gnabbr

gnabbr resolves abbreviated genera of scientific names found in documents: C. elegans -> Caenorhabditis elegans. It is a Go library for gnfinder and bhlindex.

Status: early development.

How it works

1. Find names with an abbreviated genus: a capital letter with up to two lowercase letters and a period, then an epithet (C. alba, Cae. alba).
2. Group them locally: the same abbreviation in the same section (article, chapter) of the document, split where names are far apart (in characters). Names in a group usually share a genus; a repeated name is reconciled once per group.
3. Hint: full names mentioned shortly before (Carex nigra) give genera the abbreviation may stand for.
4. Reconcile each group with gnames abbr_names, adding the taxonomic statistics of the section or the document (main taxon, dominant class), given by the caller.
5. Accept the best genus when its score reaches a threshold (8 by default, which favors precision), and expand the group's names with it (C. alba -> Carex alba). Callers verify expanded names with their other names.

Library

import (
	"github.com/gnames/gnabbr"
	"github.com/gnames/gnabbr/pkg/config"
	"github.com/gnames/gnabbr/pkg/doc"
)

ga := gnabbr.New(config.OptGnamesURL("http://localhost:8888/api/v1/"))
out, err := ga.Resolve(ctx, doc.Document{
	ID: "item-1",
	Names: []doc.Name{
		{ID: "1", Name: "Caenorhabditis elegans", Start: 0},
		{ID: "2", Name: "C. briggsae", Start: 2400},
	},
	TaxonStats: taxonStats, // gnstats stats.Stats of the document
})
// out.Reconciled: groups with an accepted genus and expanded names;
// out.Failed: groups without one, with their verbatims.

Callers do not deal with gnames: the root package gnabbr joins the core logic (pkg/gnabbr) with the internal abbr_names client. pkg/ never imports internal/ and has no configuration file or database dependencies.

gnames

gnabbr needs a gnames with the abbr_names endpoint (gnames branch 137-abbr, not released yet). The default gnames URL, https://verifier.globalnames.org/api/v1/, works only after such a gnames is deployed there; until then set it to a local gnames with config.OptGnamesURL.

Layout

gnabbr.go           entry point: New(opts...) joins pkg/gnabbr and internal/iognames
pkg/gnabbr          core: GNabbr, AbbrNamer, Resolve, Output
pkg/config          settings (functional options, no I/O)
pkg/doc             input: documents and name occurrences
pkg/abbrname        parsing abbreviated names
pkg/group           local groups
pkg/hints           genus hints
internal/iognames   abbr_names REST client (implements pkg/gnabbr.AbbrNamer)