Revised to show recent feedbacked movies in the original dataset

examples/streamlit/movielens_dashboard.py

@@ -1,3 +1,4 @@
+import datetime
 import os
 import streamlit as st
 import pandas as pd
@@ -11,6 +12,7 @@
 DATASET_URL = "http://files.grouplens.org/datasets/movielens/ml-100k/"
 RATINGS_FILE = DATASET_URL + "u.data"
 ITEMS_FILE = DATASET_URL + "u.item"
+USER_FILE = DATASET_URL + "u.user"
 
 # Caching functions
 @st.cache_data
@@ -33,6 +35,13 @@ def load_movielens_100k():
 
     return ratings_df, items_df, merged_df
 
+@st.cache_data
+def load_user_data():
+    """Load the user demographic data from the MovieLens 100K dataset."""
+    col_user = ['user_id', 'age', 'gender', 'occupation', 'zip_code']
+    user_df = pd.read_csv(USER_FILE, sep='|', names=col_user, encoding='latin-1')
+    return user_df
+
 @st.cache_data
 def load_movie_posters():
     """Load movie posters from an external source."""
@@ -100,8 +109,9 @@ def handle_feedback():
         selected_index = feedback_options.index(st.session_state[feedback_id])
         if selected_index == 0:
             st.session_state.feedback["liked"].append(title)
-        elif selected_index == 1 and title not in st.session_state.feedback["disliked"]:
+        elif selected_index == 1:
             st.session_state.feedback["disliked"].append(title)
+        st.session_state.feedback["date"].append(feedback_date)
 
     st.radio("Feedback", options=feedback_options, key=feedback_id, index=None, horizontal=True, on_change=handle_feedback)
 
@@ -111,16 +121,23 @@ def handle_feedback():
 
 # Load data
 ratings_df, items_df, merged_df = load_movielens_100k()
+user_df = load_user_data()
 poster_dict = load_movie_posters()
 
 # User input
 user_id = st.sidebar.number_input("Enter User ID", min_value=1, value=1)
+if "last_user_id" not in st.session_state:
+    st.session_state.last_user_id = user_id
+# Initialize session state
+if user_id != st.session_state.last_user_id:
+    st.session_state.feedback = {"liked": [], "disliked": [], "date": []}
+    st.session_state.last_user_id = user_id
+elif "feedback" not in st.session_state:
+    st.session_state.feedback = {"liked": [], "disliked": [], "date": []}
 top_n = st.sidebar.slider("Number of Recommendations", min_value=1, max_value=10, value=5)
 recent_feedback_n = st.sidebar.slider("Recent Feedback Items (N)", min_value=1, max_value=20, value=5)
 
-# Initialize session state
-if "feedback" not in st.session_state:
-    st.session_state.feedback = {"liked": [], "disliked": []}
+feedback_date = st.sidebar.date_input("Feedback date", datetime.date(1998, 4, 23))
 
 # Tabs for navigation
 tab1, tab2, tab3 = st.tabs(["Recommendations", "Recently Feedbacked Movies", "Search Movies"])
@@ -134,20 +151,58 @@ def handle_feedback():
         display_movie(row["movie_title"], row["movie_id"], poster_dict)
 
 with tab2:
-    st.write("### Recently Feedbacked Movies")
+    st.write("### User Information and Recently Feedbacked Movies")
+
+    # Fetch user details
+    user_info = user_df[user_df["user_id"] == user_id]
+    if not user_info.empty:
+        cols = st.columns(4)  # Create four columns for horizontal layout
+        cols[0].write(f"**User ID:** {user_id}")
+        cols[1].write(f"**Gender:** {user_info.iloc[0]['gender']}")
+        cols[2].write(f"**Age:** {user_info.iloc[0]['age']}")
+        cols[3].write(f"**Occupation:** {user_info.iloc[0]['occupation']}")
+    else:
+        st.write(f"User ID {user_id} not found in the dataset.")
+
+    st.divider()
+
+    # Get recently rated movies
+    user_ratings = ratings_df[ratings_df["user_id"] == user_id]
+    if not user_ratings.empty:
+        user_ratings = user_ratings.sort_values(by="unix_timestamp", ascending=False)
+        recent_ratings = user_ratings.merge(items_df[["movie_id", "movie_title"]], on="movie_id").head(recent_feedback_n)
+
+        # Format data for display
+        recent_ratings["date_rated"] = pd.to_datetime(recent_ratings["unix_timestamp"], unit="s").dt.date
+        recent_ratings_table = recent_ratings[["movie_title", "rating", "date_rated"]].rename(
+            columns={
+                "movie_title": "Movie Title",
+                "rating": "Rating",
+                "date_rated": "Date Rated",
+            }
+        )
+
+        st.write(f"#### Last {recent_feedback_n} Rated Movies")
+        st.dataframe(recent_ratings_table, use_container_width=True)
+    else:
+        st.write("No ratings found for this user.")
+
+    st.divider()
+
     liked_movies = st.session_state.feedback["liked"][-recent_feedback_n:]
     disliked_movies = st.session_state.feedback["disliked"][-recent_feedback_n:]
+    feedback_dates = st.session_state.feedback["date"][-recent_feedback_n:]
 
     if liked_movies or disliked_movies:
         if liked_movies:
             st.write("#### Liked Movies")
-            for movie in liked_movies:
-                st.write(f"👍 {movie}")
+            for movie, date in zip(liked_movies, feedback_dates):
+                st.write(f"👍 {movie} at 🗓️ {date})")
 
         if disliked_movies:
             st.write("#### Disliked Movies")
-            for movie in disliked_movies:
-                st.write(f"👎 {movie}")
+            for movie, date in zip(disliked_movies, feedback_dates):
+                st.write(f"👎 {movie} at 🗓️ {date}")
     else:
         st.write("No feedback given yet.")