chore: auto-commit (6 archivos)

- CONVENTIONS.md
- tools/dedup_persons.py
- tools/extract_entities.py
- tools/migrate_external_orgs.py
- tools/normalize_person_frontmatter.py
- tools/person_datapoints.py

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

tools/person_datapoints.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python3
+"""Reporte de datapoints y score de fiabilidad/completitud por persona en osint.
+
+Para cada ficha personas/<slug>.md calcula:
+  - score de completitud: campos de identidad presentes / 7 * 100
+    (sexo, fecha_nacimiento, dni, telefono, email, direccion, pais)
+  - datapoints totales: campos de identidad presentes + nº documentos + nº attachments + relaciones
+  - campos faltantes (cuando el score < 100%)
+
+Salida: tabla ordenada por score asc (las menos fiables primero) + totales globales.
+Con --json imprime el detalle como JSON. Read-only.
+"""
+import sys, os, glob, json
+
+sys.path.insert(0, "/home/enmanuel/fn_registry/python/functions")
+from obsidian import read_obsidian_note
+
+OSINT = "/home/enmanuel/Obsidian/osint"
+IDENT = ["sexo", "fecha_nacimiento", "dni", "telefono", "email", "direccion", "pais"]
+
+
+def main():
+    as_json = "--json" in sys.argv
+    rows = []
+    tot_dp = 0
+    for fp in sorted(glob.glob(f"{OSINT}/personas/*.md")):
+        slug = os.path.splitext(os.path.basename(fp))[0]
+        if slug.startswith("_"):
+            continue
+        fm = read_obsidian_note(fp)["frontmatter"]
+        present = [k for k in IDENT if fm.get(k) not in (None, "", [])]
+        missing = [k for k in IDENT if k not in present]
+        score = round(len(present) / len(IDENT) * 100)
+        ndocs = len(glob.glob(f"{OSINT}/personas/{slug}/*.md"))
+        natt = len(glob.glob(f"{OSINT}/attachments/personas/{slug}/*"))
+        nrel = len(fm.get("relaciones") or [])
+        dp = len(present) + ndocs + natt + nrel
+        tot_dp += dp
+        rows.append({"slug": slug, "score": score, "datapoints": dp,
+                     "ident": len(present), "docs": ndocs, "attachments": natt,
+                     "relaciones": nrel, "faltan": missing})
+
+    rows.sort(key=lambda r: (r["score"], -r["datapoints"]))
+    if as_json:
+        print(json.dumps({"total_datapoints": tot_dp, "personas": rows}, ensure_ascii=False, indent=2))
+        return
+
+    print(f"PERSONAS: {len(rows)} | datapoints totales: {tot_dp} | "
+          f"score medio: {round(sum(r['score'] for r in rows)/len(rows))}%\n")
+    print(f"{'persona':38} {'score':>5} {'dp':>4} {'id':>3} {'doc':>4} {'att':>4} {'rel':>4}  faltan")
+    print("-" * 100)
+    for r in rows:
+        flag = "" if r["score"] == 100 else "  <-- " + ",".join(r["faltan"])
+        print(f"{r['slug']:38} {r['score']:>4}% {r['datapoints']:>4} {r['ident']:>3} "
+              f"{r['docs']:>4} {r['attachments']:>4} {r['relaciones']:>4}{flag}")
+    bajo = [r for r in rows if r["score"] < 100]
+    print(f"\nfichas por debajo del 100%: {len(bajo)}/{len(rows)} "
+          f"({round(len(bajo)/len(rows)*100)}%)")
+
+
+if __name__ == "__main__":
+    main()