Update app.py

app.py

@@ -7,9 +7,15 @@ from apscheduler.schedulers.background import BackgroundScheduler
 from huggingface_hub import HfApi
 
 import gradio as gr
-import datasets  # Added import for datasets
+import datasets  # Ensure the datasets library is imported
 
-from datetime import timezone  # Ensure timezone is imported
+from datetime import timezone
+import atexit  # To gracefully shut down the scheduler
+import logging  # For logging purposes
+
+# --- Logging Configuration ---
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
 
 # --- Data Loading and Processing ---
 
@@ -20,33 +26,44 @@ def get_df() -> pd.DataFrame:
     Loads and merges the papers and stats datasets, preprocesses the data by removing unnecessary columns,
     and adds a 'paper_page' link for each paper.
     """
-    # Load datasets
-    df_papers = datasets.load_dataset("hysts-bot-data/daily-papers", split="train").to_pandas()
-    df_stats = datasets.load_dataset("hysts-bot-data/daily-papers-stats", split="train").to_pandas()
-    
-    # Merge datasets on 'arxiv_id'
-    df = pd.merge(left=df_papers, right=df_stats, on="arxiv_id", suffixes=('_papers', '_stats'))
-    
-    # Reverse the DataFrame to have the latest papers first
-    df = df[::-1].reset_index(drop=True)
-    
-    # Ensure 'date' is in datetime format and handle missing dates
-    df["date"] = pd.to_datetime(df["date"], errors='coerce')
-    df["date"] = df["date"].dt.strftime("%Y-%m-%d").fillna(datetime.datetime.now(timezone.utc).strftime("%Y-%m-%d"))
-    
-    # Prepare the DataFrame by removing 'abstract'
-    paper_info = []
-    for _, row in tqdm.auto.tqdm(df.iterrows(), total=len(df)):
-        info = row.copy()
-        if "abstract" in info:
-            del info["abstract"]
-        paper_info.append(info)
-    df_prepared = pd.DataFrame(paper_info)
-    
-    # Add 'paper_page' links
-    df_prepared["paper_page"] = df_prepared["arxiv_id"].apply(lambda x: f"https://huggingface.co/papers/{x}")
-    
-    return df_prepared
+    try:
+        # Load datasets
+        logger.info("Loading 'daily-papers' dataset.")
+        df_papers = datasets.load_dataset("hysts-bot-data/daily-papers", split="train").to_pandas()
+        logger.info("Loading 'daily-papers-stats' dataset.")
+        df_stats = datasets.load_dataset("hysts-bot-data/daily-papers-stats", split="train").to_pandas()
+        
+        # Merge datasets on 'arxiv_id'
+        logger.info("Merging datasets on 'arxiv_id'.")
+        df = pd.merge(left=df_papers, right=df_stats, on="arxiv_id", suffixes=('_papers', '_stats'))
+        
+        # Reverse the DataFrame to have the latest papers first
+        df = df[::-1].reset_index(drop=True)
+        
+        # Ensure 'date' is in datetime format and handle missing dates
+        logger.info("Processing 'date' column.")
+        df["date"] = pd.to_datetime(df["date"], errors='coerce')
+        df["date"] = df["date"].dt.strftime("%Y-%m-%d").fillna(datetime.datetime.now(timezone.utc).strftime("%Y-%m-%d"))
+        
+        # Prepare the DataFrame by removing 'abstract'
+        logger.info("Removing 'abstract' column if present.")
+        paper_info = []
+        for _, row in tqdm.auto.tqdm(df.iterrows(), total=len(df)):
+            info = row.copy()
+            if "abstract" in info:
+                del info["abstract"]
+            paper_info.append(info)
+        df_prepared = pd.DataFrame(paper_info)
+        
+        # Add 'paper_page' links
+        logger.info("Adding 'paper_page' links.")
+        df_prepared["paper_page"] = df_prepared["arxiv_id"].apply(lambda x: f"https://huggingface.co/papers/{x}")
+        
+        logger.info("DataFrame preparation complete.")
+        return df_prepared
+    except Exception as e:
+        logger.error(f"Error in get_df: {e}")
+        return pd.DataFrame()  # Return empty DataFrame on error
 
 
 class Prettifier:
@@ -192,7 +209,7 @@ class PaperManager:
             
             # Convert 'date' column to datetime
             df_sorted = df.copy()
-            df_sorted['date_parsed'] = pd.to_datetime(df_sorted['date'], errors='coerce').dt.tz_localize(timezone.utc)
+            df_sorted['date_parsed'] = pd.to_datetime(df_sorted['date'], errors='coerce').dt.tz_localize(timezone.utc, ambiguous='NaT', nonexistent='NaT')
             df_sorted = df_sorted[df_sorted['date_parsed'] >= time_threshold]
             df_sorted = df_sorted.sort_values(by='upvotes', ascending=False).drop(columns=['date_parsed'])
         else:
@@ -210,11 +227,11 @@ class PaperManager:
         """
         if method not in ["hot", "new", "top"]:
             method = "hot"
-        print(f"Setting sort method to: {method}")
+        logger.info(f"Setting sort method to: {method}")
         self.sort_method = method
         if method == "top" and time_frame:
             self.top_time_frame = time_frame.lower()
-            print(f"Setting top time frame to: {self.top_time_frame}")
+            logger.info(f"Setting top time frame to: {self.top_time_frame}")
         self.sort_papers()
         return True  # Assume success
 
@@ -277,6 +294,9 @@ class PaperManager:
         """
         if self.current_page < self.total_pages:
             self.current_page += 1
+            logger.info(f"Navigated to page {self.current_page}.")
+        else:
+            logger.info("Already on the last page.")
         return self.get_current_page_papers()
 
     def prev_page(self) -> str:
@@ -285,12 +305,16 @@ class PaperManager:
         """
         if self.current_page > 1:
             self.current_page -= 1
+            logger.info(f"Navigated to page {self.current_page}.")
+        else:
+            logger.info("Already on the first page.")
         return self.get_current_page_papers()
 
     def refresh(self) -> str:
         """
         Refreshes the current list of papers.
         """
+        logger.info("Refreshing papers.")
         self.sort_papers()
         return self.get_current_page_papers()
 
@@ -317,6 +341,7 @@ def setup_paper_manager():
     df = get_df()
     paper_list = PaperList(df)
     paper_manager = PaperManager(paper_list)
+    logger.info("PaperManager setup complete.")
 
 
 # Initialize PaperManager at the start
@@ -328,9 +353,14 @@ def update_paper_manager() -> str:
     Updates the global PaperManager with the latest DataFrame.
     """
     global paper_manager
+    logger.info("Updating PaperManager with latest data.")
     df = get_df()
+    if df.empty:
+        logger.warning("DataFrame is empty. Skipping update.")
+        return paper_manager.get_current_page_papers()
     paper_manager.paper_list = PaperList(df)
     paper_manager.sort_papers()
+    logger.info("PaperManager updated successfully.")
     return paper_manager.get_current_page_papers()
 
 
@@ -344,6 +374,11 @@ scheduler_data.add_job(
     misfire_grace_time=60,
 )
 scheduler_data.start()
+logger.info("BackgroundScheduler started.")
+
+# Ensure the scheduler shuts down gracefully on exit
+atexit.register(lambda: scheduler_data.shutdown())
+logger.info("Scheduler shutdown registered.")
 
 
 # --- Gradio Interface Functions ---
@@ -352,6 +387,7 @@ def change_sort_method_ui(method: str, time_frame: str = "all time") -> str:
     """
     Changes the sort method and, if 'top' is selected, sets the time frame.
     """
+    logger.info(f"Changing sort method to: {method} with time frame: {time_frame}")
     if method.lower() == "top":
         paper_manager.set_sort_method(method.lower(), time_frame)
     else:
@@ -362,7 +398,8 @@ def change_sort_method_ui(method: str, time_frame: str = "all time") -> str:
 # --- CSS Styling ---
 
 css = """
-/* Existing CSS remains unchanged */
+/* Hacker News-like CSS */
+
 body {
     background-color: white;
     font-family: Verdana, Geneva, sans-serif;
@@ -492,7 +529,6 @@ table {
 }
 """
 
-
 # --- Initialize Gradio Blocks ---
 
 demo = gr.Blocks(css=css)
@@ -507,9 +543,22 @@ with demo:
 
             Once your paper is submitted, it will automatically appear in this demo.
             """)
-        # Header without Refresh Button
+        # Hacker News-like Header
         with gr.Row():
-            gr.Markdown("<b>Daily Papers</b>")
+            gr.HTML("""
+            <table border="0" cellpadding="0" cellspacing="0" class="header-table">
+                <tr>
+                    <td>
+                        <span class="pagetop">
+                            <b class="hnname"><a href="#">Daily Papers</a></b>
+                        </span>
+                    </td>
+                    <td align="right">
+                        <!-- Future Navigation Links Can Be Added Here -->
+                    </td>
+                </tr>
+            </table>
+            """)
         # Sort Options and Time Frame (conditionally visible)
         with gr.Row():
             sort_radio = gr.Radio(
@@ -568,4 +617,4 @@ with demo:
 # --- Launch the App ---
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()