DataTableCompilation Task
Category: Writing
Compile structured data tables from multiple files using AI to identify rows, columns, and extract cell data.
Configuration
| Parameter | Type | Description |
|---|---|---|
| file_patterns | List<String> | List of file glob patterns to include in the data compilation (e.g., "**/*.md"). |
| output_file | String | REQUIRED: Output file path where the compiled data table will be saved (CSV or JSON). |
| row_identification_instructions | String | Instructions for the AI to identify what constitutes a row in the data. |
| column_identification_instructions | String | Instructions for the AI to identify what columns should be extracted. |
| cell_extraction_instructions | String | Instructions for extracting specific data values for each cell. |
Usage
This task is useful for aggregating scattered information into a structured format. For example, creating a summary table of experiments from multiple log files, or compiling a feature comparison matrix from product documentation.
Example Configuration
{
"task_type": "DataTableCompilation",
"file_patterns": ["reports/**/*.txt"],
"output_file": "summary_table.csv",
"row_identification_instructions": "Each row should represent a unique project mentioned in the reports.",
"column_identification_instructions": "Extract Project Name, Start Date, Status, and Budget.",
"cell_extraction_instructions": "Ensure dates are in YYYY-MM-DD format."
}