oseqorg / api

Audiopedia API

View the docs here: https://audiopedia-t4sg.github.io/api-docs/

---

Local installation instructions:

• Use Python 3.6.5
  • Recommended: use virtualenv -p ~/.pyenv/shims/python3.6 ~/.virtualenvs/api to create the virtual environment (must install virtualenv and pyenv and run pyenv install -v 3.6.5 for this to work)
  • source ~/.virtualenvs/api/bin/activate to activate the virtual environment
  • deactivate to deactivate
• Add secure.py (not included in GitHub repo) to audiopedia/audiopedia directory
• From the audiopedia/ directory run pip install -r requirements.txt
• From the audiopedia/ directory run python manage.py migrate

To run the project locally, run python manage.py runserver and navigate to http://localhost/graphql to interact with the GraphiQL explorer. See documentation for example queries/mutations.

api's Issues

Authentication

Creating, updating, and deleting data should have restricted access so that only authorized (authenticated) users — i.e. admins like Marcel — can access those mutations.

• These Graphene docs might help
• Update documentation appropriately

Basic setup for Graphene-Django

• Read Onboarding Meeting Notes. In particular, look through and the Gatsby docs to see how this would be used by the PWA and if this approach would be feasible.
• Read this introduction and follow this tutorial to implement a basic GraphQL server.
  • No need for really accurate/fleshed out models, but use relevant model/field names

Basic setup for Django REST Framework

• Read Onboarding Meeting Notes. In particular, look through this repo and the Gatsby docs to see if this approach would be feasible.
• Follow this tutorial (disregard the React part) or just use the Django REST Framework docs to set up a basic RESTful API (get the DRF-generated UI running!).
  • No need for really accurate/fleshed out models, but use relevant model/field names

scraped CSV to model instances

• Clean CSV of scraped audiopedia.wiki data
• Import data from CSV to convert into model instances (see this stackoverflow discussion
  • Note: I would recommend working with a small test CSV first

Fix foreign key issue

I’m not sure if this is just lacking in the documentation or if we’ve had different understandings of how we’d implement the mutations for creating tracks/playlists/topics, but at the moment it doesn’t seem like there’s a way to create a track and attach it to a given playlist, create a playlist and attach it to a given topic, or create a topic and attach it to a given language — this is important and would be the whole reason why we have implemented these as foreign keys. From what I understood, the inclusion of an array of tracks in a playlist mutation is not for the creation of said tracks or for their attachment to the new playlist (this would be incredibly inefficient) — this is solely for ordering the tracks already attached to the playlist (therefore only relevant to updating as opposed to creating a playlist).

Deployment

• Read up on Zappa
• Watch this video

Moar helpful resources:

• Zappa and Django
• using Zappa with CloudFront, RDS and VPC — we may not use all of these recommendations but still good to look into
• Regarding this, see these Graphene API specific notes

Documentation

• Add descriptions to schema (for viewing on GraphiQL).
• If there's time, start a CONTRIBUTING.md file for this repo: write up instructions for setting up one's environment and running tests.

Design UI for content management (MVP)

UPDATE: Figma sketches here

@BachToTheFuture @benwubenwu @paulazoo @tp0119 please add feedback in the comments (on this issue or on this Notion doc)

---

Browse audiopedia.app and audiopedia.io, as well as the old codebase. Design (hand-drawn sketch is fine) the UI for content management:

• uploading playlists as zipped folders of tracks (mp3s)
• updating tracks: adding transcripts
• grouping playlists by topic
• changing order of playlists within topic
• changing order of playlist topics
• updating playlists: adding/removing tracks, changing order of tracks within playlist

Implement models + basic queries

Implement models as specified in #4 (note revisions in this comment), and update schema to include queries for
- [ ] fetching playlist by id
- [ ] fetching topic by id
- [ ] fetching tracks by playlist
- [ ] fetching playlists by topic

Design database

Browse audiopedia.app and audiopedia.io, as well as the old codebase. Sketch a diagram of the models/relationships you think we'll need (refer to this tutorial for information on database structure and design notation, but no need to use the tool — hand-drawn is fine).

Automated Testing

• Create tests for queries/mutations described in #9 and #12
• Refer to "Snapshot testing" in the Graphene-Django docs

Implement basic mutations

After #9 is done, implement mutations for
- [ ] creating/updating a topic (title, audio url)
- [ ] creating/updating a playlist (title, audio url, topic)
- [ ] bulk creating tracks (title, audio url, transcript optional, playlist)
- [ ] updating a track by id
- [ ] deleting a topic/playlist/track
- [ ] "removing"/archiving a topic/playlist/track

Language not accounted for in these specs; if you are confident that having a Language model is sufficient/good practice you can implement that and the mutations accordingly, otherwise we can add that in a future migration.

Create a new issue for other mutations you think will be necessary.