Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/oliverswitzer/better_prs


https://github.com/oliverswitzer/better_prs

Last synced: 10 days ago
JSON representation

Awesome Lists containing this project

README 

        
# Experiments with trying to make a world with better pull requests



## Goals of this project



The end goal of this project is to make it more enjoyable to consume other people's PRs by revealing more about what happened while someone was working on their PR.



A working concept for this idea is a chrome extension that reorders the files in a PR not by alphabetic order but instead by when they were first created/updated in the lifecycle of working on the branch. The idea with this is that following the file creation/update path that the developer took will make it easier to understand what was going through the developers header at the time of working on the branch.



# Implementation (so far)



Using fswatch to watch the file current directory for writes to files in the project, taking a diff each time a write happens, and printing that to a file. That file is called history.txt, and is saved in a namespaced folder within `~/.cache/pr_history/*`



This is all happening inside of `demo.sh`



I experimented with using GitHub's Blob API to try to store the file history so that we can later retrieve it when viewing PRs, but ran into some weird encoding issues where when I decode the blob again it has a bunch of strange characters.... it looks like the Blob API supports JSON so maybe it'd be possible to parse the raw diff output into json to avoid these decoding issues... *shrug*



Details aside, the idea here would be to later fetch this blob on the github PR page (via a chrome extension that will be built as apart of this repo and that the user will have installed), decode the history.txt blob, and re-order the files that are displayed on the PR page based on the order in which they were created/modified while working on the branch.



A later implementation might actually use the raw diffs over time to display a high-level overview of what occured on the branch while the PR-creator was working on it. For example:



* Oliver first created a new test file `test/features/foo.exs`

* Oliver then added a test in `test/features/foo.exs`

* Oliver then created a new file `src/foo.exs`

* Oliver then deleted 20 files in `src/images/*`



All of this is an attempt to tell a story about how the code was worked on, rather than just hand you a giant flat file structure to wade through...



## GitHub Blobs API calls



Here were the API calls I was using (using the GH CLI tools):



For creating a new blob for the history.txt file:



```

CONTENT=$(cat ~/.cache/pr_history/%Users%oliverswitzer%workspace%better-prs-demo/history.txt)

ORGANIZATION=oliverswitzer

REPO=better_prs



gh api \

  --method POST \

  -H "Accept: application/vnd.github.v3+json" \

  /repos/$ORGANIZATION/$REPO/git/blobs \

  -f content=$CONTENT

 -f encoding='utf-8'

```

This returns a response with a SHA that you will need to save to later retreive the blob.



For fetching the blob:



```

 gh api \

  -H "Accept: application/vnd.github.v3+json" \

  /repos/oliverswitzer/better_prs/git/blobs/

```



If chrome extension that the person uses has authorization to retrieve blobs on behalf of the user, it can fetch the blob according to some namespace of org/repo/branch, process the history.txt file to display what files were created in what order. 



Later, we can display high level stats about what happened in the lifecycle of working on this branch... "Oliver created a test: test/test.exs". Oliver implemented a test. Oliver created a new file "src/test.exs" etc...