Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,4 +1,8 @@
|
|
| 1 |
-
import
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import requests
|
| 3 |
from bs4 import BeautifulSoup
|
| 4 |
from urllib.parse import urljoin, urlparse
|
|
@@ -13,7 +17,13 @@ import sqlite3
|
|
| 13 |
from contextlib import contextmanager
|
| 14 |
from threading import local
|
| 15 |
import time
|
|
|
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
logging.basicConfig(level=logging.INFO)
|
| 18 |
logger = logging.getLogger(__name__)
|
| 19 |
|
|
@@ -166,19 +176,68 @@ async def process_url(url, depth):
|
|
| 166 |
logger.error(f"Error in process_url: {str(e)}")
|
| 167 |
return f"An error occurred: {str(e)}"
|
| 168 |
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 181 |
)
|
| 182 |
-
|
| 183 |
-
if
|
| 184 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import dash
|
| 2 |
+
from dash import dcc, html, Input, Output, State
|
| 3 |
+
import dash_bootstrap_components as dbc
|
| 4 |
+
from dash.exceptions import PreventUpdate
|
| 5 |
+
import base64
|
| 6 |
import requests
|
| 7 |
from bs4 import BeautifulSoup
|
| 8 |
from urllib.parse import urljoin, urlparse
|
|
|
|
| 17 |
from contextlib import contextmanager
|
| 18 |
from threading import local
|
| 19 |
import time
|
| 20 |
+
import os
|
| 21 |
|
| 22 |
+
# Initialize Dash app
|
| 23 |
+
app = dash.Dash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])
|
| 24 |
+
server = app.server
|
| 25 |
+
|
| 26 |
+
# Logging setup
|
| 27 |
logging.basicConfig(level=logging.INFO)
|
| 28 |
logger = logging.getLogger(__name__)
|
| 29 |
|
|
|
|
| 176 |
logger.error(f"Error in process_url: {str(e)}")
|
| 177 |
return f"An error occurred: {str(e)}"
|
| 178 |
|
| 179 |
+
# App layout
|
| 180 |
+
app.layout = dbc.Container([
|
| 181 |
+
dbc.Navbar(
|
| 182 |
+
dbc.Container([
|
| 183 |
+
html.A(
|
| 184 |
+
dbc.Row([
|
| 185 |
+
dbc.Col(html.Img(src="/assets/logo.png", height="30px")),
|
| 186 |
+
dbc.Col(dbc.NavbarBrand("Website to PDF Converter", className="ms-2")),
|
| 187 |
+
],
|
| 188 |
+
align="center",
|
| 189 |
+
className="g-0",
|
| 190 |
+
),
|
| 191 |
+
href="/",
|
| 192 |
+
style={"textDecoration": "none"},
|
| 193 |
+
)
|
| 194 |
+
]),
|
| 195 |
+
color="#116F70",
|
| 196 |
+
dark=True,
|
| 197 |
+
),
|
| 198 |
+
|
| 199 |
+
dbc.Card(
|
| 200 |
+
dbc.CardBody([
|
| 201 |
+
html.H1("Website to PDF Converter", className="text-center mb-4"),
|
| 202 |
+
html.P("Enter docs URL and crawl depth to convert documentation pages into a PDF. Be responsible for sites you have permission to do this", className="text-center mb-4"),
|
| 203 |
+
dbc.Input(id="url-input", type="text", placeholder="Enter website URL (e.g., https://www.gradio.app/docs)", className="mb-3"),
|
| 204 |
+
dcc.Slider(id="depth-slider", min=1, max=10, step=1, value=3, marks={i: str(i) for i in range(1, 11)}, className="mb-3"),
|
| 205 |
+
dbc.Button("Convert to PDF", id="submit-button", color="primary", className="mb-3 w-100"),
|
| 206 |
+
dbc.Spinner(html.Div(id="output-area"), color="primary", type="grow"),
|
| 207 |
+
]),
|
| 208 |
+
className="mt-4"
|
| 209 |
+
)
|
| 210 |
+
], fluid=True)
|
| 211 |
+
|
| 212 |
+
@app.callback(
|
| 213 |
+
Output("output-area", "children"),
|
| 214 |
+
Input("submit-button", "n_clicks"),
|
| 215 |
+
State("url-input", "value"),
|
| 216 |
+
State("depth-slider", "value"),
|
| 217 |
+
prevent_initial_call=True
|
| 218 |
)
|
| 219 |
+
def update_output(n_clicks, url, depth):
|
| 220 |
+
if not url:
|
| 221 |
+
return "Please enter a valid URL."
|
| 222 |
+
|
| 223 |
+
pdf_path = asyncio.run(process_url(url, depth))
|
| 224 |
+
|
| 225 |
+
if pdf_path.startswith("An error occurred"):
|
| 226 |
+
return pdf_path
|
| 227 |
+
|
| 228 |
+
with open(pdf_path, "rb") as f:
|
| 229 |
+
encoded = base64.b64encode(f.read()).decode()
|
| 230 |
+
|
| 231 |
+
os.unlink(pdf_path) # Remove the temporary file
|
| 232 |
+
|
| 233 |
+
return html.Div([
|
| 234 |
+
html.H4("PDF Generated Successfully"),
|
| 235 |
+
html.A(
|
| 236 |
+
dbc.Button("Download PDF", color="success", className="mt-2"),
|
| 237 |
+
href=f"data:application/pdf;base64,{encoded}",
|
| 238 |
+
download="website_content.pdf"
|
| 239 |
+
)
|
| 240 |
+
])
|
| 241 |
+
|
| 242 |
+
if __name__ == '__main__':
|
| 243 |
+
app.run_server(debug=True, host='0.0.0.0', port=7860)
|