Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -3,21 +3,21 @@ from datetime import datetime, timedelta
|
|
| 3 |
import requests
|
| 4 |
import json
|
| 5 |
|
| 6 |
-
def wayback(website
|
| 7 |
-
limit = 3000 if limit is None else limit
|
| 8 |
-
|
| 9 |
if not website:
|
| 10 |
return '😭 Error: introduce una url correcta'
|
| 11 |
|
|
|
|
| 12 |
end_date = datetime.now()
|
| 13 |
start_date = end_date - timedelta(days=365)
|
| 14 |
datefrom = start_date.strftime('%Y%m%d')
|
| 15 |
dateto = end_date.strftime('%Y%m%d')
|
| 16 |
|
| 17 |
-
|
|
|
|
| 18 |
|
| 19 |
try:
|
| 20 |
-
response = requests.get(f"http://web.archive.org/cdx/search/cdx{query}
|
| 21 |
if response.status_code != 200:
|
| 22 |
return f'😭 Error: Respuesta no exitosa con código de estado {response.status_code}'
|
| 23 |
content = json.loads(response.text)
|
|
@@ -28,6 +28,7 @@ def wayback(website, limit=None):
|
|
| 28 |
except Exception as e:
|
| 29 |
return f"😭 Error: {e}"
|
| 30 |
|
|
|
|
| 31 |
results = []
|
| 32 |
for row in content[1:]:
|
| 33 |
date, page, status = [row[i] for i in [1, 2, 4]]
|
|
@@ -36,58 +37,14 @@ def wayback(website, limit=None):
|
|
| 36 |
results.append(f"👓 {formatted_date} {formatted_wayback_url}")
|
| 37 |
return "\n".join(results)
|
| 38 |
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
|
|
|
|
|
|
|
|
|
| 44 |
|
|
|
|
| 45 |
iface.launch()
|
| 46 |
-
|
| 47 |
-
mejor deja simplemente el de meter la url
|
| 48 |
-
|
| 49 |
-
import gradio as gr
|
| 50 |
-
from datetime import datetime, timedelta
|
| 51 |
-
import requests
|
| 52 |
-
import json
|
| 53 |
-
|
| 54 |
-
def wayback(website, limit=None):
|
| 55 |
-
limit = 3000 if limit is None else limit
|
| 56 |
-
|
| 57 |
-
if not website:
|
| 58 |
-
return '😭 Error: introduce una url correcta'
|
| 59 |
-
|
| 60 |
-
end_date = datetime.now()
|
| 61 |
-
start_date = end_date - timedelta(days=365)
|
| 62 |
-
datefrom = start_date.strftime('%Y%m%d')
|
| 63 |
-
dateto = end_date.strftime('%Y%m%d')
|
| 64 |
-
|
| 65 |
-
query = f"?url={website}&output=json&from={datefrom}&to={dateto}"
|
| 66 |
-
|
| 67 |
-
try:
|
| 68 |
-
response = requests.get(f"http://web.archive.org/cdx/search/cdx{query}&limit={limit}")
|
| 69 |
-
if response.status_code != 200:
|
| 70 |
-
return f'😭 Error: Respuesta no exitosa con código de estado {response.status_code}'
|
| 71 |
-
content = json.loads(response.text)
|
| 72 |
-
if len(content) <= 1:
|
| 73 |
-
return '😭 Error: no hay datos para esta página web'
|
| 74 |
-
except json.JSONDecodeError as e:
|
| 75 |
-
return f'😭 Error al analizar JSON: {e}'
|
| 76 |
-
except Exception as e:
|
| 77 |
-
return f"😭 Error: {e}"
|
| 78 |
-
|
| 79 |
-
results = []
|
| 80 |
-
for row in content[1:]:
|
| 81 |
-
date, page, status = [row[i] for i in [1, 2, 4]]
|
| 82 |
-
formatted_date = datetime.strptime(date, '%Y%m%d%H%M%S').strftime('%d/%m/%Y')
|
| 83 |
-
formatted_wayback_url = f"https://web.archive.org/web/{date}/{page}"
|
| 84 |
-
results.append(f"👓 {formatted_date} {formatted_wayback_url}")
|
| 85 |
-
return "\n".join(results)
|
| 86 |
-
|
| 87 |
-
iface = gr.Interface(fn=wayback,
|
| 88 |
-
inputs=["text", gr.Number(label="Límite", value=3000)],
|
| 89 |
-
outputs="text",
|
| 90 |
-
title="Wayback Machine Lookup",
|
| 91 |
-
description="Busca instantáneas de una página web en la Wayback Machine.")
|
| 92 |
-
|
| 93 |
-
iface.launch()
|
|
|
|
| 3 |
import requests
|
| 4 |
import json
|
| 5 |
|
| 6 |
+
def wayback(website):
|
|
|
|
|
|
|
| 7 |
if not website:
|
| 8 |
return '😭 Error: introduce una url correcta'
|
| 9 |
|
| 10 |
+
# Rango de tiempo fijo de 365 días hacia atrás
|
| 11 |
end_date = datetime.now()
|
| 12 |
start_date = end_date - timedelta(days=365)
|
| 13 |
datefrom = start_date.strftime('%Y%m%d')
|
| 14 |
dateto = end_date.strftime('%Y%m%d')
|
| 15 |
|
| 16 |
+
# Construir la consulta a la Wayback Machine con un límite fijo de 3000 resultados
|
| 17 |
+
query = f"?url={website}&output=json&from={datefrom}&to={dateto}&limit=3000"
|
| 18 |
|
| 19 |
try:
|
| 20 |
+
response = requests.get(f"http://web.archive.org/cdx/search/cdx{query}")
|
| 21 |
if response.status_code != 200:
|
| 22 |
return f'😭 Error: Respuesta no exitosa con código de estado {response.status_code}'
|
| 23 |
content = json.loads(response.text)
|
|
|
|
| 28 |
except Exception as e:
|
| 29 |
return f"😭 Error: {e}"
|
| 30 |
|
| 31 |
+
# Procesar los resultados y devolverlos
|
| 32 |
results = []
|
| 33 |
for row in content[1:]:
|
| 34 |
date, page, status = [row[i] for i in [1, 2, 4]]
|
|
|
|
| 37 |
results.append(f"👓 {formatted_date} {formatted_wayback_url}")
|
| 38 |
return "\n".join(results)
|
| 39 |
|
| 40 |
+
# Crear la interfaz de Gradio
|
| 41 |
+
iface = gr.Interface(
|
| 42 |
+
fn=wayback,
|
| 43 |
+
inputs="text",
|
| 44 |
+
outputs="text",
|
| 45 |
+
title="Wayback Machine Lookup",
|
| 46 |
+
description="Busca instantáneas de una página web en la Wayback Machine. Introduce solo la URL."
|
| 47 |
+
)
|
| 48 |
|
| 49 |
+
# Lanzar la aplicación
|
| 50 |
iface.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|