add scripts to output usage stats 📈

This commit is contained in:
platane
2023-01-09 08:07:43 +01:00
parent e2eb91cf8f
commit d9d2fa1b52
6 changed files with 757 additions and 3 deletions

2
packages/usage-stats/.gitignore vendored Normal file
View File

@@ -0,0 +1,2 @@
.env
cache

View File

@@ -0,0 +1,79 @@
import * as fs from "fs";
import fetch from "node-fetch";
import { load as CheerioLoad } from "cheerio";
const getPackages = async (repo: string) => {
const pageText = await fetch(
`https://github.com/${repo}/network/dependents`
).then((res) => res.text());
const $ = CheerioLoad(pageText);
return $("#dependents .select-menu-list a")
.toArray()
.map((el) => {
const name = $(el).text().trim();
const href = $(el).attr("href");
const u = new URL(href!, "http://example.com");
return { name, id: u.searchParams.get("package_id")! };
});
};
const getDependentByPackage = async (repo: string, packageId: string) => {
const repos = [] as string[];
const pages = [];
let url:
| string
| null = `https://github.com/${repo}/network/dependents?package_id=${packageId}`;
while (url) {
console.log(url, repos.length);
await wait(1000 + Math.floor(Math.random() * 500));
const $ = CheerioLoad(await fetch(url).then((res) => res.text()));
const rs = $(`#dependents [data-hovercard-type="repository"]`)
.toArray()
.map((el) => $(el).attr("href")!.slice(1));
repos.push(...rs);
const nextButton = $(`#dependents a`)
.filter((_, el) => $(el).text().trim().toLowerCase() === "next")
.eq(0);
const href = nextButton ? nextButton.attr("href") : null;
pages.push({ url, rs, next: href });
fs.writeFileSync(
__dirname + `/out-${packageId}.json`,
JSON.stringify(pages)
);
url = href ? new URL(href, "https://github.com").toString() : null;
}
return repos;
};
export const getDependents = async (repo: string) => {
const packages = await getPackages(repo);
const ps: (typeof packages[number] & { dependents: string[] })[] = [];
for (const p of packages)
ps.push({ ...p, dependents: await getDependentByPackage(repo, p.id) });
return ps;
};
const wait = (delay = 0) => new Promise((r) => setTimeout(r, delay));
(async () => {
const res = await getDependents("platane/snk");
fs.writeFileSync(__dirname + "/cache/out.json", JSON.stringify(res));
})();

View File

@@ -0,0 +1,125 @@
import * as fs from "fs";
import fetch from "node-fetch";
import { Octokit } from "octokit";
require("dotenv").config();
// @ts-ignore
import packages from "./out.json";
const octokit = new Octokit({ auth: process.env.GITHUB_TOKEN });
const getLastRunInfo = async (repo_: string) => {
const [owner, repo] = repo_.split("/");
try {
const {
data: { workflow_runs },
} = await octokit.request(
"GET /repos/{owner}/{repo}/actions/runs{?actor,branch,event,status,per_page,page,created,exclude_pull_requests,check_suite_id,head_sha}",
{ owner, repo }
);
for (const r of workflow_runs) {
const { run_started_at, head_sha, path, conclusion } = r as {
run_started_at: string;
head_sha: string;
path: string;
conclusion: "failure" | "success";
};
const workflow_url = `https://raw.githubusercontent.com/${owner}/${repo}/${head_sha}/${path}`;
const workflow_file = await fetch(workflow_url).then((res) => res.text());
const [_, dependency, __, version] =
workflow_file.match(/uses\s*:\s*(Platane\/snk(\/svg-only)?@(\w*))/) ??
[];
const cronMatch = workflow_file.match(/cron\s*:([^\n]*)/);
if (dependency)
return {
dependency,
version,
run_started_at,
conclusion,
cron: cronMatch?.[1].replace(/["|']/g, "").trim(),
workflow_file,
workflow_url,
};
}
} catch (err) {
console.error(err);
}
};
const wait = (delay = 0) => new Promise((r) => setTimeout(r, delay));
const getRepos = () => {
try {
return JSON.parse(fs.readFileSync(__dirname + "/cache/out.json").toString())
.map((p: any) => p.dependents)
.flat() as string[];
} catch (err) {
return [];
}
};
const getReposInfo = () => {
try {
return JSON.parse(
fs.readFileSync(__dirname + "/cache/stats.json").toString()
) as any[];
} catch (err) {
return [];
}
};
const saveRepoInfo = (rr: any[]) => {
fs.writeFileSync(__dirname + "/cache/stats.json", JSON.stringify(rr));
};
(async () => {
const repos = getRepos();
const total = repos.length;
const reposInfo = getReposInfo().slice(0, -20);
for (const { repo } of reposInfo) {
const i = repos.indexOf(repo);
if (i >= 0) repos.splice(i, 1);
}
while (repos.length) {
const {
data: { rate },
} = await octokit.request("GET /rate_limit", {});
console.log(rate);
if (rate.remaining < 100) {
const delay = rate.reset - Math.floor(Date.now() / 1000);
console.log(
`waiting ${delay} second (${(delay / 60).toFixed(
1
)} minutes) for reset `
);
await wait(Math.max(0, delay) * 1000);
}
const rs = repos.splice(0, 20);
await Promise.all(
rs.map(async (repo) => {
reposInfo.push({ repo, ...(await getLastRunInfo(repo)) });
saveRepoInfo(reposInfo);
console.log(
reposInfo.length.toString().padStart(5, " "),
"/",
total,
repo
);
})
);
}
})();

View File

@@ -0,0 +1,15 @@
{
"name": "@snk/usage-stats",
"version": "1.0.0",
"dependencies": {},
"devDependencies": {
"sucrase": "3.29.0",
"cheerio": "1.0.0-rc.12",
"node-fetch": "2.6.7",
"octokit": "2.0.11",
"dotenv": "16.0.3"
},
"scripts": {
"start": "sucrase-node stats.ts"
}
}

View File

@@ -0,0 +1,59 @@
import * as fs from "fs";
type R = { repo: string } & Partial<{
dependency: string;
version: string;
run_started_at: string;
conclusion: "failure" | "success";
cron?: string;
workflow_file: string;
}>;
(async () => {
const repos: R[] = JSON.parse(
fs.readFileSync(__dirname + "/cache/stats.json").toString()
);
const total = repos.length;
const recent_repos = repos.filter(
(r) =>
new Date(r.run_started_at!).getTime() >
Date.now() - 7 * 24 * 60 * 60 * 1000
);
const recent_successful_repos = recent_repos.filter(
(r) => r?.conclusion === "success"
);
const versions = new Map();
for (const { dependency } of recent_successful_repos) {
versions.set(dependency, (versions.get(dependency) ?? 0) + 1);
}
console.log(`total ${total}`);
console.log(
`recent_repos ${recent_repos.length} (${(
(recent_repos.length / total) *
100
).toFixed(2)}%)`
);
console.log(
`recent_successful_repos ${recent_successful_repos.length} (${(
(recent_successful_repos.length / total) *
100
).toFixed(2)}%)`
);
console.log("versions");
for (const [name, count] of Array.from(versions.entries()).sort(
([, a], [, b]) => b - a
))
console.log(
`${(name as string).split("Platane/")[1].padEnd(20, " ")} ${(
(count / recent_successful_repos.length) *
100
)
.toFixed(2)
.padStart(6, " ")}% ${count} `
);
})();