stat matching between LLM and SeedData. Finished frontend

projects/Mood2SpotifyRec/.streamlit/config.toml

@@ -1,3 +1,3 @@
 # Forcing light theme on streamlit app
 [theme]
-base="light" 
+base="dark" 

projects/Mood2SpotifyRec/SeedData.py

@@ -27,3 +27,25 @@ def get_test_seed_data(self):
             target_danceability=0.7,
             target_acousticness=0.3,
         )
+
+    def to_seed_data(self, in_metric):
+        return SeedData(
+            seed_genres=["indie", "pop"],
+            seed_artists=[
+                "4NHQUGzhtTLFvgF5SZesLK"
+            ],  # Example artist ID for Tame Impala
+            target_valence=in_metric["valence"],
+            target_energy=in_metric["energy"],
+            target_danceability=in_metric["danceability"],
+            target_acousticness=in_metric["acousticness"],
+        )
+
+    def view_seed_data(self):
+        return {
+            "seed_genres": self.seed_genres,
+            "seed_artists": self.seed_artists,
+            "target_valence": self.target_valence,
+            "target_energy": self.target_energy,
+            "target_danceability": self.target_danceability,
+            "target_acousticness": self.target_acousticness,
+        }

projects/Mood2SpotifyRec/llm.py

@@ -1,7 +1,7 @@
 import google.generativeai as genai
 
 
-class NatLang2Latext:
+class Mood2SpotifyRec:
     def __init__(
         self, google_api_key: str | None = None, model_name: str = "gemini-pro"
     ):
@@ -55,7 +55,7 @@ def convert(self, text: str) -> dict[str, float]:
         genai.configure(api_key=self.google_api_key)
         model = genai.GenerativeModel(model_name=self.model_name)
         prompt = f"<MOOD>{self.prompt}{text}</MOOD>\n<METRICS>\n"
-        raw_ouput = model.generate_content(
+        raw_output = model.generate_content(
             contents=prompt,
             generation_config=genai.types.GenerationConfig(
                 candidate_count=1,
@@ -64,11 +64,23 @@ def convert(self, text: str) -> dict[str, float]:
                 stop_sequences=["</METRICS>"],
             ),
         ).text
-
+        # metrics = {}
+        # for line in raw_ouput.split("\n"):
+        #     if ":" in line:
+        #         key, value = line.split(":")
+        #         metrics[key.strip()] = float(value.strip())
+        # return metrics
         metrics = {}
-        for line in raw_ouput.split("\n"):
+        for line in raw_output.split("\n"):
             if ":" in line:
                 key, value = line.split(":")
-                metrics[key.strip()] = float(value.strip())
+                range_values = value.strip().split("-")
+                if len(range_values) == 2:
+                    # Calculate the average of the range values
+                    float_values = [float(x) for x in range_values]
+                    metrics[key.strip()] = sum(float_values) / len(float_values)
+                else:
+                    # If there's only one value, convert it directly
+                    metrics[key.strip()] = float(value.strip())
 
         return metrics

projects/Mood2SpotifyRec/main.py

@@ -2,16 +2,84 @@
 import streamlit.components.v1 as components
 import spotify
 import SeedData
+import os
+from llm import Mood2SpotifyRec
+from dotenv import load_dotenv
+from PIL import Image
 
+
+# Load environment variables
+load_dotenv()
+
+# Streamlit page background
+page_bg_img = f"""
+<style>
+[data-testid="stAppViewContainer"] > .main {{
+background-image: url("https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQlo3P16G9dPsRjdEhx77zlzr2idJqnXZpLuQ90xNYrpQ&s");
+background-size: cover;
+background-position: center center;
+background-repeat: no-repeat;
+background-attachment: local;
+}}
+[data-testid="stHeader"] {{
+background: rgba(0,0,0,0);
+}}
+</style>
+"""
+st.markdown(page_bg_img, unsafe_allow_html=True)
+
+# =========================================================================================================
+# Logo and Navigation
 st.title("Mood 2 Spotify Recommendation")
+mood2spotifyrec_logo = Image.open(
+    "resources/Mood2SpotifyRec_Logov2-fotor-bg-remover-20240423143851.png"
+)
+st.image(mood2spotifyrec_logo, width=300)
+# change the size of the logo and center
+
 
 # Inputs for Client ID and Client Secret
-client_id = st.text_input("Client ID", "")
-client_secret = st.text_input("Client Secret", "")
+client_id = os.getenv("SPOTIPY_CLIENT_ID")
+client_secret = os.getenv("SPOTIPY_CLIENT_SECRET")
 sp = spotify.SpotifyTools(client_id, client_secret)
 access_token = sp.get_access_token()
 
+# =========================================================================================================
+# The user input and recommendation section
+converter = Mood2SpotifyRec(google_api_key=os.getenv("GOOGLE_API_KEY"))
+# User input for mood
+st.subheader("Enter your mood:")
+mood = st.text_input("Mood", "")
 
+# Output
+if st.button("Get Tracks Recommendations:"):
+    # Validate inputs
+    if not client_id or not client_secret:
+        st.warning("Please fill in all fields.")
+    if not mood.strip():
+        st.write("*Please complete the missing fields.")
+    else:
+        st.header("Output")
+        st.write("This is the Spotify metrics based on your mood:")
+        mood_text = mood.strip()
+        metrics = converter.convert(mood_text)
+        st.write(f"`{metrics}`")  # Display metrics in code block
+        rec_seed_data = SeedData.SeedData().to_seed_data(metrics)
+        user_id = sp.user_id
+        playlist_id = sp.add_recommend_tracks_to_playlist(
+            "Mood2Spotify Playlist", rec_seed_data
+        )
+        # Your playlist embed URL
+        playlist_embed_url = "https://open.spotify.com/embed/playlist/" + playlist_id
+        # Embed the playlist using an iframe
+        components.html(
+            f"""<iframe src="{playlist_embed_url}" width=100% height=700 frameborder="0" allowtransparency="true" allow="encrypted-media"></iframe>""",
+            height=700,
+        )
+
+
+# =========================================================================================================
+# API test and LLM suggestion test
 # Section for Audio Features
 # st.subheader("Get Audio Features")
 # track_id = st.text_input("Spotify Track ID", "")
@@ -50,39 +118,62 @@
 #         else:
 #             st.error("Failed to retrieve access token for recommendations")
 
-# Section for track recommendations
-st.subheader("Get TEST Track Recommendations")
-test_seed_data = SeedData.SeedData().get_test_seed_data()
-st.write(test_seed_data)
-if st.button("Get TEST Tracks Recommendations"):
-    if not client_id or not client_secret:
-        st.warning("Please fill in all fields.")
-    else:
-        if access_token:
-            recommended_tracks = sp.get_test_track_recommendation()
-            track_uris = [track["uri"] for track in recommended_tracks["tracks"]]
-            # Display the tracks as clickable links
-            for track in recommended_tracks["tracks"]:
-                track_name = track["name"]
-                artists = ", ".join(
-                    artist["name"] for artist in track["album"]["artists"]
-                )
-                spotify_url = track["external_urls"]["spotify"]
-                st.write(f"[{track_name} by {artists}]({spotify_url})")
-
-        else:
-            st.error("Failed to retrieve access token for recommendations")
+# Spotify recommendation API test
+# st.subheader("Get TEST Track Recommendations")
+# test_seed_data = SeedData.SeedData().get_test_seed_data()
+# st.write(test_seed_data)
+# if st.button("Get TEST Tracks Recommendations"):
+#     if not client_id or not client_secret:
+#         st.warning("Please fill in all fields.")
+#     else:
+#         if access_token:
+#             recommended_tracks = sp.get_test_track_recommendation()
+#             track_uris = [track["uri"] for track in recommended_tracks["tracks"]]
+#             # Display the tracks as clickable links
+#             for track in recommended_tracks["tracks"]:
+#                 track_name = track["name"]
+#                 artists = ", ".join(
+#                     artist["name"] for artist in track["album"]["artists"]
+#                 )
+#                 spotify_url = track["external_urls"]["spotify"]
+#                 st.write(f"[{track_name} by {artists}]({spotify_url})")
+
+#         else:
+#             st.error("Failed to retrieve access token for recommendations")
 
 # Show embed TEST recommendation playlist
-if st.button("Show playlist"):
-    user_id = sp.user_id
-    playlist_id = sp.add_recommend_tracks_to_playlist(
-        "Mood2Spotify Playlist", test_seed_data
-    )
-    # Your playlist embed URL
-    playlist_embed_url = "https://open.spotify.com/embed/playlist/" + playlist_id
-    # Embed the playlist using an iframe
-    components.html(
-        f"""<iframe src="{playlist_embed_url}" width=100% height=700 frameborder="0" allowtransparency="true" allow="encrypted-media"></iframe>""",
-        height=700,
-    )
+# if st.button("Show playlist"):
+#     user_id = sp.user_id
+#     playlist_id = sp.add_recommend_tracks_to_playlist(
+#         "Mood2Spotify Playlist", test_seed_data
+#     )
+#     # Your playlist embed URL
+#     playlist_embed_url = "https://open.spotify.com/embed/playlist/" + playlist_id
+#     # Embed the playlist using an iframe
+#     components.html(
+#         f"""<iframe src="{playlist_embed_url}" width=100% height=700 frameborder="0" allowtransparency="true" allow="encrypted-media"></iframe>""",
+#         height=700,
+#     )
+
+
+# =========================================================================================================
+# Sidebar
+st.sidebar.header("About")
+st.sidebar.markdown(
+    """
+    A project by [polskiXO](https://github.com/polskiXO) and [menamerai](https://github.com/menamerai) for EECE3092.
+    
+    Mood2SpotifyRec is made using Google Gemini LLM to generate seed param based on user input mood to feed into Spotify API recommendation API for playlist generation. 
+    """
+)
+
+st.sidebar.header("Resources")
+st.sidebar.markdown(
+    """
+- [GitHub Repository](https://github.com/polskiXO/python-beginner-projects/tree/2-natlang2latex-project/projects/Mood2SpotifyRec)
+- [Streamlit Documentation](https://docs.streamlit.io/)
+- [LaTeX Documentation](https://www.latex-project.org/help/documentation/)
+- [Google Gemini](https://ai.google.dev/)
+- [EECE3092 Course](https://www.ece.mcgill.ca/~ece309/)
+"""
+)

projects/Mood2SpotifyRec/spotify.py

@@ -114,7 +114,6 @@ def get_tracks_recommendations(self, seed_data):
         )
         return recommended_tracks
 
-    # TODO: remove when done
     def get_test_track_recommendation(self):
         testSeedData = SeedData.SeedData().get_test_seed_data()
         return self.get_tracks_recommendations(testSeedData)