crahal

 Add a line in `generate_data.py` which checks for and only retains p-values \in [0, 1].

Could be because: * Upstream data structure change * Escape character is breaking our parser\wrangler.

For speed improvements, especially c. bubble graph. Thanks to Peter/devs at GI for raising this.

If generate_data.py is interupted but raw files are already obtained, the logger reports that no new data is found the next time, although the script did not generate the required...

automatic tweet on update

2

to run at the end of generate_data.py

Some agencies are left unmapped by generate_data.py

2

Investigate why some agencies aren't getting linked to papers, e.g. in Cat_Anc_wBroader_withParents.tsv. Are there no returns from PubMed?

Related to #81, but specifically related to improving the quality of the generated reports.

Doughnut chart is always broken after filtering for any funder: ```No data found for All Parent Terms in 2022``` Shouldn't happen when there are _some_ studies in that specific year...

Turn csvs into postgres db

1

- Turn csvs into a postgres db with either timestamped views or a timestamped column - Call the postgres DB when the app is substantiated instead of loading csvs