Spaces:
Sleeping
Sleeping
RootScope v4: manuscript model (LightGBM per round, 3 seeds, layer + soft-neighbor context, radial prior)
bb112f6 verified | """Species knowledge used by the exodermis decision. | |
| The table ``data/species_exodermis.csv`` holds, for every training species, | |
| whether an exodermis was annotated in the training sections, and for a few | |
| species outside the training set what the literature reports. Perumalla, | |
| Peterson and Enstone (1990, Bot J Linn Soc 103:93) found a hypodermis with | |
| Casparian bands in 89% of 181 angiosperm species, so most species have the | |
| layer; but the exodermis is a variable, stress- and age-dependent structure | |
| and the literature disagrees even for pea. The table is therefore a hint: | |
| only the training annotations decide the labels, the literature only informs | |
| the report. | |
| """ | |
| import csv | |
| import re | |
| from pathlib import Path | |
| _TABLE = None | |
| def table(): | |
| global _TABLE | |
| if _TABLE is None: | |
| p = Path(__file__).parent / "data" / "species_exodermis.csv" | |
| with open(p, newline="") as f: | |
| _TABLE = list(csv.DictReader(f)) | |
| return _TABLE | |
| def _norm(s): | |
| s = re.sub(r"[^a-z\s-]", " ", str(s).lower().replace("_", " ")) | |
| return " ".join(s.split()) | |
| def lookup(name): | |
| """Match a user-supplied species name against the table. | |
| Accepts ``"Pisum sativum"``, ``"P. sativum"``, ``"Psativum"``, ``"pea"``-style | |
| strings are NOT resolved (no common names). Returns the row (dict) with an | |
| added ``match`` key ("species" or "genus"), or None. A genus-only match is | |
| returned only when every species of that genus in the table agrees.""" | |
| if not name or not str(name).strip(): | |
| return None | |
| q = _norm(name) | |
| rows = table() | |
| for r in rows: # full binomial | |
| if _norm(r["binomial"]) == q: | |
| return dict(r, match="species") | |
| parts = q.split() | |
| for r in rows: # abbreviated genus: "p sativum", "psativum" | |
| g, sp = _norm(r["binomial"]).split()[:2] | |
| if len(parts) == 2 and parts[1] == sp and g.startswith(parts[0][0]) and len(parts[0]) <= 2: | |
| return dict(r, match="species") | |
| if len(parts) == 1 and parts[0] == g[0] + sp: | |
| return dict(r, match="species") | |
| genus = parts[0] | |
| same = [r for r in rows if _norm(r["binomial"]).split()[0] == genus] | |
| if same: | |
| vals = {(r["exodermis_in_training"] or r["exodermis_literature"]) for r in same} | |
| if len(vals) == 1: | |
| return dict(same[0], match="genus", binomial=genus.capitalize() + " sp.") | |
| return None | |