Files
Support-Agent/labeling/auto/etiqueteur/WILDSTRUBEL_2026-09-07/outils/serie.py
T
maximeandClaude Opus 5.5 35b5ed4d74 Auto labels reviewed per station and cross-reviewed; all labeling knowledge in labeling/ (D-130)
- Every station's auto labels reviewed by an analyst (/review-auto-labels), then a
  cross-review per family across all stations (new skill /cross-review-labels),
  a finishing pass, and descriptions rewritten in words (no numeric values).
- The three never-labeled ranges (WILDSTRUBEL 07/09-01/10, NORDEND 16/08-01/10)
  labeled and reviewed: 991 auto labels on 11 stations, labeling/controle.py at 0 errors.
- Labeling knowledge gathered in labeling/: regles-d-etiquetage.md (new), familles.md
  (settled boundaries), connaissances/ (notes moved from docs/, links updated),
  the label map rebuilt from reviewed labels, events and cases; events tied to their
  stations (PS_DES_COUDRES is JUNGFRAU). The labeler and reviewers read only labeling/.
- labels-map reads auto labels and events; decision D-130; living plan points 73-75.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-05 08:21:57 +01:00

108 lines
4.2 KiB
Python

"""Séries corr_x, corr_y des cibles (mesures acceptées), axe borné aux centiles, et tilt de la caméra.
Usage : .venv/bin/python <ce fichier> DEBUT FIN [--table] [--capteurs A,B] [--nom N]
--table : imprime, à chaque instant, corr_x/corr_y de chaque cible (r = rejetée) et le nombre de références inlier.
Écrit le PNG dans le dossier parent de outils/.
"""
import csv, glob, os, sys
from datetime import datetime, timedelta, timezone
ICI = os.path.dirname(os.path.abspath(__file__))
SORTIE = os.path.dirname(ICI)
DATA = "data/WILDSTRUBEL"
CIBLES = ["PA2_2", "PA3_2", "PA_9", "PB1_2", "PB2_2", "PB4", "PB_5"]
def t(s):
return datetime.strptime(s[:16], "%Y-%m-%dT%H:%M").replace(tzinfo=timezone.utc)
def arrondi(s):
d = datetime.fromisoformat(s.replace(" ", "T"))
m = (d.minute + 7) // 15 * 15
return d.replace(minute=0, second=0, microsecond=0) + timedelta(minutes=m)
def lire(debut, fin):
rows = {}
jour = debut.date() - timedelta(days=1)
while jour <= fin.date():
f = f"{DATA}/displacement/WILDSTRUBEL_advanced_{jour:%Y%m%d}.csv"
if os.path.exists(f):
for r in csv.DictReader(open(f)):
ti = arrondi(r["time_utc"])
if debut <= ti <= fin:
rows[(ti, r["sensor_name"])] = r
jour += timedelta(days=1)
cam = {}
jour = debut.date() - timedelta(days=1)
while jour <= fin.date():
f = f"{DATA}/camera/WILDSTRUBEL_camera_{jour:%Y%m%d}.csv"
if os.path.exists(f):
for r in csv.DictReader(open(f)):
ti = arrondi(r["time_utc"])
if debut <= ti <= fin:
cam[ti] = r
jour += timedelta(days=1)
return rows, cam
def main():
a = sys.argv[1:]
debut, fin = t(a[0]), t(a[1])
cibles = CIBLES
if "--capteurs" in a:
cibles = a[a.index("--capteurs") + 1].split(",")
nom = a[a.index("--nom") + 1] if "--nom" in a else ""
rows, cam = lire(debut, fin)
instants = sorted({k[0] for k in rows})
if "--table" in a:
refs = sorted({k[1] for k, r in rows.items() if r["pattern_type"] == "global_ref"})
print("instant " + " ".join(f"{c:>15}" for c in cibles) + " inl tilt")
for ti in instants:
cells = []
for c in cibles:
r = rows.get((ti, c))
if r is None:
cells.append(f"{'-':>15}")
elif float(r["error_value"]) > float(r["error_threshold"]):
cells.append(f"{'r':>15}")
else:
cells.append(f"{float(r['corr_x']):+7.3f}/{float(r['corr_y']):+7.3f}")
n = sum(1 for rf in refs if rows.get((ti, rf), {}).get("inlier") == "1.0")
tl = cam.get(ti, {}).get("median_tilt")
print(f"{ti:%m-%d %H:%M} " + " ".join(cells) + f" {n:3d} {float(tl):.3f}" if tl else f"{ti:%m-%d %H:%M} " + " ".join(cells) + f" {n:3d} -")
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
fig, ax = plt.subplots(len(cibles) + 1, 2, figsize=(16, 2.0 * (len(cibles) + 1)), sharex=True)
for i, c in enumerate(cibles):
for j, col in enumerate(["corr_x", "corr_y"]):
xs, ys = [], []
for ti in instants:
r = rows.get((ti, c))
if r and float(r["error_value"]) <= float(r["error_threshold"]) and r[col]:
xs.append(ti); ys.append(float(r[col]))
if not ys:
continue
s = sorted(ys)
lo, hi = s[int(0.01 * len(s))], s[int(0.99 * len(s)) - 1]
m = (hi - lo) * 0.3 + 0.02
ax[i, j].plot(xs, ys, ".-", ms=2, lw=0.6)
ax[i, j].set_ylim(lo - m, hi + m)
ax[i, j].set_title(f"{c} {col}", fontsize=8)
ax[i, j].grid(alpha=0.3)
for j, col in enumerate(["median_tilt", "median_rotation"]):
xs = sorted(cam)
ax[-1, j].plot(xs, [float(cam[x][col]) for x in xs], ".-", ms=2, lw=0.6, color="g")
ax[-1, j].set_title(f"caméra {col}", fontsize=8)
ax[-1, j].grid(alpha=0.3)
fig.tight_layout()
p = f"{SORTIE}/serie_{nom}{debut:%m%d%H%M}_{fin:%m%d%H%M}.png"
fig.savefig(p, dpi=80)
print(p)
if __name__ == "__main__":
main()