Meilisearch for prefetched courses

[joschahenningsen]

Motivation and Context

The current sql fulltext search is not user friendly. To overcome this issue we can try using meilisearch as discussed here: #814 It even handles typos very well.

Description

The courses are prefetched and stored in a newly introduced search engine instead of the database. From there we can simply query them.

Steps for Testing

Run a local instance of meilisearch like this:

docker run -it --rm \
    -p 7700:7700 \
    -e MEILI_MASTER_KEY='MASTER_KEY'\
    -v $(pwd)/meili_data:/meili_data \
    getmeili/meilisearch:v0.30 \
    meilisearch --env="development"

Download and extract this json document:
prefetchedCourses.zip

Push it to meilisearch as demo data:

curl \                                      
  -H "Authorization: Bearer MASTER_KEY" -X POST 'http://localhost:7700/indexes/PREFETCHED_COURSES/documents?primaryKey=courseID' \
  -H 'Content-Type: application/json' \
  --data-binary @prefetchedCourses.json

Test the search when creating a course

The import can only be tested with a tum online token so you'll need to trust me on this :P

Screenshots

2022-12-02.15-13-22.mp4

[github-actions]

Your Testserver will be ready at https://833.test.live.mm.rbg.tum.de in a few minutes.

[joschahenningsen]

Thanks to @CommanderStorm for the input on this :)

[alexanderstephan]

When run meilisearch and add the json file, I get no results (although all requests are status 200). Do you have any idea what could be wrong?
I get [2022-12-02T20:06:08Z INFO index_scheduler] A batch of tasks was successfully completed., so I assume it is added.

[alexanderstephan]

I think I get the error missing_document_id, which might be related to a broken setup.

[joschahenningsen]

Hm I thought the ?primaryKey might do the trick 🤔
I'll check this out tomorrow!

[joschahenningsen]

@alexanderstephan I updated the zip file, it should work with the new one :)

[alexanderstephan]

Wow, it really works great. 🎉 The code looks also flawless. 👍