Skip to main content

​
Overview

The verify_environment.py script validates that all required dependencies, packages, and datasets are properly installed and accessible before running experiments. Location: scripts/verify_environment.py

​
Usage

python scripts/verify_environment.py
No command-line arguments required. The script automatically checks all dependencies and reports their status.

​
Verification Checks

​
System Information

  • Python Version: Reports the active Python interpreter version
  • Platform: Displays the operating system and platform details

​
Required Packages

Verifies installation and version of critical dependencies:
  • numpy - Numerical computing
  • matplotlib - Plotting and visualization
  • psutil - System and process utilities
  • requests - HTTP library
  • tqdm - Progress bars

​
Optional Packages

Checks for optional dependencies:
  • torch - PyTorch deep learning framework
  • pytest - Testing framework

​
Dataset Status

Validates Fashion-MNIST dataset availability:
  • Training set: Checks file existence and size
  • Test set: Checks file existence and size
  • Uses dataset configuration from dataset_config.py

​
Example Output

Python: 3.11.4
Platform: Linux-5.15.0-1028-aws-x86_64-with-glibc2.35

Required packages:
  - numpy: OK (1.24.3)
  - matplotlib: OK (3.7.1)
  - psutil: OK (5.9.5)
  - requests: OK (2.31.0)
  - tqdm: OK (4.65.0)

Optional packages:
  - torch: OK (2.0.1)
  - pytest: OK (7.4.0)

Dataset status:
  - fashion-mnist: train_exists=True size=30596944, test_exists=True size=5148144

​
Exit Behavior

The script does not exit with error codes. It reports all findings to stdout for manual review. This allows users to:
  • Identify missing packages
  • Verify version compatibility
  • Check dataset integrity
  • Diagnose environment issues before running experiments

​
Implementation Details

​
Module Checking

Uses dynamic import to test package availability: 
importlib.import_module(name)
Extracts version information via __version__ attribute when available.

​
Dataset Verification

  • Imports FASHION_MNIST_SPEC from dataset_config.py
  • Checks file existence using Path.exists()
  • Reports file size in bytes for integrity validation

​
Path Resolution

  • Automatically resolves repository root from script location
  • Adds task directory to sys.path for local imports
  • Uses pathlib.Path for cross-platform compatibility

​
Troubleshooting

​
Missing Package

If a required package shows MISSING, install it: 
pip install <package-name>

​
Dataset Unavailable

If dataset status shows train_exists=False or test_exists=False:
  1. Run the dataset download script
  2. Verify dataset_config.py paths are correct
  3. Check file permissions

​
Import Errors

If dataset_config import fails:
  • Ensure you’re running from the repository root
  • Verify the task directory structure exists
  • Check for syntax errors in dataset_config.py

​
Use Cases

​
Pre-Experiment Validation

python scripts/verify_environment.py
Run before starting experiments to catch environment issues early.

​
CI/CD Pipeline

python scripts/verify_environment.py > environment_report.txt
Capture environment state for debugging and reproducibility documentation.

​
Debugging Environment Issues

python scripts/verify_environment.py
Quickly diagnose missing dependencies or dataset problems when experiments fail.

Build docs developers (and LLMs) love

Get started for freeTalk to us