feat: Initial backend setup for Music Analyser

- Created FastAPI backend structure.
- Implemented Spotify Recently Played ingestion logic.
- Set up SQLite database with SQLAlchemy models.
- Added AI Service using Google Gemini.
- Created helper scripts for auth and background worker.
- Added Dockerfile and GitHub Actions workflow.

backend/app/services/ai_service.py

@@ -0,0 +1,40 @@
+import os
+import google.generativeai as genai
+from typing import List
+from ..models import PlayHistory, Track
+
+class AIService:
+    def __init__(self, api_key: str):
+        genai.configure(api_key=api_key)
+        self.model = genai.GenerativeModel('models/gemini-2.0-flash')
+
+    def generate_analysis(self, plays: List[PlayHistory]) -> str:
+        """
+        Generates a summary analysis of the provided play history.
+        """
+        if not plays:
+            return "No listening history available to analyze."
+
+        # Prepare a simple text representation of the history
+        history_text = "Here is my recent listening history:\n"
+        for play in plays:
+            history_text += f"- {play.track.name} by {play.track.artist} (Played at {play.played_at})\n"
+
+        prompt = f"""
+        You are a music taste analyst.
+        Analyze the following listening history and provide a short, fun, and insightful summary.
+        Identify the vibe, top artists, and any interesting patterns (e.g. "You started with high energy and chilled out").
+        Keep it under 200 words.
+
+        {history_text}
+        """
+
+        try:
+            response = self.model.generate_content(prompt)
+            return response.text
+        except Exception as e:
+            return f"AI Analysis failed: {str(e)}"
+
+# Singleton accessor
+def get_ai_service():
+    return AIService(api_key=os.getenv("GEMINI_API_KEY"))