Closed dbdimitrov closed 8 months ago
Hi @dbdimitrov, nice! I will check the output once I have some time (probably March though). Did you know that we have a PostgreSQL output adapter? Not sure we already talked about that. In any case, it will probably not be a big deal to convert that to an SQLite adapter, I will see whether I can make this work, or maybe @nilskre can have a go. We can also chat if we find some time on meeting day. :)
Yes, I can give it a go to create an output adapter for Sqlite. So if you have a chat about this, I'd be happy to join you.
Nice! Let's see if we can find some time on Monday.
Hi @slobentanzer @nilskre,
Nice. :) @slobentanzer I remember that you mentioned PostgreSQL as an output adapter, and indeed exporting sqllite3 should be straightforward given you already deal with creating/populating the tables etc.
Maybe we discuss roughly how the schema of metalinksdb looks in my case, and how it's usually via the adapter. See you on Monday.
Hey @EliasFarr @deenes @slobentanzer,
I just wanted to mention that I converted the neo4j database into an sqlite3 database. Main idea being that the latter is much lighter and an inbuilt Python module.
Hence, it's rather straightforward to query via liana: https://github.com/saezlab/liana-py/blob/dev/liana/resource/get_metalinks.py
This is a step in-between an actual solution that avoids the intermediate processing. Perhaps, it could be that using the ontologies from BioCypher we can create a relational DB, or we could have something similar accessible via pypath.
Just FYI and I'm happy to discuss thr steps forward together.