From 67cf4d0c0416bcf67e076648aec8dd14141b3dc9 Mon Sep 17 00:00:00 2001
From: Matthew Reishus <mreishus@users.noreply.github.com>
Date: Sun, 31 May 2020 20:35:20 -0500
Subject: [PATCH] my.coding.github ignores some events emitted by bots.

I use a service called dependabot ( https://dependabot.com/ ).  It
automatically creates pull requests in my repositories to upgrade
dependencies.  The modern front end javascript world moves really
quickly; projects have a ton of dependencies that are updating all the
time, so there are a lot of these pull requests.

Also, the PRs it makes have a lot of info in them.  Here's an example
one: https://github.com/mreishus/spades/pull/180 .  If you hit the
arrows, you can see it includes a lot of text in "Changelog" and
"Commits".  Now check out the list of closed PRs this project has:
https://github.com/mreishus/spades/pulls?q=is%3Apr+is%3Aclosed

Once I got everything working with my.coding.github, my Github.org
(using orger) was huge: 5MB.  I wanted to get rid of the dependabot
stuff, since it's mostly junk I'm not too interested it, and I got it
down to 130K (from 5MB) just from this commit.

Here's an example of an event I'm filtering out:
I'm looking to see if the "user" contains a "[bot]" tag in it.

  {
    "type": "pull_request",
    "url": "https://github.com/mreishus/spades/pull/96",
    "user": "https://github.com/dependabot-preview[bot]",
    "repository": "https://github.com/mreishus/spades",
    "title": "Bump axios from 0.19.1 to 0.19.2 in /frontend",
    "body": "Bumps [axios](https://github.com/axios/axios) from 0.19.1 to 0.19.2.\n<details>\n<summary>Release notes</summary [cut 5000 characters]
    "base": {
      "ref": "master",
      "sha": "a47687762887c6e5c0d5d0a38c3c9697f09cbcd6",
      "user": "https://github.com/mreishus",
      "repo": "https://github.com/mreishus/spades"
    },
    "head": {
      "ref": "dependabot/npm_and_yarn/frontend/axios-0.19.2",
      "sha": "0e79d0220002cb54cd40e13a40addcc0d0a01482",
      "user": "https://github.com/mreishus",
      "repo": "https://github.com/mreishus/spades"
    },
    "assignee": "https://github.com/mreishus",
    "assignees": [
      "https://github.com/mreishus"
    ],
    "milestone": null,
    "labels": [
      "https://github.com/mreishus/spades/labels/dependencies",
      "https://github.com/mreishus/spades/labels/javascript"
    ],
    "review_requests": [

    ],
    "work_in_progress": false,
    "merged_at": null,
    "closed_at": "2020-01-25T14:40:27Z",
    "created_at": "2020-01-22T13:37:17Z"
  },

Maybe this should be a config option, but I didn't know how to make them
cleanly in HPI, and I'm not sure if anyone would ever want this stuff.
---
 my/coding/github.py | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/my/coding/github.py b/my/coding/github.py
index 508801f..3f5dd63 100644
--- a/my/coding/github.py
+++ b/my/coding/github.py
@@ -25,6 +25,7 @@ class Event(NamedTuple):
     eid: str
     link: Optional[str]
     body: Optional[str]=None
+    is_bot: bool = False
 
 
 # TODO hmm. need some sort of abstract syntax for this...
@@ -129,32 +130,38 @@ def _parse_repository(d: Dict) -> Event:
 
 def _parse_issue_comment(d: Dict) -> Event:
     url = d['url']
+    is_bot = "[bot]" in d["user"]
     return Event( # type: ignore[misc]
         **_parse_common(d),
         summary=f'commented on issue {url}',
         eid='issue_comment_' + url,
+        is_bot=is_bot,
     )
 
 
 def _parse_issue(d: Dict) -> Event:
     url = d['url']
     title = d['title']
+    is_bot = "[bot]" in d["user"]
     return Event( # type: ignore[misc]
         **_parse_common(d),
         summary=f'opened issue {title}',
         eid='issue_comment_' + url,
+        is_bot=is_bot,
     )
 
 
 def _parse_pull_request(d: Dict) -> Event:
     url = d['url']
     title = d['title']
+    is_bot = "[bot]" in d["user"]
     return Event( # type: ignore[misc]
         **_parse_common(d),
         # TODO distinguish incoming/outgoing?
         # TODO action? opened/closed??
         summary=f'opened PR {title}',
         eid='pull_request_' + url,
+        is_bot=is_bot,
     )
 
 
@@ -245,6 +252,8 @@ def iter_events() -> Iterator[Res[Event]]:
         if isinstance(e, Exception):
             yield e
             continue
+        if e.is_bot:
+            continue
         key = (e.dt, e.eid) # use both just in case
         # TODO wtf?? some minor (e.g. 1 sec) discrepancies (e.g. create repository events)
         if key in emitted: