Extend awards action (#1937)

* remove debug job, restrict create-pull-request to only awards.txt, add documentation

* make create-pull-request use a custom branch, and filter that out, so PRs generated by action don't invoke action again
This commit is contained in:
Scroogey-SN 2025-03-03 18:47:31 +00:00 committed by GitHub
parent 8a764f0f75
commit 27104302d5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 190 additions and 0 deletions

32
.github/workflows/extend-awards.yml vendored Normal file
View File

@ -0,0 +1,32 @@
name: extend-awards
run-name: Extending awards
on:
pull_request:
types: [ closed ]
branches:
- master
jobs:
if_merged:
if: |
github.event_name == 'pull_request' &&
github.event.action == 'closed' &&
github.event.pull_request.merged == true &&
github.event.pull_request.head.ref != 'extend-awards/patch'
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: '3.13'
- run: pip install requests
- run: python extend-awards.py
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
GITHUB_CONTEXT: ${{ toJson(github) }}
- uses: peter-evans/create-pull-request@v7
with:
add-paths: awards.csv
branch: extend-awards/patch
commit-message: Extending awards.csv
title: Extending awards.csv
body: A PR was merged that solves an issue and awards.csv should be extended.

54
docs/dev/extend-awards.md Normal file
View File

@ -0,0 +1,54 @@
# Automatically extend awards.csv
## Overview
Whenever a pull request (PR) is merged in the [stacker.news](https://github.com/stackernews/stacker.news) repository, a [GitHub Action](https://docs.github.com/en/actions) is triggered:
If the merged PR solves an issue with [award tags](https://github.com/stackernews/stacker.news?tab=readme-ov-file#contributing),
the amounts due to the PR and issue authors are calculated and corresponding lines are added to the [awards.csv](https://github.com/stackernews/stacker.news/blob/master/awards.csv) file,
and a PR is opened for this change.
## Action
The action is defined in [.github/workflows/extend-awards.yml](.github/workflows/extend-awards.yml).
Filters on the event type and parameters ensure the action is [triggered only on merged PRs](https://stackoverflow.com/questions/60710209/trigger-github-actions-only-when-pr-is-merged).
The primary job consists of several steps:
- [checkout](https://github.com/actions/checkout) checks out the repository
- [setup-python](https://github.com/actions/setup-python) installs [Python](https://en.wikipedia.org/wiki/Python_(programming_language))
- [pip](https://en.wikipedia.org/wiki/Pip_%28package_manager%29) installs the [requests](https://docs.python-requests.org/en/latest/index.html) module
- a script (see below) is executed, which appends lines to [awards.csv](awards.csv) if needed
- [create-pull-request](https://github.com/peter-evans/create-pull-request) looks for modified files and creates (or updates) a PR
## Script
The script is [extend-awards.py](extend-awards.py).
The script extracts from the [environment](https://en.wikipedia.org/wiki/Environment_variable) an authentication token needed for the [GitHub REST API](https://docs.github.com/en/rest/about-the-rest-api/about-the-rest-api) and the [context](https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/accessing-contextual-information-about-workflow-runs) containing the event details including the merged PR (formatted in [JSON](https://en.wikipedia.org/wiki/JSON)).
In the merged PR's title and body it searches for the first [GitHub issue URL](https://github.com/stackernews/stacker.news/issues/) or any number with a hash symbol (#) prefix, and takes this as the issue being solved by the PR.
Using the GitHub REST API it fetches the issue and analyzes its tags for difficulty and priority.
It fetches the issue's timeline and counts the number of reviews completed with status 'changes requested' to calculate the amount reduction.
It calculates the amounts due to the PR author and the issue author.
It reads the existing awards.csv file to suppress appending redundant lines (same user, PR, and issue) and fill known receive methods (same user).
Finally, it appends zero, one, or two lines to the awards.csv file.
## Diagnostics
In the GitHub web interface under 'Actions' each invokation of the action can be viewed, including environment and [output and errors](https://en.wikipedia.org/wiki/Standard_streams) of the script. First, the specific invokation is selected, then the job 'if_merged', then the step 'Run python extend-awards.py'. The environment is found by expanding the inner 'Run python extended-awards.py' on the first line.
The normal output includes details about the issue number found, the amount calculation, or the reason for not appending lines.
The error output may include a [Python traceback](https://realpython.com/python-traceback/) which helps to explain the error.
The environment contains in GITHUB_CONTEXT the event details, which may be required to understand the error.
## Security considerations
The create-pull-request step requires [workflow permissions](https://github.com/peter-evans/create-pull-request#workflow-permissions).

104
extend-awards.py Normal file
View File

@ -0,0 +1,104 @@
import json, os, re, requests
difficulties = {'good-first-issue':20000,'easy':100000,'medium':250000,'medium-hard':500000,'hard':1000000}
priorities = {'low':0.5,'medium':1.5,'high':2,'urgent':3}
ignored = ['huumn', 'ekzyis']
fn = 'awards.csv'
sess = requests.Session()
headers = {'Authorization':'Bearer %s' % os.getenv('GITHUB_TOKEN') }
awards = []
def getIssue(n):
url = 'https://api.github.com/repos/stackernews/stacker.news/issues/' + n
r = sess.get(url, headers=headers)
j = json.loads(r.text)
return j
def findIssueInPR(j):
p = re.compile('(#|https://github.com/stackernews/stacker.news/issues/)([0-9]+)')
for m in p.finditer(j['title']):
return m.group(2)
if not 'body' in j or j['body'] is None:
return
for s in j['body'].split('\n'):
for m in p.finditer(s):
return m.group(2)
def addAward(user, kind, pr, issue, difficulty, priority, count, amount):
if amount >= 1000000 and amount % 1000000 == 0:
amount = str(int(amount / 1000000)) + 'm'
elif amount >= 1000 and amount % 1000 == 0:
amount = str(int(amount / 1000)) + 'k'
for a in awards:
if a[0] == user and a[1] == kind and a[2] == pr:
print('found existing entry %s' % a)
if a[8] != amount:
print('warning: amount %s != %s' % (a[8], amount))
return
if count < 1:
count = ''
addr = '???'
for a in awards:
if a[0] == user and a[9] != '???':
addr = a[9]
print('adding %s,%s,%s,%s,%s,%s,%s,,%s,%s,???' % (user, kind, pr, issue, difficulty, priority, count, amount, addr))
with open(fn, 'a') as f:
print('%s,%s,%s,%s,%s,%s,%s,,%s,%s,???' % (user, kind, pr, issue, difficulty, priority, count, amount, addr), file=f)
def countReviews(pr):
url = 'https://api.github.com/repos/stackernews/stacker.news/issues/%s/timeline' % pr
r = sess.get(url, headers=headers)
j = json.loads(r.text)
count = 0
for e in j:
if e['event'] == 'reviewed' and e['state'] == 'changes_requested':
count += 1
return count
def checkPR(i):
pr = str(i['number'])
print('pr %s' % pr)
n = findIssueInPR(i)
if not n:
print('pr %s does not solve an issue' % pr)
return
print('solves issue %s' % n)
j = getIssue(n)
difficulty = ''
amount = 0
priority = ''
multiplier = 1
for l in j['labels']:
for d in difficulties:
if l['name'] == 'difficulty:' + d:
difficulty = d
amount = difficulties[d]
for p in priorities:
if l['name'] == 'priority:' + p:
priority = p
multiplier = priorities[p]
if amount * multiplier <= 0:
print('issue gives no award')
return
count = countReviews(pr)
if count >= 10:
print('too many reviews, no award')
return
if count > 0:
print('%d reviews, %d%% reduction' % (count, count * 10))
award = amount * multiplier * (10 - count) / 10
print('award is %d' % award)
if i['user']['login'] not in ignored:
addAward(i['user']['login'], 'pr', '#' + pr, '#' + n, difficulty, priority, count, award)
if j['user']['login'] not in ignored:
count = 0
addAward(j['user']['login'], 'issue', '#' + pr, '#' + n, difficulty, priority, count, int(award / 10))
with open(fn, 'r') as f:
for s in f:
s = s.split('\n')[0]
awards.append(s.split(','))
j = json.loads(os.getenv('GITHUB_CONTEXT'))
checkPR(j['event']['pull_request'])