update deps

README.md

@@ -12,6 +12,11 @@ license: mit
 
 # Annotate Web Languages
 
+## Requirements
+
+- Python 3.10 (same as HF space)
+- Chromium
+
 ## Usage
 
 Run the Gradio app
@@ -30,8 +35,13 @@ conda activate seed-crawl-annotator
 pip install -r requirements.txt
 ```
 
-
 Build Docker image
 ```bash
 docker build -t seed-crawl-annotator  .
+```
+
+Run app within container
+```bash
+docker run --rm -p 7860:7860 -e HF_TOKEN=$HF_TOKEN seed-crawl-annotator gradio app.py
+
 ```

app.py

@@ -18,6 +18,16 @@ from languages import ISO_CODE_TO_LANGUAGE_NAME
 
 OFFLINE = os.environ.get("OFFLINE", False)
 
+def get_candidate_urls():
+    return [
+        "http://example.com",
+        "https://wikipedia.org/",
+        "https://occiglot.eu",
+        "https://ostendorff.org",
+        "https://fr.wikipedia.org/",
+        "https://amazon.com/"
+    ]
+    
 def pil_image_to_base64(image):
     # Save the image to a BytesIO buffer
     buffer = BytesIO()
@@ -96,23 +106,26 @@ with gr.Blocks(fill_height=True) as demo:
     profile_state = gr.State([])
     gr.LoginButton()
 
-
     with gr.Column(visible=False) as wrapper_col:
+        login_status = gr.Markdown("no")
+
         def handle_login(profile: gr.OAuthProfile | None) -> dict:
             if profile:
                 gr.Info(f"Logged in as {profile.username}")
                 return {
                     profile_state: f"{profile.username}", 
                     wrapper_col: gr.update(visible=True),
+                    login_status:  "yes",
                 }
             else:
                 gr.Warning(f"You need to login to use this app.")
                 return {
-                    profile_state: None,
+                    profile_state: [],
                     wrapper_col: gr.update(visible=False),
+                    login_status:  "no",
                 }
             
-        demo.load(handle_login, inputs=None, outputs=[profile_state, wrapper_col])
+        demo.load(handle_login, inputs=None, outputs=[profile_state, wrapper_col, login_status])
 
         url_field = gr.Textbox(label="Website URL", placeholder="Enter a URL you want to annotate", interactive=True)
 
@@ -149,14 +162,7 @@ with gr.Blocks(fill_height=True) as demo:
 
 
         def set_random_url():
-            candidate_urls = [
-                "http://example.com",
-                "https://wikipedia.org/",
-                "https://occiglot.eu",
-                "https://ostendorff.org",
-                "https://fr.wikipedia.org/",
-                "https://amazon.com/"
-            ]
+            candidate_urls = get_candidate_urls()
             selected_url = random.choice(candidate_urls)
             return selected_url    
             
@@ -183,11 +189,11 @@ with gr.Blocks(fill_height=True) as demo:
         def do_crawl(profile_state, url, language_codes, categories, do_crawl=True):
 
             if profile_state:
-                html_str = f"<b>Thanks {profile_state}, we have saved your feedback!</b>"
-                gr.Info("Thanks for your feedback")
+                # html_str = f"<b>Thanks {profile_state}, we have saved your feedback!</b>"
+                gr.Info("✅ Thanks for your feedback")
             else:
-                gr.Error("Feedback could not be saved")
-                html_str = f"<b>Feedback could not be saved.</b> You are not authenticated."
+                gr.Error("❌ Feedback could not be saved")
+                # html_str = f"<b>Feedback could not be saved.</b> You are not authenticated."
 
             return {
                 url_field: "",

requirements.txt

@@ -1,5 +1,13 @@
-selenium >=4.0.0, < 5.0.0
+
 gradio==5.9.1
+gradio[oauth]
+
+selenium >=4.0.0, < 5.0.0
 Pillow>=8.3.1,<9.0
+
 trafilatura==2.0.0
-gradio[oauth]
+
+# trafilatura fix
+# ImportError: lxml.html.clean module is now a separate project lxml_html_clean.
+lxml_html_clean
+