Files
2026-11-18 10:51:31 +01:00

4.0 KiB

labeling

The AIMsight displacement labels and all the domain knowledge to place them, in a single copy on the workstation.

  • manual/<STATION>.jsonl: the expert's labels, for him alone: no process and no agent uses them; auto/<STATION>.jsonl: those of the Support-Agent etiqueteur and those submitted by the aimsight-app analyst; review is the date the user validated a label in aimsight-app (quality over quantity: only the analyses discussed and validated there), and any change removes it; the agents' reviews of 2026-11-04 and 2026-11-05 were removed on 2026-11-17 (in the history); one label per line, {"id", "start", "end", "family", "description", ...}, UTC times, family = the family.
  • auto/analyses/<STATION>_<start>_<end>/: the analyses launched from the Data page of aimsight-app (work/ outside git).
  • auto/support/<YYYY-MM-DD>_<subject>/: the folders of the support agent (agents repository): the request, the answer, the analyst's work when it was called (work/ outside git).
  • auto/cases/<id>.json: the cases, the phenomena, false friends and lessons drawn from the labels, one JSON file each (docs/superpowers/specs/2026-11-18-cases-design.md); built and refreshed by the cases skill (/cases, .claude/skills/cases/), which never writes review nor commits; review is the date the user validated a case, written by the user in its JSON, removed by the skill when it changes the case's text; auto/cases/.read.json, the skill's state (which case each label went to).
  • The knowledge, one source per kind, maintained by hand by the user, but for the cases; the analyst reads the four files in full and searches the cases:
    Kind of knowledge Source
    The cases: phenomena, false friends, lessons auto/cases/, through cases.py
    The labels themselves, as precedents auto/<STATION>.jsonl, through precedents.py
    The families and their boundaries families.md
    How to place a label labeling-rules.md
    What the data mean: measurement chain, gaps, reading rules reading-rules.md
    The installation of each station stations.md
    What the data are: files, columns, image quality, images tools/DATA.md (tools repository)
  • scripts/: validate.py (the labels and the cases against the rules; check for one label), ids.py (the stable id of each), client.py (what the client is told of a detectors' event).
  • scripts/cases.py: the RAG of the cases. The cases closest in meaning to a text, each in full with its provenance (validated by the user or not) and its label sources dated; --list every case, one line each, without Ollama; --pending and --record, the bookkeeping of a refresh by the skill. bge-m3 model of the Ollama container ollama-bge-m3 (started by the script itself if it does not answer; port 11435); the .index/ index, outside git, updates itself at every call (only new or modified cases go through the model). The labels also feed the cases.
  • scripts/precedents.py: the RAG of the labels. The labels of auto/ closest in meaning to a text, with their provenance (validated by the user or not); --exclude STATION START END leaves out the analyzed range; its own index in .index/. An agent uses it, the case search, or both.

aimsight-app and Optimizer reach it through a link labeling -> ~/aimsight-labeling, ignored by git: aimsight-app reads and writes there (labels, analyses), Optimizer only reads; commits are made here. The scripts run from here (python3 scripts/validate.py) or through the link (python3 labeling/scripts/validate.py). Descriptions are written in English from now on.

History taken from Support-Agent (the labeling/ folder, up to commit 48a3092), then the state of aimsight-app on 2026-11-13 (462ba9e): the archives of the etiqueteur and of the reviews (auto/etiqueteur/, auto/revue/, auto/revue-croisee/) and the old labeling.py tool remain in the history.

docs/archive/: the design and plan of the labeler (not built), the label map and knowledge/, archived on 2026-11-16.