Git on mkz.me

Removing files from repository

Wed, 23 Apr 2025 08:44:05 +0200

There is a nice command to help rewriting whole repository without much hassle: ‘git-filter-repo’. In this use case, I used it to remove some directory files entirely (so I could push the repository to github)

$ git filter-repo --force --filename-callback '
  if b"fake_data/" in filename:
    return None
  else:
    return filename
  '

If for each entry the name is None, nothing will be kept in the history as all commits with the filename will be rewritten, removing the file in the objects of the commit. If commit becomes empty, it is removed from the history.

Retrieve Github contributions with GraphQL

Wed, 25 Sep 2024 22:26:12 +0200

The Github API provides an interface to retrieve a lot of informations about your account and especially daily contributions. This requires a simple github token (with user:read permissions)

Retrieve with a simple shell/curl query Link to heading

$ curl -H "Authorization: bearer <your token>" -X POST -d '{"query": "query { user(login: \"mycroft\") { contributionsCollection { contributionCalendar { totalContributions } } } }"}' https://api.github.com/graphql
{"data":{"user":{"contributionsCollection":{"contributionCalendar":{"totalContributions":1088}}}}}

Complete Python code Link to heading

import requests
from datetime import datetime, timedelta
from collections import defaultdict

def get_contributions(username, start_date, end_date, access_token):
    query = """
    query ($username: String!, $from: DateTime!, $to: DateTime!) {
      user(login: $username) {
        contributionsCollection(from: $from, to: $to) {
          contributionCalendar {
            weeks {
              contributionDays {
                date
                contributionCount
              }
            }
          }
        }
      }
    }
    """

    variables = {
        "username": username,
        "from": start_date.isoformat(),
        "to": end_date.isoformat()
    }

    headers = {
        "Authorization": f"Bearer {access_token}",
        "Content-Type": "application/json",
    }

    response = requests.post('https://api.github.com/graphql', 
                             json={'query': query, 'variables': variables}, 
                             headers=headers)

    if response.status_code == 200:
        return response.json()
    else:
        raise Exception(f"Query failed with status code: {response.status_code}. {response.text}")

username = 'mycroft'
access_token = '<your token>'
end_date = datetime.now()
start_date = end_date - timedelta(days=365)

try:
    data = get_contributions(username, start_date, end_date, access_token)
    monthly_contributions = group_contributions_by_month(data)
    
    for month, count in monthly_contributions.items():
        print(f"{month}: {count} contrib.")
except Exception as e:
    print(f"An error occurred: {e}")

This will output a simple dictionnary to parse to retrieve daily/monthly/yearly contributions.

Reading the object tree

Tue, 24 Sep 2024 13:25:59 +0200

git cat-file allows opening and reading commits/tree/blogs without any hassle. Starting with a commit hash, it is possible to find out what’s there is in the given commit, to which tree it is attached, parent, etc.

$ git rev-parse HEAD
39f40cf0d0b31b16e235f07853d74167c72c1d8b

$ git cat-file -p 39f40cf0d0b31b16e235f07853d74167c72c1d8b
tree 5ee8a97718dd4662373ea5e722b008f03b7d229e
parent 79b04c1d8b3f275c37a6f4bd0b42d35ccaf42505
author Patrick Marie <pm@mkz.me> 1726849801 +0200
committer Patrick Marie <pm@mkz.me> 1726849801 +0200

Updating scylla/scylla-operator from v1.13.0 to v1.14.0

$ git cat-file -p 5ee8a97718dd4662373ea5e722b008f03b7d229e
100644 blob e2b5b73f808c496be73f4e399090a6ae63818bf2	.drone.yml
100644 blob 55e7b6e6698a2a415ef1f4d5df92a8ec77817cc9	.gitignore
100644 blob 20f1ead4b88ab9abadb951cf2a226aecd7aa633b	.golangci.yaml
100644 blob bcc630c55a74759a730809306bdedcae93a5c68c	.justfile
...

Reading a blob is similar:

Setting a replication between gitea and github

Sat, 21 Sep 2024 12:34:00 +0200

You’ll need some github credentials (login and api key) with proper authorization.

Setting up a token in github Link to heading

In github’s tokens/settings, create a classic token with “public_repo” access and a meaningful name, such as “gitea to github replication”.

Also, create an empty repository that will receive the replication. Note the http url.

Set up the replication Link to heading

In your gitea instance, in an existing repository settings, in the Mirror settings section, set the Git Remote Repository URL to the github repository url, set username to your github’s username and use the token (prefixed with ghp_) as a password.