1
0
Fork 0

some refinements

This commit is contained in:
Xavier Fontanet 2024-07-10 14:12:28 +02:00
parent 32319e1f25
commit e53bb63ff8
2 changed files with 18 additions and 10 deletions

View File

@ -8,6 +8,8 @@ import webbrowser
import mysql.connector import mysql.connector
from datetime import datetime from datetime import datetime
cerca = "altura"
cnx = mysql.connector.connect( cnx = mysql.connector.connect(
user='root', password='1234', user='root', password='1234',
host='192.168.1.22', port=33066, host='192.168.1.22', port=33066,
@ -23,16 +25,16 @@ class Anunci:
self.preu = preu_capturat self.preu = preu_capturat
def te_altura(self): def te_altura(self):
regex = ".*.tic(o|\s|\.).*" regex = ".*(À|Á|à|á|a)tic.*"
compilador = re.compile(regex, re.IGNORECASE) compilador = re.compile(regex, re.IGNORECASE)
atico = compilador.match(desc) atico = compilador.match(titol)
regex = ".*d.plex.*" regex = ".*(d|D).plex.*"
compilador = re.compile(regex, re.IGNORECASE) compilador = re.compile(regex, re.IGNORECASE)
duplex = compilador.match(desc) duplex = compilador.match(titol)
return atico or duplex return atico or duplex
def te_piscina(self): def te_piscina(self):
regex = ".*pi(c|s|z)*ina.*" regex = ".*pi(c|s|z|sc)ina.*"
compilador = re.compile(regex, re.IGNORECASE) compilador = re.compile(regex, re.IGNORECASE)
te = compilador.match(desc) te = compilador.match(desc)
return te return te
@ -41,6 +43,10 @@ class Anunci:
regex = ".*p.r(qu|k)in.*" regex = ".*p.r(qu|k)in.*"
compilador = re.compile(regex, re.IGNORECASE) compilador = re.compile(regex, re.IGNORECASE)
te = compilador.match(desc) te = compilador.match(desc)
if te == False:
regex = ".*aparca.*"
compilador = re.compile(regex, re.IGNORECASE)
te = compilador.match(desc)
return te return te
def te_traster(self): def te_traster(self):
@ -50,18 +56,17 @@ class Anunci:
return te return te
def es_chalet(self): def es_chalet(self):
regex = ".*(chalet|casa).*" regex = ".*(C|c)(halet|asa) .*"
compilador = re.compile(regex, re.IGNORECASE) compilador = re.compile(regex, re.IGNORECASE)
te = compilador.match(titol) te = compilador.match(titol)
return te return te
cerca = "parquing_altura"
headers = { headers = {
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.149 Safari/537.36" "user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.149 Safari/537.36"
} }
base_url = "https://www.departiculares.com/alquiler/barcelona&priceMax=1000" base_url = "https://www.departiculares.com/alquiler/barcelona/el-barcelones&priceMax=1200&sizeMin=60"
def tractar_anunci(anunci): def tractar_anunci(anunci):
c0 = cnx.cursor() c0 = cnx.cursor()
@ -106,7 +111,8 @@ while pagina < pagina_horitzo:
titol = anunci.titol titol = anunci.titol
desc = anunci.descripcio desc = anunci.descripcio
if anunci.te_parquing(): #if anunci.te_parquing() and anunci.te_piscina():
if anunci.te_altura():
anuncis_list.append(anunci) anuncis_list.append(anunci)
for anunci_seleccionat in anuncis_list: for anunci_seleccionat in anuncis_list:

View File

@ -15,7 +15,9 @@ c1.execute("SELECT url FROM anuncis WHERE actiu=1 AND veure_mes_tard=0")
myresult = c1.fetchall() myresult = c1.fetchall()
for x in myresult: for x in myresult:
webbrowser.open(x[0]) url = x[0]
print('opening '+url)
webbrowser.open(url)
c1.close() c1.close()