Repository Stats

This simple tool aims on open-source projects providing simple repository stats which are a bit out of scope of base Git and need some more information about issues and PRs.

Highlighted features

• cumulative caching (no need to full download, just incremental/needed update)
• collection of overall user contributions to issues/PRs
• visualization of aggregated timeline of past contributions

Installation

Simple install with setuptools/pip as

pip install https://github.com/Borda/pyRepoStats/archive/main.zip

or after cloning the repository

python setup.py install

Sample usage

The CLI provides two commands: scrape (fetch data from GitHub) and analyze (analyze cached data).

Basic command structure

# General format
repostat <command> <repository> [options]

# Or using python -m
python -m repo_stats <command> <repository> [options]

Available commands

1. scrape - Fetch repository data from GitHub (always requires internet connection)
2. analyze - Analyze previously fetched data (works offline by default)

Examples

• If you just clone this repo without installation, you need to install dependencies and call script:

  pip install -r requirements.txt
  python src/repo_stats/__main__.py scrape PyTorchLightning/pytorch-lightning-bolts
  

• If you have already installed the package with pip or with setup.py you can call the executable:

  # Scrape (fetch) data from GitHub
  repostat scrape PyTorchLightning/pytorch-lightning-bolts --auth_token <your-personal-token>
  
  # Analyze the scraped data
  repostat analyze PyTorchLightning/pytorch-lightning-bolts --users_summary+ "all"
  

• Or use package with a pythonic way:

  # Scrape data
  python -m repo_stats scrape PyTorchLightning/pytorch-lightning-bolts
  
  # Analyze data
  python -m repo_stats analyze PyTorchLightning/pytorch-lightning-bolts --min_contribution 5
  

Note: When using python -m repo_stats, consider passing --output_path argument for output path, otherwise all caches and results will be saved in installation folder, most likely site-packages

To simplify the token passing in each call, you can export the token to environment variables export GH_API_TOKEN=<your-personal-token> for Github.

Github use-case

For GitHub users we recommend using your personal GitHub token which significantly increases request limit per hour.

Command-specific options

Use --help to see all available options for each command:

# See all commands
python -m repo_stats --help

# See options for scrape command
python -m repo_stats scrape --help

# See options for analyze command
python -m repo_stats analyze --help

Analyze command options

The analyze command provides various options for visualizing and summarizing repository data:

• User summary table: Use --users_summary+ "merged PRs" --users_summary+ "commented PRs" --users_summary+ "opened issues" --users_summary+ "commented issues" where the first column is used for sorting rows with users:

  Example command:

  repostat analyze Borda/pyRepoStats --users_summary+ "merged PRs" --users_summary+ "opened issues"
  

  user	merged PRs	commented PRs	opened issues	commented issues
  williamFalcon	74	21	14	8
  Borda	42	35	4	18
  akihironitta	17	1	5	5
  ananyahjha93	14	2	6	21
  annikabrundyn	12	0	0	2
  djbyrne	11	2	4	4
  nateraw	9	1	6	8
  teddykoker	3	2	0	0

• Contribution filter: Use --min_contribution N to filter users by minimum number of contributions:

  repostat analyze Borda/pyRepoStats --min_contribution 10 --users_summary+ "all"
  

• Time frame filtering: Define a time frame with --date_from and --date_to for filtering events - created issues, merged PRs and comments/reviews:

  repostat analyze Borda/pyRepoStats --date_from "2023-01-01" --date_to "2023-12-31"
  

• Contribution aggregation over time: Use --user_comments+ with time granularity (D=Day, W=Week, M=Month, Y=Year) to visualize contribution patterns:

  # Weekly aggregation
  repostat analyze Borda/pyRepoStats --user_comments+ W
  
  # Weekly aggregation for issues and PRs separately
  repostat analyze Borda/pyRepoStats --user_comments+ W --user_comments+ issue --user_comments+ pr
  
  # Multiple time sampling (weekly and monthly)
  repostat analyze Borda/pyRepoStats --user_comments+ W --user_comments+ M
  

  This draws double charts: (a) cumulative aggregation over all users and (b) heatmap-like image with time on Y-axis and users on X-axis.

  

To deny showing figures set environment variable export SHOW_FIGURES=0.

Contribution

Any help or suggestions are welcome, pls use Issues :]