added from github repo, branch ~peter, added chatbot

Home.py

@@ -0,0 +1,169 @@
+import streamlit as st
+import os
+import random
+import time
+from module.__custom__ import *
+from streamlit_extras.switch_page_button import switch_page
+
+# Openai API Key
+import openai 
+import json
+def read_api_key_from_secrets(file_path='secrets.json'):
+    try:
+        with open(file_path, 'r') as secrets_file:
+            secrets_data = json.load(secrets_file)
+            openai_api_key = secrets_data.get('openai_api_key')
+            
+            if openai_api_key is not None:
+                return openai_api_key
+            else:
+                raise KeyError("'openai_api_key' not found in secrets.json")
+    except FileNotFoundError:
+        raise FileNotFoundError(f"The file {file_path} was not found.")
+    except json.JSONDecodeError:
+        raise ValueError(f"Error decoding JSON in {file_path}. Please check the file format.")
+
+# Example usage
+try:
+    key = read_api_key_from_secrets()
+    openai.api_key = key
+    os.environ['OPENAI_API_KEY'] = key
+    print(f"OpenAI API Key Found")
+except (FileNotFoundError, ValueError, KeyError) as e:
+    print(f"Error: {e}")
+
+from langchain.vectorstores import Chroma
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.chains.query_constructor.base import AttributeInfo
+from langchain.retrievers.self_query.base import SelfQueryRetriever
+embedding = OpenAIEmbeddings()
+# from langchain.embeddings.sentence_transformer import SentenceTransformerEmbeddings
+# embedding = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
+
+# LLM
+from langchain.chat_models import ChatOpenAI
+llm_name = "gpt-3.5-turbo"
+llm = ChatOpenAI(model_name=llm_name, temperature=0)
+
+# load from disk
+db_cos = Chroma(
+    persist_directory="./data/docs/chroma_cos", 
+    embedding_function=embedding
+)
+db_plot = Chroma(
+    persist_directory="./data/docs/chroma_plot", 
+    embedding_function=embedding
+)
+
+metadata_field_info = [
+    AttributeInfo(
+        name="name",
+        description="The name of the video game on steam",
+        type="string",
+    )
+]
+document_content_description = "Brief summary of a video game on Steam"
+
+
+with st.sidebar: is_plot = st.toggle('Enable Plot')
+db_selected = db_cos
+if is_plot: db_selected = db_plot
+
+
+retriever = SelfQueryRetriever.from_llm(
+    llm,
+    db_selected,
+    document_content_description,
+    metadata_field_info,
+    enable_limit=True, 
+)
+
+emoji = '🕹️ GameInsightify'
+st.header(emoji)
+
+# Initialize chat history
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+if 'gamenames' not in st.session_state:
+    st.session_state.gamenames = []
+
+# Slider on range and button to clear chat history
+col1, col2= st.columns([8,2])
+with col1: 
+    st.title("Game Recommender")
+with col2: 
+    if st.button("Clear chat"):
+        st.session_state.messages = []
+        st.session_state.gamenames = []
+
+
+# Display chat messages from history on app rerun
+tab1, tab2= st.tabs(['Chatbot', ' '])
+with tab1:          # this tab exist becasue i have to limit the height of chatbot
+    for message in st.session_state.messages:
+        with st.chat_message(message["role"]):
+            st.markdown(message["content"])
+with tab2: pass    # this tab exist becasue i have to limit the height of chatbot
+
+
+# Accept user input
+if prompt := st.chat_input("Need a game recommendation?"):
+    st.session_state.messages.append({"role": "user", "content": prompt})       # Add user message to chat history
+    with st.chat_message("user"):                                               # Display user message in chat message container
+        st.markdown(prompt)
+
+
+    with st.chat_message("assistant"):                                          # Display assistant response in chat message container
+        message_placeholder = st.empty()
+        
+        # docs = db.max_marginal_relevance_search(prompt,k=query_num, fetch_k=10) # Sending query to db
+        docs = retriever.invoke(prompt)                                         # retrieve response from chatgpt
+        full_response = random.choice(                                          # 1st sentence of response
+            ["I recommend the following games:\n",
+            f"Hi, human! These are the {len(docs)} best games:\n",
+            f"I bet you will love these {len(docs)} games:\n",]
+        )
+        
+        # formatting response from db
+        top_games = []
+        assistant_response = ""
+        for idx, doc in enumerate(docs):
+            gamename = doc.metadata['name']
+            top_games.append(gamename)
+            assistant_response += f"{idx+1}. {gamename}\n"
+        
+        # separating response into chunk of words
+        chunks = []
+        for line in assistant_response.splitlines():
+            for word in line.split() : chunks.append(word)
+            chunks.append('\n')
+        chunks = chunks[0:-1]
+        
+        # Simulate stream of response with milliseconds delay
+        for chunk in chunks:
+            full_response += chunk + " "
+            time.sleep(0.05)
+            message_placeholder.markdown(full_response + "▌")   # Add a blinking cursor to simulate typing
+        message_placeholder.markdown(full_response)
+
+    # Add assistant response to chat history
+    st.session_state.messages.append({"role": "assistant", "content": full_response})
+    if is_plot: st.session_state.gamenames.append(top_games)
+                
+col1, col2, col3= st.columns([4,2,4])
+with col2:
+    if is_plot and db_selected==db_plot:
+        if st.button("Plot Games"):     # button in center column
+            switch_page('Overall')
+
+
+# Styling on Tabs
+css=f'''
+div.stTabs {{
+    height: 40vh;
+    overflow-y: scroll;
+    overflow-x: hidden;
+}}
+</style>
+'''
+st.markdown(f'<style>{css}</style>', unsafe_allow_html=True)

README.md

@@ -1,10 +1,10 @@
 ---
 title: Streamlit
-emoji: 🐠
-colorFrom: pink
+emoji: 🕹️
+colorFrom: blue
 colorTo: green
 sdk: streamlit
 sdk_version: 1.27.0
-app_file: Overall.py
+app_file: Home.py
 pinned: false
 ---

module/__custom__.py

@@ -39,13 +39,36 @@ def name(target):
 # Appending favorite game selected by user to filtered list
     """use after search box
     """
-def add_top_games(top_games, favorite_game, ranges, df_ax):
-    if favorite_game in top_games: 
-        top_games = df_ax.gamename.unique()[ranges[0]-1:ranges[1]+1]
-    elif favorite_game!=None: 
-        top_games = np.append(top_games, favorite_game)
+def add_top_games(top_games, fav_games, ranges, df_ax):
+    ranges_last = ranges[1]
+    for fav in fav_games:
+        if fav in top_games: 
+            range_last+=1
+            top_games = df_ax.gamename.unique()[ranges[0]-1:range_last]
+        elif fav!=None: 
+            top_games = np.append(top_games, fav)
     return top_games
 
+def add_list(favorite_game, rec_games):
+    fav_games = []
+    if favorite_game[0]: fav_games = favorite_game
+    if rec_games: 
+        fav_games = list(set(fav_games + rec_games))
+    return fav_games
+
+# fav_games = []
+# if favorite_game: fav_games = [favorite_game]
+# if st.session_state.gamenames[-1]: 
+#     rec_games = st.session_state.gamenames[-1]
+#     fav_games = list(set(fav_games + rec_games))
+
+# def add_top_games(top_games, favorite_game, ranges, df_ax):
+#     if favorite_game in top_games: 
+#         top_games = df_ax.gamename.unique()[ranges[0]-1:ranges[1]+1]
+#     elif favorite_game!=None: 
+#         top_games = np.append(top_games, favorite_game)
+#     return top_games
+
 # Linear search over all gamenames
 def search(target):
     gamenames = df['gamename'].unique()                 # all unique gamenames
@@ -65,16 +88,39 @@ def searchbox():
     )
     return selected_game
 
+# Overloaded with target name 
+def searchbox(target):
+    col1, col2= st.columns([1,1])
+    with col1:  
+        selected_game = st_searchbox(
+            search_function=search,
+            key="gamename_searchbox",
+            default_options=target, 
+            placeholder="Compare with your Favorite Game...",
+        )
+    return selected_game
 
 ########## PAGE SECTION ##########
 # Datafram Section
     """Dataframe
     together with Title
     """
-def dfbox(ax_name, y_name, df_ax):
-    title = f"1.1 Dataset of :blue[{ax_name}] Games Sorted by :blue[{y_name}]:"
-    st.subheader(title)
-    st.dataframe(df_ax)
+def dfbox(ax_name, y_name, df_ax, ranges, order_name):
+    title = f"1.2 :blue[{ax_name}] Games with the {order_name} :blue[{y_name}]:"
+    with st.sidebar:
+        gamenames = df_ax.gamename.unique()
+        df_names = pd.DataFrame(gamenames, columns=['gamename'])
+        st.write(title)
+        st.dataframe(df_names[ranges[0]:ranges[1]+1])
+
+def rec_dfbox():
+    title = f"1.1 :blue[Recommended] by :green[GameInsightify]"
+    if len(st.session_state.gamenames) > 0:
+        with st.sidebar:
+            rec_games = st.session_state.gamenames[-1]
+            df_names = pd.DataFrame(rec_games, columns=['gamename'])
+            st.write(title)
+            st.dataframe(df_names[0:len(rec_games)])
 
 # plot 1 Section
     """Plot contains the top ranked games
@@ -89,10 +135,15 @@ def plot1_box(ax, y, order_name, ranges, df_ax, top_games):
     st.subheader(title)
 
     # Plot 1 - select box
-    favorite_game = searchbox()                                                     # search box to add a user favorite game on Plot 1
-    top_games = add_top_games(top_games, favorite_game, ranges, df_ax)        
+    rec_games = []
+    if len(st.session_state.gamenames) > 0 : rec_games = st.session_state.gamenames[-1]
+    favorite_game = searchbox(None)                                                     # search box to add a user favorite game on Plot 1
+    fav_games = add_list([favorite_game], rec_games)
+    fav_options = st.multiselect('Select Recommended Games', fav_games)
+    
     options = top_games
-    selected_options = st.multiselect('Select Video Games', options)
+    selected_tops = st.multiselect('Select Video Games', options)
+    selected_options = add_top_games(selected_tops, fav_options, ranges, df_ax) 
 
     # Plot 1
     title_names = ','.join(selected_options)
@@ -178,6 +229,81 @@ def plot2_box(theme, y, genres, df_bx):
     with tab3:
         st.plotly_chart(fig_sc)
 
+# Plot 3 - Pie chart
+import plotly.express as px
+def plot3_box(theme, labels):
+    title = f"2.1 Ratio of Games Among :blue[{theme}]"
+    st.subheader(title)
+    
+    if (type(labels)==str):
+        values = []
+        index = df[labels].unique().tolist()
+        for idx, value in enumerate(index):
+            count = len(df[df[labels] == value])
+            values.append(count)
+            if(count/len(df) < 0.02): index[idx] = 'Other'
+        
+        df_p = pd.DataFrame(data = values,  
+                index = index,  
+                columns = ['counts']) 
+        df_p = df_p.reset_index().rename(columns={'index':labels})
+        fig_ratio = px.pie(df_p, values='counts', names=labels)
+        st.plotly_chart(fig_ratio)
+        
+    else:
+        values = []
+        for label in labels:
+            value = len(df[df[label]==1])
+            values.append(value)
+        
+        fig_ratio = go.Figure(data=[go.Pie(labels=labels, values=values)])
+        st.plotly_chart(fig_ratio)
+
+# Could not overload function, so renamed it
+def plot3_box_limit(theme, labels, limit_perc):
+    title = f"2.1 Ratio of Games Among :blue[{theme}] over :blue[{limit_perc*100}%]"
+    st.subheader(title)
+    
+    values = []
+    index = df[labels].unique().tolist()
+    for idx, value in enumerate(index):
+        count = len(df[df[labels] == value])
+        values.append(count)
+        if(count/len(df) < limit_perc): index[idx] = 'Other'
+    
+    df_p = pd.DataFrame(data = values,  
+            index = index,  
+            columns = ['counts']) 
+    df_p = df_p.reset_index().rename(columns={'index':labels})
+    fig_ratio = px.pie(df_p, values='counts', names=labels)
+    st.plotly_chart(fig_ratio)
+
+def plot_chat_box(y, query_num, top_games):
+    y_name = name(y)            # formeting strings
+    
+    title = f"1.2 Comparison on The {query_num} Best Recommended Games on :blue[{y_name}]"
+    st.subheader(title)
+
+    # Plot 1 - select box                                                    # search box to add a user favorite game on Plot 1
+    options = top_games
+    selected_options = st.multiselect('Select Video Games', options)
+
+    # Plot 1
+    title_names = ','.join(selected_options)
+    plot_title = f"Monthly {y_name} of {title_names} Over Time"
+    gb = df.sort_values(by='date')
+    gb_list = {game: gb[gb["gamename"] == game] for game in selected_options}
+
+    fig_1 = go.Figure()
+    fig_1.update_layout(
+        title = plot_title, 
+        xaxis_title = 'Date',
+        yaxis_title = y_name,
+    )
+    for game, gb in gb_list.items():
+        fig_1 = fig_1.add_trace(go.Scatter(x=gb["date"], y=gb[y], name=game, mode='lines'))
+    st.plotly_chart(fig_1)
+
 
 ##### Execute Page #####
 def exec_page(emoji, theme, page_genres):
@@ -217,7 +343,7 @@ def exec_page(emoji, theme, page_genres):
     max = len(max)-1
     ranges = st.slider(
         label=f'Select range of the {order_name.lower()} games',
-        value = (1, 5), 
+        value = (1, 3), 
         min_value=1, max_value=30, 
         # min_value=1, max_value=max, 
     )
@@ -225,7 +351,8 @@ def exec_page(emoji, theme, page_genres):
     
 
     # Dataframe preview
-    dfbox(ax_name, y_name, df_ax)
+    rec_dfbox()
+    dfbox(ax_name, y_name, df_ax, ranges, order_name)  
 
     ##### PLOT 1 #####
     # Plot 1 - markdown
@@ -242,7 +369,7 @@ def exec_page_home(theme):
     st_page_selectbox(theme)
     
     # Header
-    st.header(f"👋 ForcaSteam")
+    st.header("👋")
     st.title("Customized Plot on :blue[General Features]")
 
     ##### FILTER #####
@@ -266,14 +393,15 @@ def exec_page_home(theme):
     max = len(max)-1
     ranges = st.slider(
         label=f'Select range of the {order_name.lower()} games',
-        value = (1, 5),
+        value = (1, 3),
         min_value=1, max_value=30, 
         # min_value=1, max_value=max, 
     )
     top_games = df_ax.gamename.unique()[ranges[0]-1:ranges[1]]
 
     # Dataframe preview
-    dfbox("", y_name, df_ax)
+    rec_dfbox()
+    dfbox("", y_name, df_ax, ranges, order_name)
 
 
     ##### PLOT 1 #####
@@ -284,10 +412,15 @@ def exec_page_home(theme):
 
 
     # Plot 1 - select box
-    favorite_game = searchbox()                                                     # search box to add a user favorite game on Plot 1
-    top_games = add_top_games(top_games, favorite_game, ranges, df_ax) 
+    rec_games = []
+    if len(st.session_state.gamenames)>0: rec_games = st.session_state.gamenames[-1]
+    favorite_game = searchbox(None)                                                     # search box to add a user favorite game on Plot 1
+    fav_games = add_list([favorite_game], rec_games)
+    fav_options = st.multiselect('Select Recommended Games', fav_games)
+    
     options = top_games
-    selected_options = st.multiselect('Select Video Games', options)
+    selected_tops = st.multiselect('Select Video Games', options)
+    selected_options = add_top_games(selected_tops, fav_options, ranges, df_ax) 
 
     # Plot 1
     title_names = ','.join(selected_options)
@@ -359,11 +492,12 @@ def exec_page_pub(emoji, theme, main_genre):
     top_games = df_ax.gamename.unique()[ranges[0]-1:ranges[1]]
 
     # Dataframe preview
-    dfbox(ax_name, y_name, df_ax)
+    rec_dfbox()
+    dfbox(ax_name, y_name, df_ax, ranges, order_name)
 
-    title = f"1.2 {ranges[1]} :blue[{ax_name}] Games with the :red[{order_name}] Monthly :blue[{y_name}]:"
+    title = f"1.2 5 :blue[{theme}s] with the :red[{order_name}] Monthly :blue[{y_name}]:"
     st.subheader(title)
-    st.write(top_games)
+    st.dataframe(genres[0:5])
 
 
 

module/__selectpage__.py

@@ -2,18 +2,13 @@ import streamlit as st
 from streamlit_extras.switch_page_button import switch_page 
 import os
 
-# pages = ["Overall"]
+pages = ["Overall"]
 
-# files = os.listdir('.\pages')
-# for name in files:
-#     name = name.split()[1].replace('.py', '').replace('_', ' ')
-#     pages.append(name)
+files = os.listdir('.\pages')
+for name in files:
+    name = name.split()[1].replace('.py', '').replace('_', ' ')
+    pages.append(name)
 
-pages = [
-    "Overall", 
-    "Category", "Genre", "DLC Counts", "Price", "Age Restriction", 
-    "Settings", "OS", "Publisher", "Developer", 
-]
 
 def st_page_selectbox(current_page):
     current_index = pages.index(current_page)

pages/{01 Category.py → 02 Category.py}

@@ -5,6 +5,8 @@ from module.__custom__ import *
 theme = "Category"
 page_genres = ['multi_player', 'pvp', 'co-op', 'single_player']
 add_genre = ['single_player', 'multi_player']
+labels = ['pvp', 'co-op', 'single_player']
 
 add_opp_features(add_genre)
-exec_page('🤼', theme, page_genres)
+exec_page('🤼', theme, page_genres)
+plot3_box(theme, labels)

pages/03 DLC_Counts.py

@@ -11,8 +11,11 @@ add_genres = [[0, 1, 'dlc_0', 'dlc_count'],
               [41, 101, 'dlc_100', 'dlc_count'],
               [101, 301, 'dlc_300', 'dlc_count'],
               [301, 9999, 'dlc_most', 'dlc_count']]
+labels = 'dlc_count'
+
 
 # ### adding single-player feature ###
 for add in add_genres:
        add_range_features(add)
-exec_page('🪜', theme, page_genres)
+exec_page('🪜', theme, page_genres)
+plot3_box(theme, labels)

pages/{02 Genre.py → 04 Genre.py}

@@ -7,6 +7,16 @@ page_genres = ['genre_action', 'genre_adventure', 'genre_casual',
        'genre_sexual_content', 'genre_strategy', 'genre_sports',
        'genre_racing', 'genre_rpg', 'genre_simulation', 'indie', 'mainstream']
 add_genre = ['mainstream', 'indie']
+labels = ['genre_action', 'genre_adventure', 'genre_casual',
+       'genre_sexual_content', 'genre_strategy', 'genre_sports',
+       'genre_racing', 'genre_rpg', 'genre_simulation']
+labels_2 = add_genre
 
 add_opp_features(add_genre)
-exec_page('🦹‍♀️', theme, page_genres)
+exec_page('🦹‍♀️', theme, page_genres)
+
+tab1, tab2= st.tabs(['Genre ratio', 'Indie versus mainstream ratio'])
+with tab1:
+       plot3_box(theme, labels)
+with tab2:
+       plot3_box(theme, labels_2)

pages/{04 Price.py → 05 Price.py}

@@ -12,8 +12,10 @@ add_genres = [[0, 0.001, 'price_free', 'price'],
               [20.001, 40.001, 'price_40', 'price'],
               [40.001, 60.001, 'price_60', 'price'], 
               [60.001, 9999, 'price_most', 'price']]
+labels = 'price'
 
 # ### adding single-player feature ###
 for add in add_genres:
        add_range_features(add)
-exec_page('💸', theme, page_genres)
+exec_page('💸', theme, page_genres)
+plot3_box(theme, labels)

pages/{05 Age_Restriction.py → 06 Age_Restriction.py}

@@ -1,7 +1,10 @@
 import streamlit as st
 
-from module.__custom__ import exec_page
+from module.__custom__ import *
 
 theme = "Age Restriction"
 page_genres = ['age_0_plus', 'age_13_plus', 'age_18_plus']
-exec_page('👴', theme, page_genres)
+labels = page_genres
+
+exec_page('👴', theme, page_genres)
+plot3_box(theme, labels)

pages/{06 Settings.py → 07 Settings.py}

@@ -5,8 +5,16 @@ from module.__custom__ import *
 theme = "Settings"
 page_genres = ['full_audio', 'full_controller_support', 'not_full_audio', 'not_full_controller']
 add_genres = [['not_full_audio', 'full_audio'], ['not_full_controller', 'full_controller_support'] ]
+labels = ['not_full_controller', 'full_controller_support']
+labels_2 = ['not_full_audio', 'full_audio']
 
 # ### adding single-player feature ###
 for add in add_genres:
        add_opp_features(add)
-exec_page('🎧', theme, page_genres)
+exec_page('🎧', theme, page_genres)
+tab1, tab2= st.tabs(['Controller support ratio', 'Full audio ratio'])
+with tab1:
+       plot3_box(theme, labels)
+with tab2:
+       plot3_box(theme, labels_2)
+

pages/{07 OS.py → 08 OS.py}

@@ -5,8 +5,18 @@ from module.__custom__ import *
 theme = "OS"
 page_genres = ['windows', 'mac', 'linux', 'not_windows', 'not_mac', 'not_linux']
 add_genres = [['not_windows', 'windows'], ['not_mac', 'mac'], ['not_linux', 'linux'] ]
+labels = ['not_windows', 'windows']
+labels_2 = ['not_mac', 'mac']
+labels_3 = ['not_linux', 'linux']
 
 # ### adding single-player feature ###
 for add in add_genres:
        add_opp_features(add)
-exec_page('🖥', theme, page_genres)
+exec_page('🖥', theme, page_genres)
+tab1, tab2, tab3 = st.tabs(['Windows game ratio', 'Mac game ratio', 'Linux game ratio'])
+with tab1:
+       plot3_box(theme, labels)
+with tab2:
+       plot3_box(theme, labels_2)
+with tab3:
+       plot3_box(theme, labels_3)

pages/{08 Publisher.py → 09 Publisher.py}

@@ -4,5 +4,7 @@ from module.__custom__ import *
 
 theme = "Publisher"
 main_genre = "publishers"
+labels = main_genre
 
-exec_page_pub('🛒', theme, main_genre)
+exec_page_pub('🛒', theme, main_genre)
+plot3_box_limit(theme, labels, 0.01)

pages/{09 Developer.py → 10 Developer.py}

@@ -4,5 +4,7 @@ from module.__custom__ import *
 
 theme = "Developer"
 main_genre = "developers"
+labels = main_genre
 
-exec_page_pub('💽', theme, main_genre)
+exec_page_pub('💽', theme, main_genre)
+plot3_box_limit(theme, labels, 0.005)