Files
aimsight-labeling/auto/analyses/NORDEND_2026-10-13T0745_2026-10-24T1630/scripts/build_requalified.py
T
maximeandClaude Opus 5.5 ddb17c2189 auto: labels requalified under the 5-noise rules (only what stands out from the noise): 32 analyses refined, the 6 paused periods verified, the long labels requalified with the fixes the period analysts found; 1153 -> 961 labels
- 32 analyses: each label kept, rewritten or removed by its analyst (section Requalification of the report; mm magnitudes where the targets have them)
- SNCF01 x3, WILDSTRUBEL 2026-09-01 -> 2026-10-06 x3: verified from their existing labels under the new rules
- long labels (no analysis): auto/analyses/long-labels/<STATION>/; NORDEND-169 and -195 run to 07/11, so NORDEND-283 and -284 (their continuation) are removed
- reading-rules.md, stations.md: the citations of removed or rewritten labels point to their successor, or to [expert]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-17 23:58:00 +01:00

23 lines
3.7 KiB
Python

"""Writes labels.jsonl from work/requalify_input.jsonl: removed labels dropped, rewritten ones replaced (review field
dropped), the others kept exactly; the id field removed."""
import json, sys
D = sys.argv[1]
REMOVED = {"NORDEND-183", "NORDEND-188", "NORDEND-191", "NORDEND-194", "NORDEND-199", "NORDEND-201", "NORDEND-207",
"NORDEND-274", "NORDEND-277"}
NEW = {
"NORDEND-186": "In mid-afternoon, the sun lights the rock behind the references CTR02 and CTR03 and the local reference CTR01 (zone thumbnails of 14:30, 15:30 and 16:15: the rock bright around them at 15:30, back in shadow at 16:15), the wall of the targets staying in shadow. The error_value of CTR03 rises markedly without it being discarded, CTR01 (read as a target) is rejected; the references move apart a little on corr_y and the corr_y of PR13, PR12 and PR11 drop clearly together for a measurement or two, then come back. Cause seen on the thumbnails: the sun on the rock behind the references, which disturbs the base. Not a movement; ignore the points of the range.",
"NORDEND-208": "Weak thermal cycle of the camera: overcast day, the housing temperature barely rises (highest around 14:00); the pan barely drops, the tilt rises barely more; the camera comes back in the evening and at night, when the housing cools down below its morning level, and the pan goes beyond its level of the day before. By day the compensation absorbs it, the corrected curves staying close to their night level; but in the evening the corr_y of PR06, PR05, PR01, PR11 and PR12 drop a little together and stay low all night, until the start of the next morning: what the compensation leaves of the night overshoot of the camera. The jumps of the faint targets by day are labelled separately. Deduced cause: the pose follows the housing temperature, and the rock face of the quarry does not move with the temperature. The structure does not move, nothing to do.",
"NORDEND-210": "Weak thermal cycle of the camera: overcast day, the housing temperature only rises a little (highest around 14:00); the pan drops a little, the tilt rises a little less; but the camera does not come back at night: on the morning of the 23rd, before the rain, the pan is still lower than the day before, by almost its whole drop of the day. The compensation does not absorb it all: in the morning the corr_y of PR01, PR12, PR05, PR10, PR11 and PR06 rise a little together and stay up all afternoon and night, like the pan, and the corr_x of PR18 drops a little in the afternoon and evening. The jumps of the morning and of midday are labelled separately. Deduced cause: the pose follows the housing temperature, and the rock face of the quarry does not move with the temperature. The structure does not move, nothing to do.",
"NORDEND-214": "Weak thermal cycle of the camera: overcast day, the housing temperature only rises a little (highest around 12:00); the pan drops a little, the tilt rises as much; the camera comes back in the afternoon, until the night of rain (separate label). The compensation does not absorb it all: from late morning the corr_y of PR13 rises clearly, those of PR06, PR12 and PR18 a little, and they are not yet back at their night level when the label ends. The two one-measurement jumps of the afternoon are labelled separately. Deduced cause: the pose follows the housing temperature, and the rock face of the quarry does not move with the temperature. The structure does not move, nothing to do.",
}
with open(f"{D}/labels.jsonl", "w") as out:
for line in open(f"{D}/work/requalify_input.jsonl"):
d = json.loads(line)
i = d.pop("id")
if i in REMOVED:
continue
if i in NEW:
d["description"] = NEW[i]
d.pop("review", None)
out.write(json.dumps(d, ensure_ascii=False) + "\n")