- 20 Feb, 2024 1 commit
-
-
Przemyslaw Kaminski authored
-
- 02 Feb, 2024 1 commit
-
-
Przemyslaw Kaminski authored
-
- 17 Nov, 2023 1 commit
-
-
Przemyslaw Kaminski authored
-
- 16 Nov, 2023 1 commit
-
-
Przemyslaw Kaminski authored
-
- 08 Nov, 2023 1 commit
-
-
Przemyslaw Kaminski authored
-
- 28 Sep, 2023 2 commits
-
-
Przemyslaw Kaminski authored
For ngrams table, we set ngramsValidPatch = mempty because the table is refreshed and we have issue gargantext/purescript-gargantext#594 However, we need to keep that field, it turns out, because we have gargantext/purescript-gargantext#598 Recently applied patches are needed for the document ngrams highlighting. This all would be unnecessary (or at least moved to the backend) if we had gargantext/purescript-gargantext#553
-
Przemyslaw Kaminski authored
This reverts commit 058fa2fe.
-
- 08 Sep, 2023 2 commits
-
-
Przemyslaw Kaminski authored
-
Przemyslaw Kaminski authored
The issue was this: When user clicks around the ngrams table, only local state is updated (`ngramsLocalPatch`). After he clicks 'Sychronize', this is uploaded to server and `ngramsVersion` number is bumped up. What happened is that `ngramsValidPatch` inside internal state was updated and so the patch was sychronized locally, without refreshing the table. However, this resulted in errors later, when user changed the same term, because then, when user changes view to 'Stop terms', the table is refreshed with the newly applied terms and `ngramsValidPatch` doesn't make sense anymore. The solution is to set `ngramsValidPatch` to `mempty` and, after state is updated, refresh the current table. It could be argued whether it's the most optimal thing to do and why don't we keep things "offline". Well, clicking on 'Sychronize' requires the client to be online so why not sync the table already at this time? I guess this makes things bit simpler and I think it renders `ngramsValidPatch` unnecessary. So our state could be simplified (patches is already a complex beast).
-
- 26 Jun, 2023 1 commit
-
-
Przemyslaw Kaminski authored
Also, tests fixed a bit (at least they work).
-
- 21 Jun, 2023 2 commits
-
-
Przemyslaw Kaminski authored
-
Przemyslaw Kaminski authored
-
- 19 Jun, 2023 1 commit
-
-
Alexandre Delanoë authored
-
- 06 Jun, 2023 2 commits
-
-
Przemyslaw Kaminski authored
But this will never be the same with what the backend is doing.
-
Przemyslaw Kaminski authored
-
- 05 Jun, 2023 1 commit
-
-
Przemyslaw Kaminski authored
-
- 01 Jun, 2023 1 commit
-
-
Przemyslaw Kaminski authored
Also, limit to context ngrams only, for searching highlight patterns.
-
- 30 May, 2023 3 commits
-
-
Przemyslaw Kaminski authored
-
Alexandre Delanoë authored
This reverts commit 43c5b7b5.
-
Przemyslaw Kaminski authored
-
- 26 May, 2023 1 commit
-
-
Przemyslaw Kaminski authored
-
- 24 Apr, 2023 1 commit
-
-
Alexandre Delanoë authored
-
- 01 Mar, 2023 1 commit
-
-
Przemyslaw Kaminski authored
Also, performance is fixed now.
-
- 13 Dec, 2022 2 commits
-
-
Przemyslaw Kaminski authored
-
Przemyslaw Kaminski authored
-
- 15 Nov, 2022 1 commit
-
-
Przemyslaw Kaminski authored
-
- 15 Sep, 2022 1 commit
-
-
Przemyslaw Kaminski authored
-
- 30 Jun, 2022 2 commits
-
-
Przemyslaw Kaminski authored
-
Przemyslaw Kaminski authored
-
- 21 Jun, 2022 1 commit
-
-
Przemyslaw Kaminski authored
-
- 20 Jun, 2022 1 commit
-
-
Przemyslaw Kaminski authored
-