prfiesta 🦜🥳

Collect and Analyze Individual Contributor Pull Requests

prfiesta allows you to collect, analyze and celebrate pull requests made by an individual 🎉.

It can be used by engineers or managers to gain insights into all the great work the contributor has made over a specified period of time. A great use case of this tool is during a performance review process when you want to perform analysis on all the contributions made over the year.

Install

python -m pip install prfiesta

Dependencies:

• Python 3.8+

Usage

# Authenticate yourself
export GITHUB_TOKEN=... # or GITHUB_ENTERPRISE_TOKEN

# Get all pull requests for a user
prfiesta -u kiran94

# Get all pull requests for a user created after a date
prfiesta -u kiran94 --after 2023-01-01

# Get all pull requests for a user created between two dates
prfiesta -u kiran94 --after 2023-01-01 --before 2023-06-01

# Get all pull requests for a user updated after a date
prfiesta -u kiran94 --after 2023-01-01 --use-updated

# Get all pull requests with a custom output file name
prfiesta -u kiran94 --output my_pull_requests.csv

# Get all pull requests in parquet format with a custom file name
prfiesta -u kiran94 --output-type parquet --output my_pull_requests.parquet

# Get all pull requests and export to a duckdb database
prfiesta -u kiran94 --output-type duckdb --output mydatabase.duckdb

# Get all pull requests for more then one user
prfiesta -u kiran94 -u user2

# Get all pull requests and drop specific columns from the output
prfiesta -u kiran94 -dc events_url -dc comments_url -dc node_id

# Get all pull requests where the user was involved (as opposed to just authored)
prfiesta -u kiran94 --use-involves

# Get all pull requests where the user reviewed it rather then being the author
prfiesta -u charliermarsh --after 2023-05-01 --use-reviewed-by

# Get all pull requests where the user was requested a review rather then being the author
prfiesta -u charliermarsh --after 2023-05-01 --use-review-requested

# Get help
prfiesta --help

# Show the current version
prfiesta --version

You can also leverage prfiesta directly in your own application:

import pandas as pd

from datetime import datetime
from prfiesta.collectors.github import GitHubCollector

github = GitHubCollector()
frame: pd.DataFrame = github.collect('kiran94', 'user2', after=datetime(2023, 1, 1))

print(frame)

Output

You can control the output type using the --output-type option. Supported options:

• csv (default)
• parquet
• duckdb

You can also customize the output file name using the --output option. When using duckdb, this argument is the duckdb database that we should export into. You can see an example of a duckdb workflow here.

User Filter

By default, prfiesta will take the users provided in the --user option and search the Git provider for any pull requests that the user authored. Within more collaborative environments, this may not be what you want as you may want to also gain some visibility into all secondary contributions a user made (e.g commenting on others pull requests).

The options listed here are mutually exclusive.

User Involvement

prfiesta exposes the --use-involves flag which will search for pull requests that were:

• Created by a certain user
• Assigned to that user
• Mention that user
• commented on by that user

Learn more about involves here.

User Reviewed

prfiesta exposes a --use-reviewed-by flag which will collect pull requests where the user has reviewed other's pull requests.

Learn more about searching review requests here

User Requested Review

prfiesta exposes a --use-review-requested flag which will collect pull requests where the user was requested a review from other collaborators.

Learn more about searching review requests here

Date Filter

When using the --after and --before date filters, by default prfiesta will use the created date dimension with these filters on the Git provider (e.g GitHub). This may not fit your use case and you may want to filter on when a pull request was updated instead. To do this you can use the --use-updated flag.

Learn more about date filters here.

Analysis

prfiesta ships with built in plots to help analyze your pull request data. These serve as a starting point in your analysis. See more information on the build in plots and views here.

Using GitHub Enterprise

If you trying to fetch data from a GitHub Enterprise server, then much of the same functionality should work the same. You just need to make sure that:

• GH_HOST is set to your enterprise instance's API URL. Reach out to your internal GitHub team if you are not sure what this should be.
• GITHUB_ENTERPRISE_TOKEN a personal access token generated on your GitHub Enterprise instance.

GitHub Rate Limiting

Depending on your input parameters, you may end up in a situation where you are being Rate Limited by the GitHub API.

See this Notebook on a way to handle this.

Environment Variables

Variable	Description	Default
GITHUB_TOKEN	The Github Token to use
GITHUB_ENTERPRISE_TOKEN	Takes precedence over GITHUB_TOKEN when set
GH_HOST	The Github Host to communicate with (Override this with your company's GitHub Enterprise server if needed)	https://api.github.com
LOGGING_LEVEL	The Logging Level to use	INFO
LOGGING_FORMAT	The Logging Format to use	%(message)s
SPINNER_STYLE	The Spinner Style to use	blue

Developer Setup

Assuming you have cloned the repository and are at the root of the repository in your terminal.

poetry shell
poetry install
poetry build

This should leave you in a state where you have the virtual environment sourced, all dependencies are installed and prfiesta is installed locally.

You can then leverage the various commands in the makefile for development tasks:

# Run all unit tests
make test

# Produce code coverage reports
make coverage

# Code linting
make lint

Optionally you can also install pre-commit to run some sanity checks before your commits.

# Install it into your git hooks (one time setup)
# from this point onwards, any commits will run pre-commit checks
precommit_install

# If you want to run all checks on all files without comitting.
precommit_run

Creating Prereleases

When you create a pull request on this repository, various CI checks are run, towards the end of those checks there is a release job.

Usually when running under main, this job is responsible for publishing new versions to pypi. However when running under a pull request, this will create a special prerelease package specific to that pull request.

The versioning of this package follows PEP-440 and will look something like this:

0.8.1b125

Where

• 0.8.1 = The bumped version of what is currently within the pyproject.toml of that pull request. We don't attempt to do any analysis to figure out if we should be bumping with a higher serverity in this context.
• b = Beta; Indicates to pypi that this is a prerelease package.
• 125 = The github.run_number from GitHub Actions.

An example prerelease package looks like this: https://pypi.org/project/prfiesta/0.8.1b125/

Downstream users can then do a full end to end test with the prerelease package before the change is merged into main. This will automatically be posted into the pull request example.

You can find the full version history of package here