fix(oo): skip junk dirs in applications Discover

Stop inventing CRM companies from node_modules/tools README trees.
Only sync dossier-like folders (source-NNN-…) and invent Company from
the slug token after the numeric id.
This commit is contained in:
2026-07-23 14:43:45 +01:00
parent 09e242d7c3
commit 3969d045cc
+5 -1
View File
@@ -751,7 +751,11 @@ oo applications sync --path ./applications/2026 --verbose
oo applications sync --path ./applications/2026 --apply --verbose
```
Each `README.md` is parsed for company, role, email, phone, LinkedIn, etc.
Each dossier `README.md` is parsed for company, role, email, phone, LinkedIn, etc.
Discovery walks `--path` and **skips** junk trees (`node_modules`, `tools`, `.venv`,
`pdfs`, …). Only folders that look like CV application slugs (`source-NNN-…`, or
long hyphenated dossiers) are synced — not npm package READMEs.
The sync creates or finds contacts, opens a deal, adds members, and appends a
history note. Re-running is safe: duplicate members and duplicate deal titles
are skipped when already present.