ignore-bot

Github API Bot which makes PRs to Ignore Common Files

Well, dang it

the GitHub Search API provides up to 1,000 results for each search.

Meaning, of the 500,000 results, we can see at most 3,000 based on the different sorting methods.

Documentation

Converting cURL to Ruby
Example Query for .DS_Store
API Docs for Searching
API Docs for Search Query Syntax
API Docs for Search Query on Code Syntax

Install

Required Files:

• dbname.token
• dbuser.token
• dbpass.token
• tablename.token
• oauth.token

MySQL

Database: ignore_bot

Table: repos

• id INT NOT NULL PRIMARY KEY AUTO_INCREMENT
• full_name CHAR(128) NOT NULL
• file_name CHAR(64) NOT NULL
• status CHAR(64) NOT NULL : fork, edit, pr, delete, done

Functionality

Authentication

• Save user's OAuth token
• Add to Organization (as bot)
• Accept Invite (as user)

Hourly

• Search Repos with Bad Files (as user)
• Fork on Github to Org (as user)
• Clone, Edit Git Ignore, Push
• Create Pull Request (as user)
• Delete Fork (as user)

Numbers

500,000 Repos with .DS_Store files

API Requests

List 100 per Requests: 5,000 Requests to list all
Fork Repos: 500,000 Requests
Edit Files: 500,000 Requests
Create PRs: 500,000 Requests
Delete Forks: 500,000 Requests
Total Requests: ~2,000,000

With 100 Users @ 5000 Requests Each (500,000 Requests/Hour) it would take 4 Hours
With 1 User @ 5000 Requests it would take 17 Days

Network Traffic

List Repos (per 100): 100 B up, 500 KB down = 500 KB 5,000 = 2500 MB
Fork Repos: 100 B 500,000 = 50 MB
Edit Files: 100 B 500,000 = 50 MB
Create PRs: 100 B 500,000 = 50 MB
Delete Forks: 100 B * 500,000 = 50 MB
Total Traffic: 2700 MB

With a network speed of 1 MB/s, it would take 45 Minutes

100 B comes from approximation of each request sent.
5 MB from each request comes from counting 1 result as 5 KB times 100 results per page