL to the created DOCX file)

Authentication

No authentication required for conversion tools.

Export data to CSV

- id: fetch_data
  name: fetch_data
  tool: PYTHON_SANDBOX_RUN
  input:
    - name: data_type
      value: OBJECT
    - name: code
      value: |
        # Sample data processing
        import json
        
        # Simulate fetching data from API or database
        raw_data = [
            {"name": "John Doe", "email": "[e-mail]", "age": 30, "city": "New York"},
            {"name": "Jane Smith", "email": "[e-mail]", "age": 25, "city": "Los Angeles"},
            {"name": "Bob Johnson", "email": "[e-mail]", "age": 35, "city": "Chicago"}
        ]
        
        # Convert to 2D array format
        headers = ["Name", "Email", "Age", "City"]
        data_rows = []
        data_rows.append(headers)
        
        for item in raw_data:
            row = [item["name"], item["email"], str(item["age"]), item["city"]]
            data_rows.append(row)
        
        print(json.dumps({"rows": data_rows}))

- id: convert_to_csv
  name: convert_to_csv
  tool: CONVERTER_ARRAY_TO_CSV
  needs: [fetch_data]
  input:
    - name: input
      value: "{{steps.fetch_data.result.data.rows}}"

- id: save_csv_file
  name: save_csv_file
  tool: OUTPUT_FILE
  needs: [convert_to_csv]
  input:
    - name: content
      value: "{{steps.convert_to_csv.result}}"
    - name: filename
      value: "user_data_export.csv"
    - name: fileType
      value: "csv"

Example: Sales Report Generation

- id: calculate_sales_metrics
  name: calculate_sales_metrics
  tool: PYTHON_SANDBOX_RUN
  input:
    - name: data_type
      value: OBJECT
    - name: code
      value: |
        import json
        from datetime import datetime, timedelta
        
        # Sample sales data
        sales_data = [
            {"product": "Widget A", "sales": 1500, "region": "North", "month": "Jan"},
            {"product": "Widget B", "sales": 2300, "region": "South", "month": "Jan"},
            {"product": "Widget A", "sales": 1800, "region": "North", "month": "Feb"},
            {"product": "Widget B", "sales": 2100, "region": "South", "month": "Feb"}
        ]
        
        # Create summary table
        summary = [["Product", "Region", "Month", "Sales", "Growth"]]
        
        for i, item in enumerate(sales_data):
            growth = "N/A"
            if i >= 2:  # Calculate growth for Feb data
                prev_sales = sales_data[i-2]["sales"]
                growth = f"{((item['sales'] - prev_sales) / prev_sales * 100):.1f}%"
            
            summary.append([
                item["product"],
                item["region"], 
                item["month"],
                f"${item['sales']:,}",
                growth
            ])
        
        print(json.dumps({"rows": summary}))

- id: export_sales_csv
  name: export_sales_csv
  tool: CONVERTER_ARRAY_TO_CSV
  needs: [calculate_sales_metrics]
  input:
    - name: input
      value: "{{steps.calculate_sales_metrics.result.data.rows}}"

- id: create_downloadable_report
  name: create_downloadable_report
  tool: OUTPUT_FILE
  needs: [export_sales_csv]
  input:
    - name: content
      value: "{{steps.export_sales_csv.result}}"
    - name: filename
      value: "sales_report.csv"
    - name: fileType
      value: "csv"

Convert an image to PDF

- id: input_images
  name: input_images
  tool: INPUT_IMAGE_FILE
  input:
    - name: value
      value: "https://example.com/chart1.png"
    - name: description
      value: "Sales chart image"

- id: convert_chart_to_pdf
  name: convert_chart_to_pdf
  tool: CONVERTER_IMAGE_TO_PDF
  input:
    - name: file_link
      value:
        - "https://example.com/chart1.png"

- id: convert_multiple_charts
  name: convert_multiple_charts
  tool: CONVERTER_IMAGE_TO_PDF
  input:
    - name: file_link
      value:
        - "https://example.com/chart1.png"
        - "https://example.com/chart2.png"
        - "https://example.com/chart3.png"

Convert HTML to PDF or DOCX

- id: generate_html_report
  name: generate_html_report
  tool: OPENAI_INVOKE
  config:
    - name: version
      value: gpt-5.2
  input:
    - name: prompt
      value: |
        Generate an HTML report summarizing this month's sales results.
        Return only the HTML content.

- id: convert_report_to_pdf
  name: convert_report_to_pdf
  tool: CONVERTER_HTML_TO_PDF
  needs: [generate_html_report]
  input:
    - name: html_content
      value: "{{steps.generate_html_report.result.content}}"
    - name: filename
      value: "monthly_sales_report"

- id: convert_report_to_docx
  name: convert_report_to_docx
  tool: CONVERTER_HTML_TO_DOCX
  needs: [generate_html_report]
  input:
    - name: html_content
      value: "{{steps.generate_html_report.result.content}}"
    - name: filename
      value: "monthly_sales_report"

# steps.convert_report_to_pdf.result is a signed URL to the PDF file
# steps.convert_report_to_docx.result is a signed URL to the DOCX file

Example: Complex Data Transformation

- id: input_excel
  name: input_excel
  tool: INPUT_FILE
  input:
    - name: description
      value: "Excel file to transform"

- id: extract_complex_data
  name: extract_complex_data
  tool: EXCEL_GET_ROWS
  needs: [input_excel]
  input:
    - name: url
      value: "{{steps.input_excel.result}}"
    - name: range
      value: "A1:F1000"

- id: transform_and_clean_data
  name: transform_and_clean_data
  tool: PYTHON_SANDBOX_RUN
  needs: [extract_complex_data]
  input:
    - name: data_type
      value: OBJECT
    - name: code
      value: |
        import json
        import re
        
        # Parse the Excel data
        excel_data = json.loads(r'''{{steps.extract_complex_data.result}}''')
        
        # Clean and transform data
        cleaned_data = []
        headers = ["ID", "Name", "Email", "Phone", "Status", "Notes"]
        cleaned_data.append(headers)
        
        for row in excel_data[1:]:  # Skip header row
            if len(row) >= 6:
                # Clean email
                email = row[2].strip().lower() if row[2] else ""
                
                # Format phone number
                phone = re.sub(r'[^\d+]', '', row[3]) if row[3] else ""
                
                # Clean and escape notes (may contain commas, quotes)
                notes = str(row[5]).replace('"', '""') if row[5] else ""
                
                cleaned_row = [
                    str(row[0]) if row[0] else "",
                    str(row[1]).title() if row[1] else "",
                    email,
                    phone,
                    str(row[4]).upper() if row[4] else "PENDING",
                    notes
                ]
                cleaned_data.append(cleaned_row)
        
        print(json.dumps(cleaned_data))

- id: convert_cleaned_data
  name: convert_cleaned_data
  tool: CONVERTER_ARRAY_TO_CSV
  needs: [transform_and_clean_data]
  input:
    - name: input
      value: "{{steps.transform_and_clean_data.result.data}}"

- id: save_cleaned_export
  name: save_cleaned_export
  tool: OUTPUT_FILE
  needs: [convert_cleaned_data]
  input:
    - name: content
      value: "{{steps.convert_cleaned_data.result}}"
    - name: filename
      value: "cleaned_data_export.csv"
    - name: fileType
      value: "csv"

Data Format Support

CSV Conversion

• Input: 2D arrays, JSON arrays, structured data
• Output: RFC 4180 compliant CSV
• Features: Automatic escaping, encoding handling

Image to PDF

• Supported formats: TIFF, PNG, JPEG, GIF, and more
• Input: Up to 10 image URLs per conversion
• Output: PDF document URL

HTML to PDF / DOCX

• Input: HTML content string and optional filename (without extension)
• Output: Signed URL to the generated PDF or DOCX file
• Features: Useful for turning LLM-generated HTML into downloadable documents

Best Practices

CSV Handling

• Always include headers in your data arrays
• Handle special characters (commas, quotes, newlines)
• Use consistent data types within columns
• Validate data before conversion

Image Conversion

• Ensure images are accessible via public URLs
• Consider file size limitations
• Optimize images before conversion for better performance
• Use appropriate filenames for organization

Use Cases

• Data Export: Export database queries to CSV for Excel
• Report Generation: Convert processed data to spreadsheet format
• Document Creation: Convert charts and images to PDF
• Data Migration: Transform data between different systems
• Analytics Export: Export analysis results for stakeholders
• Backup and Archive: Convert data to portable formats
• Integration: Prepare data for external system import