[mdoering]

Fixes every open issue except #30 (help wanted). One commit per issue, or per group where the issues share a fix.

Read together, the issues overlapped in five groups: separators and blacklists (#209, #210, #211), notes in the authorship (#217, #220, #222), author tokens (#213, #218, #219, #221), qualifiers and ranks (#214, #215), and alternatives (#212, #225, plus the pipe from #219).

Changes

Issue	Before	After
#225, #212	Festuca rubra/arenaria: species with author arenaria	INFORMAL, genus Festuca, phrase rubra/arenaria, state partial
#225	Bombus terrestris or lucorum: trinomial	the same, phrase terrestris or lucorum
#212	Uria aalge/Alca torda, Anser/Branta: first name with an author	unparsable OTHER
#210	Apidae-223957883: unparsable	uninomial Apidae, phrase 223957883
#209	Sus_spec_2, Acrocirridae_gen1: unparsable	genus Sus, phrase spec_2; uninomial Acrocirridae, phrase gen1
#211	Genus sp., ADN Suelo, Wurm-8005097	unparsable; five new blacklisted epithets
#213	Priscula Dup9, Actia Janzen05: code as author	phrase; (Linnaeus1758) splits off the year; new warning for an author still holding a digit
#214	Canis lupus familiaris cf.: quadrinomial	qualifier on the last epithet; also (cf.) and Engl. cf. var. giganteum
#215	DC. subp. japonica: unparsed	subspecies
#216	(BIN BOLD:ADH3884): basionym author BIN Bold	phrase BOLD:ADH3884
#217	L.s.l., Øvst.ined: one author	author and note
#218	vanAchterberg, (deGeer, 1778): epithet, unparsed	authors van Achterberg, de Geer
#219	Singer | M.M. Moser: one author	two authors
#220	W.J. Kress, sp. nov., ined. 2013: author sp.	note sp. nov. ined., year 2013
#221	Rech. fil. subsp. bourgeaui: unparsed	subspecies
#222	Fisher & Orth, 1978: note Orth.	authors Fisher, Orth

Decisions to check

• trailing qualifier or if between author and infraspecificEpithet not always recognized #214, not done as asked: Ulnaria cf. and Neobisium aff. stay the phrase of an unnamed species, as decided in fix(parser): review batch 4 — six FIXME(review) markers #57. The issue asked for an epithet qualifier here too. Changing it is one branch in name_tokens.rs if you prefer the issue's reading.
• Slash between epithets: parse as an informal name with the genus and the rest as phrase #225, authorship: a separate authorship joins the phrase (rubra/arenaria Hack.), so it matches the embedded form and nothing is lost. The issue's "same result" could also mean dropping it.
• Slash between epithets: parse as an informal name with the genus and the rest as phrase #225, rank: these names are always UNRANKED, whatever rank hint is passed.
• Slash between epithets: parse as an informal name with the genus and the rest as phrase #225, or with one genus: Larus marinus or Larus hyboreus moves from OTHER (Name parser: Treatment of alternative names #194) to informal, to match the slash.
• treat slash as unparsable if it occurs after first or second word #212, intergrades: Bubo virginianus pallescens / Bubo virginianus pinorum intergrade anchors on the genus, as Tachybaptus ruficollis tricolor/vulcanorum does in Slash between epithets: parse as an informal name with the genus and the rest as phrase #225. treat slash as unparsable if it occurs after first or second word #212 wished for the species level.
• author not recognized if space is missing after prefix “van”, “de” #218, normalised spelling: a glued particle always gets its space, so duPont becomes du Pont. One test in gna_01 and two golden rows changed for this.
• authors with attached number create no warning #213, year-like codes: after a uninomial, Word1234 is a code (Capitonius Malaise1924), because nearly all of the 562 such rows in the candidate set are trap or interim codes. The three that look like real authors (Thiele1928, Navas1919, Dirsh1953) become phrases too.
• New warning: author with attached number exists only in this engine. It travels as a string, so there is no ABI or model change, but name-parser-api's Warnings class has no constant for it.

Left as they are

• A slash right after the genus with a lower-case word behind it: Belaustium/microsmaris, Hypsithocus / hynsithocus Bergroth, 1927.
• Alternatives with authors or rank markers in the first name: Sorbus udvardyana Somlyay et Sennikov / S. pelsoensis Németh, Viola curtisii ssp. coniophila|Viola tricolor.
• Sus_spec_2 renders as Sus sp. spec_2 in name-parser-api's own NameFormatter, which has a separate copy of the phrase marker. No ChecklistBank name has this shape.

Impact

• ChecklistBank verbatim names, on candidate sets of about 650,000 rows picked by the shapes above: 15,385 distinct names change. About 2,280 become informal alternatives, about 330 become unparsable alternatives, and about 7,260 carry the new warning.
• COL XR 26.6, all 6.4M names: 954 rows change, all reviewed by group. 787 Anon.ined. (now an anonymous author with the note), 85 glued particles, 49 Orth, 19 with the new warning, 10 Genus … and 4 others.
• The candidate sets are shape-based, so changes outside those shapes in ChecklistBank were not measured. The COL XR comparison covers all of its names.