[jt55401]

What this adds

Adds a Brazil-range collared peccary record with a species-specific foraging FSM and BR activity curve.

Closes #988

Evidence and scope

GBIF usage key 2440995 and iNaturalist taxon 42113 both resolve to accepted species Pecari tajacu, checked 2026-11-01. The Mammal Diversity Database and Wikipedia use Dicotyles tajacu following the 2020 nomenclatural treatment; the record follows the requested GBIF/iNaturalist provider names and records that disagreement.

Mass (21.13 kg) comes from PanTHERIA AdultBodyMass_g (21133.69 divided by 1000 and rounded to 0.01 kg). Typical adult mass range (16–27 kg) and body length (1.0–1.5 m; 1.25 m midpoint estimate) come from Wikipedia. A Paraguay Dry Chaco field study supports group size (5–20) and local activity timing. Diet weights and model geometry are marked as authored simulation judgments. The optional shoulder-height field is omitted because ADW (30–50 cm) and Wikipedia (51–61 cm) conflict.

The required lifespan_y is 7 years, taken from the ADW account's explicit average wild lifespan statement. Its separate 15-year wild high and 31-year captive high were not used. ADW is CC BY-NC-SA 3.0; after checking permissive alternatives that only supplied captive or maximum values, it is retained as a last resort for lifespan_y and the year-round breeding statement. The ADW account and search limitations are recorded in provenance. The Paraguay paper is CC BY 4.0 and supports Dry Chaco activity and group-size claims. The range retains Uruguay as introduced based on the cited source.

Activity and behaviour

The BR curve uses 1,552 georeferenced GBIF occurrences from Brazil under the CC0/CC BY query, with all 12 monthly counts nonzero and usable: true. The iNaturalist Brazil query returned 439 observations at tier C, below the configured 500 minimum; it is recorded as unusable and excluded from the curve. The final curve is GBIF's monthly share, not a measure of abundance.

The local program fsm-pecari-tajacu-forager-v1 makes browse, fruit and invertebrate feeding actions reachable, walks while foraging, trots to flee, and rests during the 55–90° sun-elevation interval corresponding to the Dry Chaco study's hot midday rest. The study reports activity after sunset, morning peaks and midday rest. Arizona seasonal day/night reversals are not generalized to Brazil. The program is species-local because the shared programs do not represent these food tags and the local rest interval together.

Validate report

Tested commit: ebbc22de9e608584a69436c0737f24207b76c3da.

uv run tools/validate.py species/pecari-tajacu --write-attribution --report: 0 errors, 2 warnings; 34/34 provenance checks. Warnings: optional shoulder height omitted due to conflicting measurements; ADW NC source is limited to lifespan and year-round breeding and is justified above. Attribution generation is committed and idempotent.

uv run tools/validate.py --self-check: 643 files scanned, 0 findings.
uv run tools/validate.py --report: all species, 0 errors.
uv run tools/validate.py --locale all --report: all locales, 0 errors.
uv run --with pytest --with jsonschema pytest tools/tests -q: 349 passed, 8 skipped.

Preview evidence

Built with python3 tools/build_site.py and inspected in Firefox via the local preview. Checked idle, walk, and trot; all declared poses were rendered and showed no visible clipping. The shoulder collar is a simple pale block band in this stylized model.

Checklist

Species

• Closes the linked species-request issue; issue claimed before work
• Checked current main and open PRs for this taxon
• Taxon identifiers and source claims checked at their providers
• Evidence locators, units/statistics and authored transformations documented
• Species validation has no errors; warning resolutions stated above
• Attribution regenerated and committed
• BR activity curve generated; query scope, licence tier, counts and usability inspected
• Executed feeding, gaits, daytime rest and regional scope reviewed
• New record; no existing shared-record fields or regional curves were changed
• Repository self-check, all-species/all-locale validation and tests pass on this commit
• Every declared pose previewed: idle, walk, trot
• CC BY and CC BY-SA sources preferred; ADW NC use justified above
• No private paths, hostnames, or tokens in the diff
• One species in this PR

Non-CC0/CC-BY sources

The ADW CC BY-NC-SA 3.0 source supports only traits.lifespan_y and seasonal.phases. AnAge's accessible species data and the checked AZA source provided captive longevity or maxima, not a typical wild lifespan. ADW is the only checked source that explicitly distinguishes typical wild lifespan from its wild and captive maxima. Its required datum is 7 years; the 15-year wild and 31-year captive maxima are excluded. No ADW morphology or diet values rely on the NC source.

Wikipedia is CC BY-SA 4.0 and supplies several lower-risk descriptive traits; it is retained with attribution. The Dry Chaco paper and GBIF/iNaturalist provider data are CC BY 4.0.

Remaining work

Maintainers should review the ADW NC last-resort justification, the conflicting optional shoulder-height sources, and the documented genus usage disagreement. No live import or release acceptance is claimed.

[jt55401]

Hourly maintainer review: ebbc22d

Disposition: changes requested for behavior, geography and provenance. No merge.

• species/pecari-tajacu/behavior.json:transitions: feeding precedes rest at equal priority, and active states have no high-sun rest exit. After idle dwell, sun +70°, hunger 0.9, nearby browse and a successful browse chance gives idle → browse; hungry browse continues through noon. Add higher-priority rest guards from ordinary active states, and outcome regressions. tools/tests/test_pecari_tajacu_behavior.py currently checks the rest parameters without testing selection under competing food conditions.
• The program calls itself “Brazilian Dry Chaco,” but Zaldivar et al., Methods/Activity patterns studied Estancia Montania, Boquerón, Paraguay, with 265 collared-peccary records. It reports low activity at 12:00–17:00, not a measured 55°–90° sun interval. Provide target-region evidence or an explicit defensible transfer and authored angular approximation; remove the false Brazilian study-site description. Do not present Paraguay timing as observed for Brazil.
• species.json:range.introduced=[UY] is justified by “reintroduced.” The cited Distribution and habitat says reintroduced after 100 years of extirpation. Reintroduction does not establish alien status; verify historical native status and correct the country classification/provenance.
• seasonal.life_stages contains Juvenile duration_days=270/scale=0.65, while its provenance claims Adult only. Source/derive those values or omit unsupported optional timing. Narrow the mass-conversion provenance with derived: true; correct the ADW source attribution to the currently named author Aaliyah Benson and document the actual iNaturalist taxon licence grant rather than assuming CC BY 4.0. Regenerate attribution.

Live accepted GBIF 2440995 and active iNaturalist 42113 agree; the Dicotyles/Pecari disagreement is disclosed. ADW Lifespan/Longevity explicitly labels seven years average wild, separately from maxima; text NC-SA terms are verified. Wikipedia supports 16–27 kg and the 1.25 m length midpoint. Cached PanTHERIA row 3393 has AdultBodyMass_g 21133.69, matching 21.13 kg; the cited archive returns 403, so I could not independently refresh that material/licence at its source. Both the Paraguay article and its journal page link CC BY 4.0. All three diet tags are reachable; no month predicates occur.

BR place 6878, country BR, Mammalia baseline and filters checked. 1,552 total/1,481 month-bearing GBIF records reproduce every monthly value; iNaturalist 439 remains excluded below the configured 500 threshold. Existing species/regions preserved. Public checks pass: focused 0 errors/2 warnings, 34/34 provenance, self-check, all-species/all-locale validation, pytest 349 passed/8 skipped, attribution freshness. Current-head validate is green. Viewed idle, walk, trot: collar/limbs remain attached, no blocking clipping; default mesh has no eyes. Private import/live-scene acceptance untested.