Finished basic functions to find RSS feed uri and pull the feed.

assets/html.py

@@ -3,9 +3,9 @@
 TITLE = (
     '''
         <center> 
-            <h1>RSS feed finder</h1>
+            <h1>RSS feed reader</h1>
         </center>
     '''
 )
 
-DESCRIPTION = 'Enter a website to crawl for RSS feed URI.'
+DESCRIPTION = 'Enter a website to extract RSS feed entry titles.'

functions/helper_functions.py

@@ -1 +1,56 @@
 '''Helper functions for MCP tools.'''
+
+import logging
+import feedparser
+from findfeed import search
+
+
+def get_feed(website_url: str) -> str:
+    '''Finds the RSS feed URI for a website given the website's url.
+    
+    Args:
+        website_url: The url for the website to find the RSS feed for
+        
+    Returns:
+        The website's RSS feed URI as a string
+    '''
+
+    logger = logging.getLogger(__name__ + '.get_content')
+
+    feeds = search(website_url)
+
+    if len(feeds) > 0:
+        return str(feeds[0].url)
+
+    else:
+        return f'No feed found for {website_url}'
+
+
+def parse_feed(feed_uri: str) -> list:
+    '''Gets content from a remote RSS feed URI.
+    
+    Args:
+        feed_uri: The RSS feed to get content from
+
+    Returns:
+        List of titles for the 10 most recent entries in the RSS feed.
+    '''
+
+    logger = logging.getLogger(__name__ + '.parse_feed')
+
+    feed = feedparser.parse(feed_uri)
+    logger.info('%s yieled %s entries', feed_uri, len(feed.entries))
+
+    titles = []
+
+    for entry in feed.entries:
+
+        logger.debug('Entry attributes: %s', list(entry.keys()))
+
+        if 'title' in entry:
+            titles.append(entry.title)
+
+        if len(titles) >= 10:
+            break
+
+    return titles

functions/tools.py

@@ -1,22 +1,26 @@
 '''Tool functions for MCP server'''
 
-from findfeed import search
+import logging
+import functions.helper_functions as helper_funcs
 
 
-def get_feed(url: str) -> str:
-    '''Finds the RSS feed URI for a website given the website's url.
+def get_content(website_url: str) -> list:
+    '''Gets RSS feed content from a given website.
     
     Args:
-        url: The url for the website to find the RSS feed for
-        
+        website_url: URL of website to extract RSS feed content from
+
     Returns:
-        The website's RSS feed URI as a string
+        List of titles for the 10 most recent entries in the RSS feed.
     '''
 
-    feeds = search(url)
+    logger = logging.getLogger(__name__ + '.get_content')
+    logger.info('Getting feed content for: %s', website_url)
+
+    feed_uri = helper_funcs.get_feed(website_url)
+    logger.info('get_feed() returned %s', feed_uri)
 
-    if len(feeds) > 0:
-        return str(feeds[0].url)
+    content = helper_funcs.parse_feed(feed_uri)
+    logger.info('parse_feed() returned %s', content)
 
-    else:
-        return f'No feed found for {url}'
+    return '\n'.join(content)

rss_server.py

@@ -1,9 +1,32 @@
 '''Main script to run gradio interface and MCP server.'''
 
+import logging
+from pathlib import Path
+from logging.handlers import RotatingFileHandler
+
 import gradio as gr
 import assets.html as html
 from functions import tools as tool_funcs
 
+# Make sure log directory exists
+Path('logs').mkdir(parents=True, exist_ok=True)
+
+# Set-up logger
+logger = logging.getLogger()
+
+logging.basicConfig(
+    handlers=[RotatingFileHandler(
+        'logs/rss_server.log',
+        maxBytes=100000,
+        backupCount=10,
+        mode='w'
+    )],
+    level=logging.DEBUG,
+    format='%(levelname)s - %(name)s - %(message)s'
+)
+
+logger = logging.getLogger(__name__)
+
 
 with gr.Blocks() as demo:
 
@@ -12,14 +35,14 @@ with gr.Blocks() as demo:
 
     gr.Markdown(html.DESCRIPTION)
     website_url = gr.Textbox('hackernews.com', label='Website URL')
-    output = gr.Textbox(label='RSS feed URI')
+    output = gr.Textbox(label='RSS entry titles', lines=10)
     submit_button = gr.Button('Submit')
 
     submit_button.click( # pylint: disable=no-member
-        fn=tool_funcs.get_feed,
+        fn=tool_funcs.get_content,
         inputs=website_url,
         outputs=output,
-        api_name='get_feed'
+        api_name='Get RSS feed content'
     )