#!/bin/bash
# uso: q.sh sql/fichero.sql [nombre] [salida.md] -> DuckDB (imagen etl:v2), datos :ro, -m 8g, sin red, timeout.
# Vistas: u = unified_20261018; idn/* identidad; p8 y actas montados :ro. Escribe solo en colusion/ (montado en /c).
# Sentencias separadas por ';\n--;'. Si se da salida, se escribe además en ese fichero.
set -u
C=<mediciones>/colusion
SQL="/c/$1"; N="colusion-${2:-sql}"; OUT="${3:-}"
AV=$(free -g | awk '/Mem:/{print $7}'); while [ "$AV" -lt 12 ]; do echo "available=${AV}G < 12G: espero" >&2; sleep 30; AV=$(free -g | awk '/Mem:/{print $7}'); done
timeout ${TMO:-3000} docker run -i --rm -m 8g --memory-swap 8g --cpus 4 --network none --user 1000:1000 --name "$N" \
  -v <datos>/:/d:ro -v <mediciones>/p8:/p8:ro \
  -v <mediciones>/actas:/actas:ro -v $C:/c \
  buscalicitaciones-etl:v2 python3 - "$SQL" <<'PY' | if [ -n "$OUT" ]; then tee "$C/$OUT"; else cat; fi
import sys, duckdb, time
sql = open(sys.argv[1]).read()
con = duckdb.connect('/c/interno/col.duckdb')
con.execute("SET memory_limit='6GB'; SET threads=4; SET temp_directory='/c/tmp/duck'; SET max_temp_directory_size='30GB'; SET preserve_insertion_order=false;")
con.execute("CREATE OR REPLACE TEMP VIEW u AS SELECT * FROM read_parquet('/d/unified_20261018/*/*/*.parquet', hive_partitioning=1, union_by_name=1)")
for t in ['adjudicatarios_filas','organos_filas','organos_tipo','ute_socios','empresas','organos']:
    con.execute(f"CREATE OR REPLACE TEMP VIEW idn_{t} AS SELECT * FROM read_parquet('/d/unified_20261018.identidad/{t}.parquet')")
con.execute("CREATE OR REPLACE TEMP VIEW vf AS SELECT * FROM read_parquet('/d/unified_20261018.vinculos/filas.parquet')")
print(f"DuckDB {duckdb.__version__} · {time.strftime('%Y-%m-%d %H:%M:%S')}\n")
for stmt in [s for s in sql.split(';\n--;') if s.strip()]:
    t=time.time()
    try:
        r = con.execute(stmt)
        if r.description:
            cols=[d[0] for d in r.description]; rows=r.fetchall()
            tit=[l for l in stmt.strip().splitlines() if l.startswith('-- @')]
            print('\n## '+(tit[0][4:].strip() if tit else stmt.strip().splitlines()[0][:110]))
            print('\n| '+' | '.join(cols)+' |\n|'+'|'.join('---' for _ in cols)+'|')
            for row in rows[:500]: print('| '+' | '.join('' if v is None else (f'{v:.4g}' if isinstance(v,float) else str(v)) for v in row)+' |')
            if len(rows)>500: print(f'... {len(rows)} filas')
    except Exception as e:
        print('## ERROR en:', stmt.strip().splitlines()[0][:110]); print(e)
    print(f'\n_{time.time()-t:.1f} s_', flush=True)
PY