I’ve updated the esqueleto bindings for esqueleto-textsearch to include a tutorial and documentation so it no longer requires guesswork. Furthermore I’ve also created new esqueleto bindings for PostGIS.
Esqueleto is a more advanced query library that builds on top of the persistent ORM3. Postgres text search brings Elasticsearch like functionality to Postgres. PostGIS is a spatial database extension for Postgres. This allows querying on longitude and latitude for example, or any geometry. As part of postgres it allows mixing normal relational data with space and geometry, which makes space itself relational!
Roughly four years ago I had drafted out an implementation for the admin pages for some company. This was part of my “trial” 2 days, where you just work as a “technical interview”. 1 Ironically the admin pages ended up having far better search than the main app, unfortunately, implementing proper text-search to the main app was never prioritized. A colleague of mine had ported that to the current library. Recently I needed this again. However, I completely had forgotten how it worked and there was essentially no documentation. 2 Since I was somehow the maintainer of this package I decided to just fix this for all Haskellers.
PostGIS has a different motivation. My current contract is about finding stuff. Now I could do this with some geometry library in pure haskell. However, considering most logic is already in the database, I decided it was worth a try to see if PostGIS would be a good fit. There was already some existing art done, such as reading from the database, but there was no off-the-shelf, functional writing. I initially handcrafted some haphazard persistent instances to see if a full solution was feasible. It worked, and one interesting outcome is that I could combine this PostGIS implementation with text-search! For example, we’ve named locations on a map, and objects within the database have some point on the map, we can now search for some location on the map and find the objects through normal full-text search. The database does all the thinking. Because this worked so well, I got motivated to do the full bindings for PostGIS, and prevent the issue I just had with text-search.
I’m slowly transforming into a database engineer š . Anyway, everyone should try these libraries out! You’re in space after all, and you lose stuff all the time, go get searching in space! Let me know if you’ve any comments or suggestions. Furthermore, let me know if you need help with database woes, I’m becoming quite good at this.
I perform way better in such a situation than a take-home test, because it’s incredibly hard for me to get motivated to do throw away work, but here I did something useful. ↩
Note that the hackage upload was also done by me, because my dear colleague had ported it to a library which wasn’t even published! It was just some github repository. After a couple years I decided to just upload it, right before I jumped ship to another company. ↩
Object relational mapping ↩