Open duncanpeacock opened 2 years ago
At the moment absolute URLs appear to be stored in just a couple of tables:
viewer_moleculetag (the table that stores tags related to molecules):
The discourse_url is stored as absolute.
Within the additional_info field there is an absolute link to a project snapshot.
Action is required to fix these and the software to prevent them being created.
viewer_computedset: method_url contains URLs to supporting information sources - these do not have to change.
viewer_sessionactions and viewer_snapshotactions: These contain the session project and snapshot json data saved by the frontend. I couldn't see any URLs stored in the session data so no action required there.
For the snapshot actions the data seemed to be encoded - if Boris can check the frontend?.
Notes: viewer_sessionprojecttag
This also has a discourse_url -> we should make sure that this is not ever filled with an absolute url.
Frontend/Backend next steps to be investigated.
Discussed with Boris:
Following that the duplicate of the data/media from production to staging should work (assuming the discourse data has also been duplicated so the relative links work on the staging discourse server).
Fixed in the frontend
Discourse links in Fragalysis are currently stored as absolute URLs linking to the production version of Discourse. To enable a copy to staging, both the Fragalysis and Discourse databases will be copied over.
Hence the Fragalysis data from production to staging will need to be able to link to the staging version of Discourse.