[FEATURE] Auto-Generate Field Definitions from Uploaded Evaluation Data Headers #142
Description
📋 Please ensure that:
- I have searched existing issues to avoid duplicates
- I have provided a clear problem statement and solution
- I understand this is a feature request and not a bug report
- I am willing to help implement this feature if needed
- I have submitted this feature request in English (otherwise it will not be processed)
🎯 Problem Statement
Currently, when creating an evaluation, users need to manually add fields one by one in the platform before uploading a CSV. This becomes very time-consuming and error-prone when dealing with datasets that contain dozens or even hundreds of fields.
We propose adding support for automatically reading column headers from uploaded CSV (or Excel) files and generating the corresponding field definitions automatically.
💡 Proposed Solution
When a user uploads a CSV/Excel file for evaluation:
Automatically parse the first row (header) as field names.
Generate field definitions dynamically in the platform without requiring manual input.
Allow users to adjust or rename fields after auto-generation if needed.
Ensure compatibility with large datasets (hundreds of fields).
📋 Use Cases
Decision Model Evaluation
A team has a decision model with hundreds of structured input fields.
They prepare an evaluation dataset with fields like user_id, age, income, credit_score, …
Currently, they must manually define all fields in the UI, which is impractical.
With auto-header parsing, the platform generates all field definitions as soon as the dataset is uploaded.
Prompt Construction with Structured Data
Many test cases require concatenating multiple structured fields into the prompt.
By reading headers automatically, users can immediately map fields into prompt templates without additional setup.
⚡ Priority
Medium - Nice to have
🔧 Component
Evaluation
🔄 Alternatives Considered
No response
🎨 Mockups/Designs
No response
⚙️ Technical Details
No response
✅ Acceptance Criteria
No response
📝 Additional Context
No response