GitLab

- Can now also handle DOIs.
- Added support for {PMID:PubMed 2019 abstract:12345678} format. The script will fetch the paper's author and year using eutils and rewrite the reference to {PMID:Author et al (Year):12345678}.
- Implemented a cache to make sure eutils doesn't get the same question twice.
- Updated lovd_callJSONService() to be able to return an array instead of an object.

Reverted ONLY_FULL_GROUP_BY changes.

- It makes no sense to order on something, if we do a GROUP BY on the PK that we also do a WHERE on. Only one line will be returned anyway.
- Added sorting to the GROUP_CONCAT() so that the gene symbols will be sorted in the email's subject.
- Don't overcomplicate the query for fetching curators.

- Don't pull out extra fields out of the database, just because we're sorting on it.
- Don't sort on a field alias if you can sort on the actual fields.

- Genes are assumed to be in sense when at least one transcript is in sense.

- Also simplified SELECTs whereever possible.
- Custom VLs with VOG as the first table (Variant VL on Screenings VE and Individual VE) no longer show the minimal VOT's effectid, but instead the VOG's effectid, which makes much more sense.
- These changes might also lead to some speedups.

- Reviewed all MIN()/MAX() uses.
- For BED files, always return the mapping on the earliest transcript.
- For the genes list, assume the gene is in sense when at least one transcript is in sense.

Fix `variants/in_gene` view.

- Removed more min() and max() calls.

- This speeds up the query quite a lot.
- Also, rename LEFT JOINs to LEFT OUTER JOINs to make more clear what they are.

- Even if we try to play nice, the totally unnecessary MIN() and MAX() calls slow down queries a lot.
- See #386.

Fixed some problems with the DOI view.

- Pagination was off, but especially if you're searching by journal you need pagination.
- Fixed the journal search; DOIs can also contain slashes and journal names can contain dots.
- Cleaned up the code a bit.

Allow for publisher and journal-specific queries using the DOI reference search.

Enlarge the DNA and protein fields to support longer insertions.

- Normally DOI reference searches are on the full DOI only (like 10.1002/humu.21438). Now, also allow for searches using the journal's DOI, like 10.1002/humu.
- Implements #257.

- This allows for larger insertions to be stored.
- LOVD+ is using the same size, so this also removes one difference between LOVD3 and LOVD+.

- This allows for larger insertions to be stored.
- LOVD+ is using the same sizes, so this also reduces one difference between LOVD3 and LOVD+.

Multiple bug fixes.

Updated the ReadingFrameChecker; it now provides the correct (new) HGVS nomenclature.

The public gene-specific downloads now have these non-public columns removed from the Phenotype object as well.

- Selected user's IDs where stored in both userid and id. Since the latter is visible, part of this selection could be influenced. Now only userid is used, making the filtering of the VL complete invisible to the user.
- Also, when selecting users as curators/collaborators, their ID wasn't added to their name, like with previously selected users.

- The dropdown became too large on large databases.
- The VL is much better equipped to handle large datasets.
- The output is restricted for curators in case they curate 25 genes or less.
- The currently selected gene is highlighted.

A download will now no longer show inactive Phenotype and VOT columns, if filters are active for Diseases or Genes, respectively.
- This cleans up the file a lot for single-gene downloads.
- It reports a warning for Phenotype data, since it's possible that data is selected unrelated to the selected Disease(s).

The public gene-specific downloads had some non-public columns included, like status, created_by, created_date, etc.
- These are now removed. They're not damaging to be shared, but they're not shown on VLs or VEs, so they should remain hidden in the download as well.

- Now, it'll take the VOG's effectid and it'll copy it.

- The old notation (c.154-?_295+?del) was not correct anymore.
- The new notation is c.(153+1_154-1)_(295+1_296-1)del.

Allow LOVD3 to nicely display JSON data in fields.

Fixed bug; lovd_getVariantInfo() didn't recognize delins variants.

- VLs just show "(data)" and don't display the data.
- VEs show a table structure displaying the JSON data. May not work for all data, but works for most.
- Implemented for the VKGL data that is now in the shared GV LOVD3.

- Also updated tests.

Fix optimizer for complex queries.

- Added unit test for this, too.

- It was looking only for aliases, but the multi value filter is putting a full column reference in there.
- Also added a unit test for this new code.

- If this still doesn't process all LOVD queries, I'll need to break it into pieces; this regexp is too complex already.