Bigsdb

This is a minor release that:

- improves the layout of the user database registration page, making use of the JQuery.multiselect plugin.
- provides an option to select only DNA or only peptide loci when making a sequence query in databases that contain both kinds.

**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.48.4...v_1.48.5

This version speeds up the display of results for an isolate query where scheme fields from large schemes (e.g. cgST for cgMLST schemes) are displayed. The scheme values used to be determined from the isolate allele designations in real time so that they were always up-to-date, but this can take 500ms+ for a large cgMLST scheme per isolate, which is noticeable when displaying a table of 25 records. Now, if 'cache_schemes' is set in config.xml, and the scheme has >100 loci then the scheme cache table is used for the lookup. Smaller schemes (e.g. MLST) are still determined using a live lookup.

If a scheme field has multiple values, which can happen for cgSTs due to missing loci, then these are now collapsed in the results table with only the cgST that has the fewest missing loci shown. Other matching cgSTs can be shown by clicking a hyperlink.


The second record below has the list expanded:


**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.48.3...v_1.48.4

This version introduces a number of performance improvements to sequence queries. The cached BLAST database now also includes sequence length in its headers to avoid a database lookup during the query - to see maximum improvement immediately, mark any BLAST caches as stale (using link in admin interface) so that they can be re-created. The query will still work using the old cache if this isn't done, and the cache will be recreated automatically once a new allele is added or when it is >7 days old.

If LIN codes are assigned for a scheme these are now used to identify the nearest matching genomes following a sequence query.


> **Please note the new dependency for Perl module Text::CSV.**


**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.48.2...v_1.48.3

This release fixes a bug when attempting to download a scheme TSV file from the API when the scheme does not have a primary key field, e.g. https://rest.pubmlst.org/db/pubmlst_neisseria_seqdef/schemes/6/profiles_csv. Due to the recent change in the way the results were streamed (introduced in v1.48.1), the check for a valid scheme was not occurring in time, and the process would crash when attempting to access a non-existent table.

**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.48.1...v_1.48.2

This version fixes a couple of issues:

- LIN code prefix lookup of nickname field values was matching multiple values, e.g. the prefix 0_0_1051 was also returning values for prefixes 0_0_105, 0_0_10 and 0_0_1.
- LIN code assignment was failing when a scheme had a field called 'profile_id'.

In additional a performance issue has been addressed for the API cgMLST profile download in TSV format. This was taking a long time (and potentially timing out).

**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.48.0...v_1.48.1

This version adds optional placeholder text to search fields in isolate queries. You can add placeholder text for any of the fields defined in config.xml by setting the `placeholder` attribute, e.g.


<field type="text" required="no" length="50" maindisplay="no" comments="region inside country" placeholder="e.g. Oxfordshire">region</field>

Placeholder text can also be set for sparse fields, scheme fields, LIN codes, and LIN code fields by setting the `placeholder` value in the appropriate table - (in the admin interface):


**Full Changelog**: https://github.com/kjolley/BIGSdb/compare/v_1.47.3...v_1.48.0