GH Metrics

Analyze some metrics around a window of time from a github repo, as i want to track Pull requests as a means of measuring team and project harmony.

Looking at

number of PRs time they sat open for pending review number of distinct contributors (using labels) what kind of PR and how many were external

Current

• using python 3.9
• requirements.txt includes
  • yaml
  • pandas
  • PyGithub

Usage

Ensure you've exported a github access token to GITHUB_TOKEN

Run the script using Python and pass the required and optional arguments as flags:

python script.py --config CONFIG_PATH [--start-date START_DATE] [--end-date END_DATE] [--absolute]

Flags:
    --config CONFIG_PATH: Path to the YAML configuration file. (optional)
    --start-date START_DATE: The start date for the analysis in the "YYYY-MM-DD" format. (optional, default is 2 weeks ago from the current date)
    --end-date END_DATE: The end date for the analysis in the "YYYY-MM-DD" format. (optional, default is the current date)
    --absolute (presence): generate statistics for entire history of repository

Example:

Analyze PRs from the past two weeks (using defaults for start and end date):

    python script.py --config config.yaml

Analyze PRs from specific dates:

    python script.py --config config.yaml --start-date 2023-01-01 --end-date 2023-01-15

Using the --absolute Flag

The --absolute flag allows you to control whether the statistics should be generated in absolute mode or for the entire history of the repository.

Output

The script outputs JSON data with the following metrics:

average_pr_open_time: Average time a PR was open
counts_per_day: PRs opened each day
open: Remaining OPEN PRs from period
pr_count: Total number of PRs.
distinct_contributors: Number of unique contributors.
label_distribution: Distribution of labels across PRs.
average_pr_open_time: Average time PRs stay open.
pr_open_time_histogram: A histogram of PR open times, categorized into predefined bins.

Security Note

Always keep your GitHub personal access tokens secure and never embed them directly in code. Use environment variables or secure vaults to handle sensitive data, especially in production environments.