Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- app.py +3 -3
- requirements.txt +1 -1
app.py
CHANGED
|
@@ -275,7 +275,7 @@ footer { display: none !important; }
|
|
| 275 |
|
| 276 |
lang_choices = [("Any Language", "any")] + [(f"{v} ({k})", k) for k, v in sorted(LANGUAGE_NAMES.items(), key=lambda x: x[1])]
|
| 277 |
|
| 278 |
-
with gr.Blocks(
|
| 279 |
gr.HTML("""
|
| 280 |
<h1 class='title'>▸ WebCrawler / Dataset Builder</h1>
|
| 281 |
<p class='subtitle'>Crawl the web and extract text datasets filtered by language or topic — ready for NLP & LLM training.</p>
|
|
@@ -328,7 +328,7 @@ with gr.Blocks(theme=THEME, css=css, title="WebCrawler · Dataset Builder") as d
|
|
| 328 |
)
|
| 329 |
|
| 330 |
with gr.Tab("📜 Logs"):
|
| 331 |
-
log_box = gr.Textbox(label="Crawl Log", lines=20, max_lines=30
|
| 332 |
|
| 333 |
with gr.Tab("💾 Download"):
|
| 334 |
gr.Markdown("### Download your dataset")
|
|
@@ -344,4 +344,4 @@ with gr.Blocks(theme=THEME, css=css, title="WebCrawler · Dataset Builder") as d
|
|
| 344 |
stop_btn.click(fn=None, cancels=[crawl_event])
|
| 345 |
|
| 346 |
if __name__ == "__main__":
|
| 347 |
-
demo.launch()
|
|
|
|
| 275 |
|
| 276 |
lang_choices = [("Any Language", "any")] + [(f"{v} ({k})", k) for k, v in sorted(LANGUAGE_NAMES.items(), key=lambda x: x[1])]
|
| 277 |
|
| 278 |
+
with gr.Blocks(title="WebCrawler · Dataset Builder") as demo:
|
| 279 |
gr.HTML("""
|
| 280 |
<h1 class='title'>▸ WebCrawler / Dataset Builder</h1>
|
| 281 |
<p class='subtitle'>Crawl the web and extract text datasets filtered by language or topic — ready for NLP & LLM training.</p>
|
|
|
|
| 328 |
)
|
| 329 |
|
| 330 |
with gr.Tab("📜 Logs"):
|
| 331 |
+
log_box = gr.Textbox(label="Crawl Log", lines=20, max_lines=30)
|
| 332 |
|
| 333 |
with gr.Tab("💾 Download"):
|
| 334 |
gr.Markdown("### Download your dataset")
|
|
|
|
| 344 |
stop_btn.click(fn=None, cancels=[crawl_event])
|
| 345 |
|
| 346 |
if __name__ == "__main__":
|
| 347 |
+
demo.launch(theme=THEME, css=css)
|
requirements.txt
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
-
gradio>=
|
| 2 |
requests>=2.31.0
|
| 3 |
beautifulsoup4>=4.12.0
|
| 4 |
langdetect>=1.0.9
|
|
|
|
| 1 |
+
gradio>=6.0.0
|
| 2 |
requests>=2.31.0
|
| 3 |
beautifulsoup4>=4.12.0
|
| 4 |
langdetect>=1.0.9
|