README.md

metadata

title: OpenEnv Data Cleaner
emoji: 🧹
colorFrom: indigo
colorTo: purple
sdk: docker
pinned: false
base_path: /web

OpenEnv Data Cleaner

An OpenEnv-compliant AI-powered data cleaning environment built on openenv-core.

Features

• OpenEnv-native: Built using openenv-core base classes
• Data Cleaning Actions: Drop nulls, fill nulls, remove duplicates, filter rows, drop columns, convert types, validate emails, outlier removal, normalization
• Task-based Learning: Three difficulty levels (easy, medium, hard)
• Grading System: Deterministic scoring based on data quality improvements
• Reward System: Structured rewards with quality, progress, and penalty components
• Web Interface: Interactive UI for manual data cleaning
• Docker Ready: Deployable to Hugging Face Spaces

Quick Start

# Install dependencies
pip install -r requirements.txt

# Run the server
python app.py

API Endpoints

• GET / - Web interface
• GET /health - Health check
• POST /reset - Initialize a new task
• POST /step - Execute a cleaning action
• POST /submit - Submit solution for grading
• POST /revert - Revert last action
• GET /tasks - List available tasks
• GET /state - Get current environment state
• GET /dataset - Get dataset information
• GET /history - Get action history

Tasks

Task ID	Difficulty	Description
easy_001	Easy	Basic cleaning: drop nulls and remove duplicates
medium_001	Medium	Intermediate: handle nulls, validate emails, remove outliers
hard_001	Hard	Advanced: full pipeline with type conversion and normalization

Deployment

Deploy to Hugging Face Spaces:

openenv push ./env