NESCent / FossilCalibrations

Fossil calibrations database
http://fossilcalibrations.org
BSD 2-Clause "Simplified" License
14 stars 4 forks source link

Incomplete Autocomplete? #71

Open Ksepka opened 9 years ago

Ksepka commented 9 years ago

Reported by Rachel:

In "advanced search" stage, using common, rather than scientific/proper names - the "by tip taxa" and "by clade" options won't let me search for items that don't appear in the autofilled list (which includes some but not all common identifiers).

For example, I can search for "lizard" in the main search box, but not in the taxon specific boxes, which only let me search for things beginning with "lizard'. "snake" is another example.

I think this is because whatever is infilling those boxes is a non-exhaustive lists of the common names - I have no problem searching for "cat", "dog", or "cat" + "dog". Another example is "chimp", which brings up results via the main box, but not the clade or taxon specific boxes, which will only recognise "chimpanzee" (although in this case the term does appear in the autofilled list).

jimallman commented 9 years ago

In cases like these, we want to match from controlled vocabularies. If a name doesn't appear in the autocomplete field, it doesn't exist (in the FCDB, at least) as a discrete term that we can match. The field on the home page is the most inclusive, drawing from all names in table AC_names_searchable, which is defined as follows in the database:

Add tables for responsive "auto-complete" widgets in client UI:
    AC_names_nodes -- NCBI taxa (and other sources) and FDC-contributed nodes?
    AC_names_taxa  -- just NCBI nodes (and other sources)?
    AC_names_extant_species -- filtered view of AC_names_nodes (or AC_names_taxa), only leaf nodes?
    AC_names_clades -- filtered view of AC_names_taxa, all but leaf nodes?
    AC_names_searchable  -- all above, plus names of authors, publications, journals, collections, fossils

As Rachel found, the common names "lizard" and "chimp" aren't included in either NCBI's taxonomy or among FCD's node names. One way to add these would be using a supplementary table (possibly the existing FCD_names) and keying them to NCBI taxa.