Add SQLAlchemy ORM with Alembic migrations

- Added SQLAlchemy 2.0 and Alembic 1.13 dependencies
- Created models.py with Channel and VideoEntry ORM models
- Created database.py for database configuration and session management
- Initialized Alembic migration system with initial migration
- Updated feed_parser.py with save_to_db() method for persistence
- Updated main.py with database initialization and new API routes:
  - /api/feed now saves to database by default
  - /api/channels lists all tracked channels
  - /api/history/<channel_id> returns video history
- Updated .gitignore to exclude database files
- Updated CLAUDE.md with comprehensive ORM and migration documentation

Database uses SQLite (yottob.db) with upsert logic to avoid duplicates.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

feed_parser.py

@@ -4,9 +4,15 @@ This module handles fetching and parsing YouTube channel RSS feeds,
 with filtering capabilities to exclude unwanted content like Shorts.
 """
 
+from datetime import datetime
 import feedparser
 from typing import Dict, List, Optional
 
+from sqlalchemy.orm import Session
+from sqlalchemy.exc import IntegrityError
+
+from models import Channel, VideoEntry
+
 
 class FeedEntry:
     """Represents a single entry in a YouTube RSS feed."""
@@ -66,3 +72,58 @@ class YouTubeFeedParser:
             "feed_link": feed.feed.link,
             "entries": [entry.to_dict() for entry in entries]
         }
+
+    def save_to_db(self, db_session: Session, feed_data: Dict) -> Channel:
+        """Save feed data to the database.
+
+        Args:
+            db_session: SQLAlchemy database session
+            feed_data: Dictionary containing feed metadata and entries (from fetch_feed)
+
+        Returns:
+            The Channel model instance
+
+        This method uses upsert logic:
+        - Updates existing channel if it exists
+        - Creates new channel if it doesn't exist
+        - Only inserts new video entries (ignores duplicates)
+        """
+        # Get or create channel
+        channel = db_session.query(Channel).filter_by(
+            channel_id=self.channel_id
+        ).first()
+
+        if channel:
+            # Update existing channel
+            channel.title = feed_data["feed_title"]
+            channel.link = feed_data["feed_link"]
+            channel.last_fetched = datetime.utcnow()
+        else:
+            # Create new channel
+            channel = Channel(
+                channel_id=self.channel_id,
+                title=feed_data["feed_title"],
+                link=feed_data["feed_link"],
+                last_fetched=datetime.utcnow()
+            )
+            db_session.add(channel)
+            db_session.flush()  # Get the channel ID
+
+        # Add video entries (ignore duplicates)
+        for entry_data in feed_data["entries"]:
+            # Check if video already exists
+            existing = db_session.query(VideoEntry).filter_by(
+                link=entry_data["link"]
+            ).first()
+
+            if not existing:
+                video = VideoEntry(
+                    channel_id=channel.id,
+                    title=entry_data["title"],
+                    link=entry_data["link"],
+                    created_at=datetime.utcnow()
+                )
+                db_session.add(video)
+
+        db_session.commit()
+        return channel