setup-workspace

Set Up Workspace

Initialize a workspace for synthetic data generation tasks. This creates folder structure, a Python requirements file, and optionally a virtual environment.

When to use

• Starting a new synthetic data project
• Want to organize outputs, inputs, configs, and reports in one place
• Need a repeatable, version-controlled setup

Inputs to gather

• Workspace path (optional; default: ./synthetic-data-workspace/ in current directory)
• Python version check (requires 3.10+; script offers venv setup)
• Tool selection (SDV, Faker, Synthcity, etc. — user can customize requirements later)

Procedure

1. Create folder structure:

   mkdir -p ./synthetic-data-workspace/{inputs,outputs,reports,configs}
   

2. Create a README.md describing the workspace layout:

   # Synthetic Data Workspace
   
   - `inputs/` — Real source datasets (CSV, Parquet, JSON)
   - `outputs/` — Generated synthetic datasets
   - `reports/` — Quality evaluation reports
   - `configs/` — YAML/JSON configuration files (schemas, synthesis params)
   
   ## Quick start
   
   1. Activate venv: `source venv/bin/activate` (if created)
   2. Place real data in `inputs/`
   3. Define schema or config in `configs/`
   4. Run synthesis skill (e.g. `tabular-from-real`)
   5. Check outputs in `outputs/` and reports in `reports/`
   

3. Create requirements.txt with common tools:

   sdv>=1.10.0
   sdmetrics>=0.11.0
   faker>=20.0.0
   mimesis>=7.0.0
   synthcity>=0.2.0
   pandas>=2.0.0
   numpy>=1.24.0
   pyarrow>=12.0.0
   scikit-learn>=1.3.0
   umap-learn>=0.5.0
   

   (User can edit to remove unnecessary packages.)

4. Check Python version:

   python3 --version  # Must be 3.10 or higher
   

5. Offer venv setup:

   cd ./synthetic-data-workspace
   python3 -m venv venv
   source venv/bin/activate
   pip install --upgrade pip
   pip install -r ../requirements.txt
   

6. Record workspace path for other skills (e.g. in a .workspace-config file or environment variable):

   echo "SYNTHETIC_DATA_WORKSPACE=$(pwd)" > .workspace-config
   

Output / side effects

• Folder structure in place: ./synthetic-data-workspace/{inputs,outputs,reports,configs}/
• requirements.txt created and ready to customize
• README.md with usage guide
• Optional: virtual environment activated, dependencies installed
• Workspace path recorded for future skill invocations

Safety / constraints

• If workspace folder already exists, check for conflicts before overwriting structure
• Recommend the user commit workspace config to version control (except large data files — add inputs/ and outputs/ to .gitignore)