rss/app.py
2025-05-29 09:41:51 +02:00

379 lines
14 KiB
Python

from flask import Flask, render_template, request, redirect, url_for, Response
from apscheduler.schedulers.background import BackgroundScheduler
from datetime import datetime
import feedparser
import hashlib
import re
import mysql.connector
import csv
from io import StringIO
app = Flask(__name__)
DB_CONFIG = {
'host': 'localhost',
'user': 'x',
'password': 'x',
'database': 'noticiasrss'
}
MAX_FALLOS = 5 # Número máximo de fallos antes de desactivar el feed
# ======================================
# Página principal: últimas noticias
# ======================================
@app.route('/')
def home():
conn = None
noticias = []
categorias = []
continentes = []
paises = []
cat_id = request.args.get('categoria_id')
cont_id = request.args.get('continente_id')
pais_id = request.args.get('pais_id')
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
cursor.execute("SELECT id, nombre FROM categorias_estandar ORDER BY nombre")
categorias = cursor.fetchall()
cursor.execute("SELECT id, nombre FROM continentes ORDER BY nombre")
continentes = cursor.fetchall()
if cont_id:
cursor.execute("SELECT id, nombre, continente_id FROM paises WHERE continente_id = %s ORDER BY nombre", (cont_id,))
else:
cursor.execute("SELECT id, nombre, continente_id FROM paises ORDER BY nombre")
paises = cursor.fetchall()
sql = """
SELECT n.fecha, n.titulo, n.resumen, n.url, n.imagen_url,
c.nombre AS categoria, p.nombre AS pais, co.nombre AS continente
FROM noticias n
LEFT JOIN categorias_estandar c ON n.categoria_id = c.id
LEFT JOIN paises p ON n.pais_id = p.id
LEFT JOIN continentes co ON p.continente_id = co.id
WHERE 1=1
"""
params = []
if cat_id:
sql += " AND n.categoria_id = %s"
params.append(cat_id)
if pais_id:
sql += " AND n.pais_id = %s"
params.append(pais_id)
elif cont_id:
sql += " AND p.continente_id = %s"
params.append(cont_id)
sql += " ORDER BY n.fecha DESC LIMIT 50"
cursor.execute(sql, params)
noticias = cursor.fetchall()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al leer noticias: {db_err}", exc_info=True)
finally:
if conn:
conn.close()
return render_template(
'noticias.html',
noticias=noticias,
categorias=categorias,
continentes=continentes,
paises=paises,
cat_id=int(cat_id) if cat_id else None,
cont_id=int(cont_id) if cont_id else None,
pais_id=int(pais_id) if pais_id else None
)
# ======================================
# Gestión de feeds en /feeds
# ======================================
@app.route('/feeds')
def feeds():
conn = None
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
# Feeds con descripción y fallos
cursor.execute("""
SELECT f.id, f.nombre, f.descripcion, f.url, f.categoria_id, f.pais_id, f.activo, f.fallos, c.nombre, p.nombre
FROM feeds f
LEFT JOIN categorias_estandar c ON f.categoria_id = c.id
LEFT JOIN paises p ON f.pais_id = p.id
""")
feeds = cursor.fetchall()
cursor.execute("SELECT id, nombre FROM categorias_estandar ORDER BY nombre")
categorias = cursor.fetchall()
cursor.execute("SELECT id, nombre FROM continentes ORDER BY nombre")
continentes = cursor.fetchall()
cursor.execute("SELECT id, nombre, continente_id FROM paises ORDER BY nombre")
paises = cursor.fetchall()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al leer feeds/categorías/países: {db_err}", exc_info=True)
feeds, categorias, continentes, paises = [], [], [], []
finally:
if conn:
conn.close()
return render_template("index.html", feeds=feeds, categorias=categorias, continentes=continentes, paises=paises)
# Añadir feed
@app.route('/add', methods=['POST'])
def add_feed():
nombre = request.form.get('nombre')
descripcion = request.form.get('descripcion')
url = request.form.get('url')
categoria_id = request.form.get('categoria_id')
pais_id = request.form.get('pais_id')
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
cursor.execute(
"INSERT INTO feeds (nombre, descripcion, url, categoria_id, pais_id) VALUES (%s, %s, %s, %s, %s)",
(nombre, descripcion, url, categoria_id, pais_id)
)
conn.commit()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al agregar feed: {db_err}", exc_info=True)
finally:
if conn:
conn.close()
return redirect(url_for('feeds'))
# Editar feed
@app.route('/edit/<int:feed_id>', methods=['GET', 'POST'])
def edit_feed(feed_id):
conn = None
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor(dictionary=True)
if request.method == 'POST':
nombre = request.form.get('nombre')
descripcion = request.form.get('descripcion')
url_feed = request.form.get('url')
categoria_id = request.form.get('categoria_id')
pais_id = request.form.get('pais_id')
activo = 1 if request.form.get('activo') == 'on' else 0
cursor.execute(
"UPDATE feeds SET nombre=%s, descripcion=%s, url=%s, categoria_id=%s, pais_id=%s, activo=%s WHERE id=%s",
(nombre, descripcion, url_feed, categoria_id, pais_id, activo, feed_id)
)
conn.commit()
return redirect(url_for('feeds'))
cursor.execute("SELECT * FROM feeds WHERE id = %s", (feed_id,))
feed = cursor.fetchone()
cursor.execute("SELECT id, nombre FROM categorias_estandar ORDER BY nombre")
categorias = cursor.fetchall()
cursor.execute("SELECT id, nombre FROM paises ORDER BY nombre")
paises = cursor.fetchall()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al editar feed: {db_err}", exc_info=True)
feed, categorias, paises = {}, [], []
finally:
if conn:
conn.close()
return render_template('edit_feed.html', feed=feed, categorias=categorias, paises=paises)
# Eliminar feed
@app.route('/delete/<int:feed_id>')
def delete_feed(feed_id):
conn = None
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
cursor.execute("DELETE FROM feeds WHERE id=%s", (feed_id,))
conn.commit()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al eliminar feed: {db_err}", exc_info=True)
finally:
if conn:
conn.close()
return redirect(url_for('feeds'))
# Backup de feeds a CSV
@app.route('/backup_feeds')
def backup_feeds():
conn = None
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
cursor.execute("""
SELECT f.id, f.nombre, f.descripcion, f.url, f.categoria_id, c.nombre AS categoria, f.pais_id, p.nombre AS pais, f.activo, f.fallos
FROM feeds f
LEFT JOIN categorias_estandar c ON f.categoria_id = c.id
LEFT JOIN paises p ON f.pais_id = p.id
""")
feeds = cursor.fetchall()
header = [desc[0] for desc in cursor.description]
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] Al hacer backup de feeds: {db_err}", exc_info=True)
return "Error generando backup.", 500
finally:
if conn:
conn.close()
si = StringIO()
cw = csv.writer(si)
cw.writerow(header)
cw.writerows(feeds)
output = si.getvalue()
si.close()
return Response(
output,
mimetype="text/csv",
headers={"Content-Disposition": "attachment;filename=feeds_backup.csv"}
)
# Restaurar feeds desde CSV
@app.route('/restore_feeds', methods=['GET', 'POST'])
def restore_feeds():
msg = ""
if request.method == 'POST':
file = request.files.get('file')
if not file or not file.filename.endswith('.csv'):
msg = "Archivo no válido."
else:
file_stream = StringIO(file.read().decode('utf-8'))
reader = csv.DictReader(file_stream)
rows = list(reader)
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
n_ok = 0
for row in rows:
try:
descripcion = row.get('descripcion') or ""
cursor.execute("""
INSERT INTO feeds (nombre, descripcion, url, categoria_id, pais_id, activo, fallos)
VALUES (%s, %s, %s, %s, %s, %s, %s)
ON DUPLICATE KEY UPDATE
nombre=VALUES(nombre),
descripcion=VALUES(descripcion),
url=VALUES(url),
categoria_id=VALUES(categoria_id),
pais_id=VALUES(pais_id),
activo=VALUES(activo),
fallos=VALUES(fallos)
""", (
row['nombre'],
descripcion,
row['url'],
row['categoria_id'],
row['pais_id'],
int(row['activo']),
int(row.get('fallos', 0))
))
n_ok += 1
except Exception as e:
app.logger.error(f"Error insertando feed {row}: {e}")
conn.commit()
conn.close()
msg = f"Feeds restaurados correctamente: {n_ok}"
return render_template("restore_feeds.html", msg=msg)
@app.route('/noticias')
def show_noticias():
return home()
# ================================
# Lógica de procesado de feeds con control de fallos
# ================================
def sumar_fallo_feed(cursor, feed_id):
cursor.execute("UPDATE feeds SET fallos = fallos + 1 WHERE id = %s", (feed_id,))
cursor.execute("SELECT fallos FROM feeds WHERE id = %s", (feed_id,))
fallos = cursor.fetchone()[0]
if fallos >= MAX_FALLOS:
cursor.execute("UPDATE feeds SET activo = 0 WHERE id = %s", (feed_id,))
return fallos
def resetear_fallos_feed(cursor, feed_id):
cursor.execute("UPDATE feeds SET fallos = 0 WHERE id = %s", (feed_id,))
def fetch_and_store():
conn = None
try:
conn = mysql.connector.connect(**DB_CONFIG)
cursor = conn.cursor()
cursor.execute("SELECT id, url, categoria_id, pais_id FROM feeds WHERE activo = TRUE")
feeds = cursor.fetchall()
except mysql.connector.Error as db_err:
app.logger.error(f"[DB ERROR] No se pudo conectar o leer feeds: {db_err}", exc_info=True)
return
for feed_id, rss_url, categoria_id, pais_id in feeds:
try:
app.logger.info(f"Procesando feed: {rss_url} [{categoria_id}] [{pais_id}]")
parsed = feedparser.parse(rss_url)
except Exception as e:
app.logger.error(f"[PARSE ERROR] Al parsear {rss_url}: {e}", exc_info=True)
sumar_fallo_feed(cursor, feed_id)
continue
if getattr(parsed, 'bozo', False):
bozo_exc = getattr(parsed, 'bozo_exception', 'Unknown')
app.logger.warning(f"[BOZO] Feed mal formado: {rss_url} - {bozo_exc}")
sumar_fallo_feed(cursor, feed_id)
continue
else:
resetear_fallos_feed(cursor, feed_id)
for entry in parsed.entries:
link = entry.get('link') or entry.get('id')
if not link:
links_list = entry.get('links', [])
if isinstance(links_list, list) and links_list:
href = next((l.get('href') for l in links_list if l.get('href')), None)
link = href
if not link:
app.logger.error(f"[ENTRY ERROR] Entrada sin link en feed {rss_url}, salto entrada.")
continue
try:
noticia_id = hashlib.md5(link.encode()).hexdigest()
titulo = entry.get('title', '')
resumen = entry.get('summary', '')
imagen_url = ''
fecha = None
if 'media_content' in entry:
imagen_url = entry.media_content[0].get('url', '')
else:
img = re.search(r'<img.+?src="(.+?)"', resumen)
if img:
imagen_url = img.group(1)
published = entry.get('published_parsed') or entry.get('updated_parsed')
if published:
try:
fecha = datetime(*published[:6])
except Exception:
fecha = None
cursor.execute(
"""
INSERT IGNORE INTO noticias
(id, titulo, resumen, url, fecha, imagen_url, categoria_id, pais_id)
VALUES (%s, %s, %s, %s, %s, %s, %s, %s)
""",
(noticia_id, titulo, resumen, link, fecha, imagen_url, categoria_id, pais_id)
)
except Exception as entry_err:
app.logger.error(
f"[ENTRY ERROR] Falló procesar entrada {link}: {entry_err}", exc_info=True
)
continue
try:
conn.commit()
except Exception as commit_err:
app.logger.error(f"[DB ERROR] Al confirmar transacción: {commit_err}", exc_info=True)
finally:
if conn:
conn.close()
app.logger.info(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Feeds procesados.")
if __name__ == '__main__':
scheduler = BackgroundScheduler()
scheduler.add_job(fetch_and_store, 'interval', minutes=2, id='rss_job')
scheduler.start()
try:
app.run(host="0.0.0.0", port=5000, debug=True)
finally:
scheduler.shutdown()