Files
feadulta/scripts/import_new_content.py

255 lines
9.0 KiB
Python

#!/usr/bin/env python3
"""
import_new_content.py
Importa los ew4r_content items no-carta nuevos (id > 9043, catid NOT IN 27,40,41)
al WordPress local (Docker).
Mapping catid → WP term_ids:
54 (Índice multimedia) → 26
77 (Videos) → 58
64 (Noticias de alcance) → 41
52 (Tablón de anuncios) → 1 (uncategorized / sin categoría)
63 (Fechas) → 40
61 (Lista completa de autores) → 38
65 (Cantoral Salomé Arricibita) → 31
otro → 1 (uncategorized)
"""
import subprocess
import sys
from datetime import datetime
JOOMLA_SSH_HOST = "134.0.10.170"
JOOMLA_SSH_USER = "feadulta"
JOOMLA_SSH_PASS = "6Rm2qOF@eundwpda"
JOOMLA_DB_HOST = "127.0.0.1"
JOOMLA_DB_USER = "fejoomla3"
JOOMLA_DB_PASS = "5FF-}5^[>7^pK4W9"
JOOMLA_DB_NAME = "fejoomla3"
WP_DOCKER = "wordpress-mysql"
WP_DB_USER = "wordpress_user"
WP_DB_PASS = "wordpress_pass"
WP_DB_NAME = "wordpress_db"
LAST_CONTENT_ID = 9043
CARTA_CATIDS = {27, 40, 41}
CATID_TO_WP = {
54: [26],
77: [58],
64: [41],
52: [1],
63: [40],
61: [38],
65: [31],
}
DRY_RUN = '--dry-run' in sys.argv
# ── Helpers ────────────────────────────────────────────────────────────────────
def joomla_query(query: str) -> list[dict]:
mysql_cmd = (f"mysql --skip-ssl -h {JOOMLA_DB_HOST} -u {JOOMLA_DB_USER} "
f"-p'{JOOMLA_DB_PASS}' {JOOMLA_DB_NAME} "
f"--default-character-set=utf8mb4 -B")
cmd = ['sshpass', '-p', JOOMLA_SSH_PASS,
'ssh', f'{JOOMLA_SSH_USER}@{JOOMLA_SSH_HOST}', mysql_cmd]
result = subprocess.run(cmd, input=query, capture_output=True,
text=True, encoding='utf-8')
if result.returncode != 0:
print(f"[ERR SSH] {result.stderr[:300]}", file=sys.stderr)
return []
lines = result.stdout.strip().split('\n')
if len(lines) < 2:
return []
headers = lines[0].split('\t')
return [dict(zip(headers, line.split('\t'))) for line in lines[1:] if line]
def wp_mysql(query: str) -> list[dict]:
cmd = ['docker', 'exec', WP_DOCKER,
'mysql', '-u', WP_DB_USER, f'-p{WP_DB_PASS}', WP_DB_NAME,
'--default-character-set=utf8mb4', '-B', '-e', query]
result = subprocess.run(cmd, capture_output=True, text=True, encoding='utf-8')
if result.returncode != 0:
return []
lines = result.stdout.strip().split('\n')
if len(lines) < 2:
return []
headers = lines[0].split('\t')
return [dict(zip(headers, line.split('\t'))) for line in lines[1:] if line]
def wp_execute(sql: str):
if DRY_RUN:
print(f" [DRY] {sql[:110]}")
return
cmd = ['docker', 'exec', WP_DOCKER,
'mysql', '-u', WP_DB_USER, f'-p{WP_DB_PASS}', WP_DB_NAME,
'--default-character-set=utf8mb4', '-e', sql]
result = subprocess.run(cmd, capture_output=True, text=True)
if result.returncode != 0:
err = result.stderr.replace('mysql: [Warning] Using a password on the command line interface can be insecure.\n', '')
if err.strip():
print(f" [ERR] {err.strip()[:200]}", file=sys.stderr)
def esc(s: str) -> str:
return s.replace('\\', '\\\\').replace("'", "\\'")
def unhex(val: str) -> str:
if not val or val == 'NULL':
return ''
try:
return bytes.fromhex(val).decode('utf-8', errors='replace')
except Exception:
return val
def main():
print(f"=== Import ew4r_content no-cartas (id > {LAST_CONTENT_ID}) "
f"{'[DRY RUN]' if DRY_RUN else '[LIVE]'} ===\n")
# Cargar user map
user_rows = wp_mysql(
"SELECT um.meta_value jid, u.ID wid FROM wp_users u "
"JOIN wp_usermeta um ON um.user_id=u.ID "
"WHERE um.meta_key='_fgj2wp_old_user_id'"
)
user_map = {}
for r in user_rows:
try:
user_map[int(r['jid'])] = int(r['wid'])
except ValueError:
pass
print(f"Usuarios mapeados: {len(user_map)}")
# Cargar term_taxonomy_ids
all_term_ids = sorted({t for cats in CATID_TO_WP.values() for t in cats} | {1})
rows = wp_mysql(
f"SELECT term_id, term_taxonomy_id FROM wp_term_taxonomy "
f"WHERE term_id IN ({','.join(map(str,all_term_ids))}) AND taxonomy='category'"
)
tt_ids = {int(r['term_id']): int(r['term_taxonomy_id']) for r in rows}
print(f"TT IDs: {tt_ids}")
# Polylang ES
pl_rows = wp_mysql(
"SELECT tt.term_taxonomy_id FROM wp_terms t "
"JOIN wp_term_taxonomy tt ON tt.term_id=t.term_id "
"WHERE tt.taxonomy='language' AND t.slug='es' LIMIT 1"
)
pl_es_tt = int(pl_rows[0]['term_taxonomy_id']) if pl_rows else None
# IDs ya en WP
existing_rows = wp_mysql(
f"SELECT meta_value FROM wp_postmeta "
f"WHERE meta_key='_fgj2wp_old_content_id' AND meta_value+0 > {LAST_CONTENT_ID}"
)
existing_ids = {int(r['meta_value']) for r in existing_rows}
print(f"IDs ya importados con id > {LAST_CONTENT_ID}: {len(existing_ids)}")
# Obtener items de Joomla
print("\nObteniendo items de Joomla...")
catids_excl = ','.join(str(c) for c in CARTA_CATIDS)
query = (
f"SELECT id, HEX(title) title, HEX(alias) alias, "
f"HEX(introtext) introtext, HEX(`fulltext`) fulltext_col, "
f"catid, created, created_by "
f"FROM ew4r_content "
f"WHERE state=1 AND id > {LAST_CONTENT_ID} AND catid NOT IN ({catids_excl}) "
f"ORDER BY id;"
)
items = joomla_query(query)
print(f"Items a importar: {len(items)}")
stats = {'ok': 0, 'skip': 0, 'err': 0}
for item in items:
joomla_id = int(item['id'])
catid = int(item['catid'])
title = unhex(item.get('title', ''))
alias = unhex(item.get('alias', ''))
intro = unhex(item.get('introtext', ''))
full = unhex(item.get('fulltext_col', ''))
created = item.get('created', '') or datetime.now().strftime('%Y-%m-%d %H:%M:%S')
created_by = int(item.get('created_by', 0) or 0)
if joomla_id in existing_ids:
print(f" [SKIP] id={joomla_id} ya existe")
stats['skip'] += 1
continue
content = intro + ('\n<!--more-->\n' + full if full.strip() else '')
# Multimedia/pensamientos/vídeos/cantoral (catid 54/77/65) son contenido
# propio de FeAdulta → autor "Fe Adulta" (WP user 890), no el webmaster
# que los sube en Joomla. El resto conserva su autor real (noticias, etc.).
FEADULTA_AUTHOR = 890
FEADULTA_CATIDS = {54, 77, 65}
wp_author = FEADULTA_AUTHOR if catid in FEADULTA_CATIDS else user_map.get(created_by, 1)
wp_cats = CATID_TO_WP.get(catid, [1])
print(f" [{joomla_id}] catid={catid} | {title[:50]}")
wp_execute(
f"INSERT INTO wp_posts "
f"(post_author, post_date, post_date_gmt, post_content, post_title, "
f"post_excerpt, post_status, comment_status, ping_status, post_name, "
f"post_type, post_modified, post_modified_gmt, comment_count, "
f"to_ping, pinged, post_content_filtered) VALUES ("
f"{wp_author}, '{created}', '{created}', '{esc(content)}', "
f"'{esc(title)}', '', 'publish', 'open', 'open', '{esc(alias[:200])}', "
f"'post', '{created}', '{created}', 0, '', '', '')"
)
if DRY_RUN:
stats['ok'] += 1
continue
new_id_rows = wp_mysql("SELECT MAX(ID) new_id FROM wp_posts")
if not new_id_rows:
stats['err'] += 1
continue
new_wp_id = int(new_id_rows[0]['new_id'])
print(f" → WP ID={new_wp_id}")
# Metas
wp_execute(f"INSERT INTO wp_postmeta (post_id, meta_key, meta_value) VALUES ({new_wp_id}, '_fgj2wp_old_content_id', '{joomla_id}')")
# Categorías
for term_id in wp_cats:
tt_id = tt_ids.get(term_id)
if tt_id:
wp_execute(
f"INSERT IGNORE INTO wp_term_relationships (object_id, term_taxonomy_id) "
f"VALUES ({new_wp_id}, {tt_id})"
)
# Polylang ES
if pl_es_tt:
wp_execute(
f"INSERT IGNORE INTO wp_term_relationships (object_id, term_taxonomy_id) "
f"VALUES ({new_wp_id}, {pl_es_tt})"
)
stats['ok'] += 1
if not DRY_RUN and stats['ok'] > 0:
print("\nActualizando counts de categorías...")
tt_str = ','.join(str(v) for v in tt_ids.values())
wp_execute(
f"UPDATE wp_term_taxonomy tt SET count = ("
f"SELECT COUNT(*) FROM wp_term_relationships tr "
f"WHERE tr.term_taxonomy_id=tt.term_taxonomy_id"
f") WHERE tt.term_taxonomy_id IN ({tt_str})"
)
print(f"\n=== Resultado: {stats['ok']} ok, {stats['skip']} skip, {stats['err']} err ===")
if __name__ == '__main__':
main()