Consider only counting unique commits

[mathbunnyru]

Describe the bug

I maintain a repository https://github.com/jupyter/docker-stacks/, which has 848 my commits.
Unfortunately, some users, instead of forking this repository, create their own copy.
This makes github-readme-stats count commits in this repos as well.

https://github-readme-stats-eight-theta.vercel.app/api?username=mathbunnyru&show_icons=true&theme=solarized-light&include_all_commits=true

This chart shows 68.6k commits, which would be awesome, but it's approximately 40x-80x larger than the correct number.
If I search my commits I see all these duplicates: https://github.com/search?o=desc&s=committer-date&type=Commits&q=author%3Amathbunnyru&p=94

Expected behaviour

I expect github-readme-stats to parse the commits and only list the number of unique SHAs.
This way it would show the real number of commits.

Screenshots / Live demo link

Additional context

No response

[rickstaa]

@mathbunnyru thanks for your bug report. I could reproduce the bug below. It is caused by the fact that we use the GitHub rest API for the total commits.

github-readme-stats/src/fetchers/stats-fetcher.js

Lines 146 to 170 in fd64333

	const fetchTotalCommits = (variables, token) => {
	return axios({
	method: "get",
	url: `https://api.github.com/search/commits?q=author:${variables.login}`,
	headers: {
	"Content-Type": "application/json",
	Accept: "application/vnd.github.cloak-preview",
	Authorization: `token ${token}`,
	},
	});
	};
	try {
	let res = await retryer(fetchTotalCommits, { login: username });
	let total_count = res.data.total_count;
	if (!!total_count && !isNaN(total_count)) {
	return res.data.total_count;
	}
	} catch (err) {
	logger.log(err);
	}
	// just return 0 if there is something wrong so that
	// we don't break the whole app
	return 0;
	};

The behaviour is therefore as expected. You could create a pull request to improve the search query we use. We use the GitHub Rest API because of resource limitations in the GraphQL API (see https://docs.github.com/en/graphql/overview/resource-limitations). Fetching all commits through the GraphQL API, although more accurate, will result is in hitting the rate limiter faster. #1691 will fix this, but we cannot merge it into the main branch because of this. You can, however, already use that pull request on your own Vercel instance, since there you will likely not hit the GraphQL limits. We will likely release it under our new GitHub action (see #2179).

Reproduce

[![Mathbunnyru GitHub stats](https://github-readme-stats.vercel.app/api?username=Mathbunnyru)](https://github.com/anuraghazra/github-readme-stats)

[![Mathbunnyru GitHub stats](https://github-readme-stats.vercel.app/api?username=Mathbunnyru&include_all_commits=true)](https://github.com/anuraghazra/github-readme-stats)

[mathbunnyru]

Thank you! I think https://github-readme-stats-rongronggg9.vercel.app works well, and the number of commits is as expected.

[rickstaa]

Thank you! I think https://github-readme-stats-rongronggg9.vercel.app works well, and the number of commits is as expected.

@mathbunnyru please don't use @Rongronggg9 Vercel instance as you will likely cause the GraphQL resource limits to be reached in the future when people follow your behaviour 😅(see https://docs.github.com/en/graphql/overview/resource-limitations). I would like to recommend you to create your own Vercel instance and merge #1691 into your main branch 🙏🏻(see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own).

[mathbunnyru]

Ok, sorry @rickstaa, I will do what you suggest 😅

[rickstaa]

Ok, sorry @rickstaa, I will do what you suggest 😅

No problem, the GraphQL limit is a real bottleneck and not well explained in the README. That's why we want to switch to a GitHub action approach (see #2179).

[mathbunnyru]

@rickstaa I don't think this is fixed yet.

[rickstaa]

@rickstaa I don't think this is fixed yet.

You're right, I have mistaken it with another issue related to the language results 👍🏻.

[rickstaa]

You can also show your support for https://github.com/orgs/community/discussions/35675 which would allow us to solve this for the Vercel version.