Excel ↔ Json Converter

Author: qiangxinglin

Version: 0.0.5

Type: tool

Repository https://github.com/qiangxinglin/excel-dify-plugin

Description

The built-in Doc Extractor would convert input .xlsx file to markdown table string for downstream nodes (e.g. LLM). But this does not cover all situations! This plugin provides 2 tools:

• xlsx → json: Read the Excel file and output the Json presentation of the data.
• json → xlsx: Convert the given json string (list of records) to xlsx blob.

Usage

[!IMPORTANT] Correctly configure the FILES_URL in your docker-compose.yaml or .env in advance.

Tools

xlsx → json

• The output is placed in the text output field rather than the json field in order to preserving the header order.
• All cells are parsed as string, no matter what it is.
• If the uploaded Excel file contains multiple sheets, the plugin will automatically convert it into a JSON object, where each key is the sheet name and the value is the data array of the corresponding sheet.

Name	Age	Date
John	18	2020/2/20
Doe	20	2020/2/2

json → xlsx

• The output filename can be configured, default Converted_Data
• If the input JSON is an object (whose values are arrays), the plugin will automatically create a multi-sheet Excel file, where each key of the object will become a sheet name.

Format Settings

The plugin supports optional formatting for row heights and column widths using the [format] reserved key.

Note: Excel prohibits sheet names containing these characters: / \ ? * : [ ] Therefore, [format] is guaranteed to never conflict with actual sheet names.

[format] Structure

{
  "[format]": {
    "defaults": {
      "rowHeight": 20,           // Default height for all rows
      "columnWidth": 15,         // Default width for all columns
      "rowHeights": {            // Specific row heights (1-based)
        "1": 30,                 // Row 1 height = 30
        "2": 25                  // Row 2 height = 25
      },
      "columnWidths": {          // Specific column widths
        "A": 25,                 // Column A width = 25
        "B": 15                  // Column B width = 15
      }
    },
    "sheets": {
      "SheetName": {             // Per-sheet overrides
        "rowHeight": 22,
        "columnWidth": 18,
        "rowHeights": {"1": 28},
        "columnWidths": {"A": 30, "B": 20}
      }
    }
  },
  "SheetName": [...]             // Actual data
}

Format Priority Rules

Settings are applied in the following order (later overrides earlier):

1. [format].defaults.rowHeight / columnWidth - Global defaults for all rows/columns
2. [format].sheets.<name>.rowHeight / columnWidth - Per-sheet defaults
3. [format].defaults.rowHeights / columnWidths - Global specific rows/columns
4. [format].sheets.<name>.rowHeights / columnWidths - Per-sheet specific rows/columns

Validation and Warnings

• Unknown sheet references: If [format].sheets references a sheet that doesn't exist in the data, a warning will be displayed and those configurations will be ignored. The Excel file will still be generated successfully.
• Type errors: If format values have incorrect types (e.g., non-dict, negative numbers), an error will be thrown and the Excel generation will fail.

Examples

Single sheet with global formatting:

{
  "[format]": {
    "defaults": {
      "rowHeight": 20,
      "columnWidth": 15
    }
  },
  "Sheet1": [
    {"Name": "John", "Age": "18"},
    {"Name": "Doe", "Age": "20"}
  ]
}

Multiple sheets with per-sheet formatting:

{
  "[format]": {
    "defaults": {
      "rowHeight": 18
    },
    "sheets": {
      "Employees": {
        "columnWidths": {"A": 25, "B": 15},
        "rowHeights": {"1": 30}
      },
      "Departments": {
        "columnWidths": {"A": 20}
      }
    }
  },
  "Employees": [{"Name": "John", "Department": "R&D"}],
  "Departments": [{"ID": "1", "Name": "HR"}]
}

Column identifiers:

You can use either Excel letters or 1-based numeric indexes:

{
  "[format]": {
    "defaults": {
      "columnWidths": {
        "A": 25,    // Letter format
        "1": 25,    // Numeric format (same as "A")
        "B": 15,
        "2": 15     // Same as "B"
      }
    }
  },
  "Sheet1": [...]
}

Used Open sourced projects

• pandas, BSD 3-Clause License

Changelog

• 0.0.5: Add [format] metadata support for controlling row heights and column widths during JSON → Excel conversion
• 0.0.4: Add missing dependency (xlrd)
• 0.0.3: Add multi-sheet support for Excel processing (closes #13)

License

• Apache License 2.0

Privacy

This plugin collects no data.

All the file transformations are completed locally. NO data is transmitted to third-party services.