Uvify

Turn python repositories to uv environments and oneliners, without diving into the code.

• Generates oneliners for quick python environment setup
• Helps with migration to uv for faster builds in CI/CD
• It works on existing projects based on: requirements.txt, pyproject.toml or setup.py, recursively.
  • Supports local directories.
  • Supports GitHub links using Git Ingest.
• It's fast!

Prerequisites

| uv

Demo

https://huggingface.co/spaces/avilum/uvify

uv is by far the fastest python and package manager.

Source: https://github.com/astral-sh/uv

You can run uvify with uv.
Let's generate oneliners for a virtual environment that has requests installed, using PyPi or from source:

# Run on a local directory
uvx uvify . | jq

# Run on requests
uvx uvify https://github.com/psf/requests | jq
# or:
# uvx uvify psf/requests | jq

[
  ...
  {
    "file": "setup.py",
    "fileType": "setup.py",
    "oneLiner": "uv run --python '>=3.8.10' --with 'certifi>=2017.4.17,charset_normalizer>=2,<4,idna>=2.5,<4,urllib3>=1.21.1,<3,requests' python -c 'import requests; print(requests)'",
    "uvInstallFromSource": "uv run --with 'git+https://github.com/psf/requests' --python '>=3.8.10' python",
    "dependencies": [
      "certifi>=2017.4.17",
      "charset_normalizer>=2,<4",
      "idna>=2.5,<4",
      "urllib3>=1.21.1,<3"
    ],
    "packageName": "requests",
    "pythonVersion": ">=3.8",
    "isLocal": false
  }
]

Parse all python artifacts in repository:

uvify psf/requests
uvify https://github.com/psf/requests

Parse specific fields in the response

uvify psf/black | jq '.[] | {file: .file, pythonVersion: .pythonVersion, dependencies: .dependencies, packageName: .packageName}'

Use existing python repos with 'uv':

uvify psf/requests | jq '.[0].oneLiner'
"uv run --with 'git+https://github.com/psf/requests' --python '3.11' python"

Install a repository with 'uv' from github sources:

uvify psf/requests | jq '.[0].dependencies'

List the dependencies.

uvify psf/requests | jq '.[].dependencies'
[
  "certifi>=2017.4.17",
  "charset_normalizer>=2,<4",
  "idna>=2.5,<4",
  "urllib3>=1.21.1,<3"
]

Filtering Options

Uvify supports filtering which files to analyze using include and exclude patterns with glob syntax.

Exclude directories from analysis

Skip test directories and any paths matching the pattern:

uvify --exclude "tests/*" --exclude "test_*" my-project/

Include only specific directories

Analyze only files in the src/ directory:

uvify --include "src/*" my-project/

Analyze only a specific subdirectory:

uvify --include "src/my_app/*" my-project/

Combine include and exclude patterns

Include everything in src/ but exclude test files:

uvify --include "src/*" --exclude "*/test_*" --exclude "*/tests/*" my-project/

GitHub repositories with filtering

The filtering also works with GitHub repositories:

# Exclude test directories from a GitHub repo
uvify --exclude "tests/*" psf/requests

# Only analyze specific subdirectories  
uvify --include "src/*" --include "lib/*" myorg/myrepo

Pattern Examples

• tests/* - Excludes any directory named "tests" and all its contents
• test_* - Excludes any file or directory starting with "test_"
• */tests/* - Excludes "tests" directories at any depth
• src/my_app/* - Includes only files within the "src/my_app/" directory
• *.py - Includes only Python files

Note: By default, uvify scans all directories for requirements.txt, pyproject.toml, and setup.py files. The include/exclude patterns filter which of these files to analyze based on their path.

Uvify HTTP Server: Using uvify with client/server architecture instead of SDK

First, install uvify with the optional API dependencies:

uv add uvify[api]
# or with pip:
# pip install uvify[api]

Then run the server:

# Run the server using the built-in serve command
uvify serve --host 0.0.0.0 --port 8000

# Or using uvicorn directly
uv run uvicorn src.uvify:api --host 0.0.0.0 --port 8000

# Using curl
curl http://0.0.0.0:8000/psf/requests | jq

# Using wget
wget -O-  http://0.0.0.0:8000/psf/requests | jq

Developing

# Install dependencies (including optional API dependencies)
uv venv
uv sync --dev --extra api
uv run pytest

# Run linter before PR 
./lint.sh

# Install editable version locally
uv run pip install --editable .
uv run python -m src.uvify --help
uv run python -m src.uvify psf/requests

# Run the HTTP API with reload
uv run uvicorn src.uvify:api --host 0.0.0.0 --port 8000 --reload 
# Or use the built-in serve command:
# uv run python -m src.uvify serve --host 0.0.0.0 --port 8000

curl http://0.0.0.0:8000/psf/requests | jq

Special Thanks

Thanks to the UV team and Astral for this amazing tool.