Minor layout changes

bertrend/demos/demos_utils/data_loading_component.py

@@ -173,34 +173,38 @@ def display_data_loading_component():
         # Deduplicate using all columns
         df = df.drop_duplicates()
 
-        # Select timeframe
-        min_date, max_date = df["timestamp"].dt.date.agg(["min", "max"])
-        register_widget("timeframe_slider")
-        start_date, end_date = st.slider(
-            "Select Timeframe",
-            min_value=min_date,
-            max_value=max_date,
-            value=(min_date, max_date),
-            key="timeframe_slider",
-            on_change=save_widget_state,
-        )
-
-        # Filter and sample the DataFrame
-        df_filtered = df[
-            (df["timestamp"].dt.date >= start_date)
-            & (df["timestamp"].dt.date <= end_date)
-        ]
-        df_filtered = df_filtered.sort_values(by="timestamp").reset_index(drop=True)
+        col1, col2 = st.columns([0.8, 0.2])
+        with col1:
+            # Select timeframe
+            min_date, max_date = df["timestamp"].dt.date.agg(["min", "max"])
+            register_widget("timeframe_slider")
+            start_date, end_date = st.slider(
+                "Select Timeframe",
+                min_value=min_date,
+                max_value=max_date,
+                value=(min_date, max_date),
+                key="timeframe_slider",
+                on_change=save_widget_state,
+            )
+
+            # Filter and sample the DataFrame
+            df_filtered = df[
+                (df["timestamp"].dt.date >= start_date)
+                & (df["timestamp"].dt.date <= end_date)
+            ]
+            df_filtered = df_filtered.sort_values(by="timestamp").reset_index(drop=True)
+
+        with col2:
+            register_widget("sample_size")
+            sample_size = st.number_input(
+                "Sample Size",
+                value=SAMPLE_SIZE_DEFAULT or len(df_filtered),
+                min_value=1,
+                max_value=len(df_filtered),
+                key="sample_size",
+                on_change=save_widget_state,
+            )
 
-        register_widget("sample_size")
-        sample_size = st.number_input(
-            "Sample Size",
-            value=SAMPLE_SIZE_DEFAULT or len(df_filtered),
-            min_value=1,
-            max_value=len(df_filtered),
-            key="sample_size",
-            on_change=save_widget_state,
-        )
         if sample_size < len(df_filtered):
             df_filtered = df_filtered.sample(n=sample_size, random_state=42)
 

bertrend/demos/demos_utils/state_utils.py

@@ -21,6 +21,11 @@ def register_widget(key):
         st.session_state[STATE_KEYS].append(key)
 
 
+def register_multiple_widget(*keys):
+    for key in keys:
+        register_widget(key)
+
+
 def save_widget_state():
     if STATE_KEYS in st.session_state.keys():
         st.session_state[WIDGET_STATE] = {

bertrend/demos/topic_analysis/data_utils.py

@@ -17,9 +17,11 @@
 from bertrend.utils.data_loading import TEXT_COLUMN, TIMESTAMP_COLUMN
 
 
-def data_overview(df: pd.DataFrame):
-    with st.container(border=True):
-        col1, col2 = st.columns([0.4, 0.6])
+def data_distribution(df: pd.DataFrame):
+    with st.expander(
+        label="Data distribution",
+        expanded=False,
+    ):
         freq = st.select_slider(
             "Time aggregation",
             options=(
@@ -34,16 +36,8 @@ def data_overview(df: pd.DataFrame):
             ),
             value="1M",
         )
-        with col1:
-            fig = plot_docs_repartition_over_time(df, freq)
-            st.plotly_chart(
-                fig, config=PLOTLY_BUTTON_SAVE_CONFIG, use_container_width=True
-            )
-        with col2:
-            st.dataframe(
-                st.session_state["time_filtered_df"][[TEXT_COLUMN, TIMESTAMP_COLUMN]],
-                use_container_width=True,
-            )
+        fig = plot_docs_repartition_over_time(df, freq)
+        st.plotly_chart(fig, config=PLOTLY_BUTTON_SAVE_CONFIG, use_container_width=True)
 
 
 def choose_data(base_dir: Path, filters: List[str]):

bertrend/demos/topic_analysis/demo_pages/explore_topics.py

@@ -104,9 +104,9 @@ def display_topic_info():
     ]["Representation"]
 
     st.write(
-        f"# Topic {st.session_state['selected_topic_number']} : {topic_docs_number} documents"
+        f"## Topic {st.session_state['selected_topic_number']} : {topic_docs_number} documents"
     )
-    st.markdown(f"## #{' #'.join(topic_words)}")
+    st.markdown(f"### #{' #'.join(topic_words)}")
 
 
 def plot_topic_over_time():
@@ -232,7 +232,6 @@ def display_representative_documents(filtered_df):
                 st.link_button(content, doc.url)
             else:
                 st.markdown(content)
-            st.divider()
 
 
 def display_new_documents():
@@ -283,57 +282,76 @@ def create_topic_documents(
     return folder_name, documents
 
 
+def _display_topic_description(filtered_df):
+    # GPT description button
+    if st.button(
+        "Generate a short description of the topic",
+        type="primary",
+        use_container_width=True,
+    ):
+        with st.spinner("Génération de la description en cours..."):
+            language_code = (
+                "fr" if SessionStateManager.get("language") == "French" else "en"
+            )
+            gpt_description = generate_topic_description(
+                st.session_state["topic_model"],
+                st.session_state["selected_topic_number"],
+                filtered_df,
+                language_code=language_code,
+            )
+        with st.container(border=True):
+            st.markdown(gpt_description)
+
+
 def main():
     """Main function to run the Streamlit topic_analysis."""
     check_model_and_prepare_topics()
+
+    st.title("Topics exploration")
+
     display_sidebar()
 
     if "selected_topic_number" not in st.session_state:
         st.stop()
 
-    display_topic_info()
-    plot_topic_over_time()
-
-    st.divider()
-
-    # Number of articles to display
-    top_n_docs = st.number_input(
-        "Number of articles to display",
-        min_value=1,
-        max_value=st.session_state["topics_info"].iloc[
-            st.session_state["selected_topic_number"]
-        ]["Count"],
-        value=st.session_state["topics_info"].iloc[
-            st.session_state["selected_topic_number"]
-        ]["Count"],
-        step=1,
-    )
-
-    representative_df = get_representative_documents(top_n_docs)
-    representative_df = representative_df.sort_values(by="timestamp", ascending=False)
+    col1, col2 = st.columns([0.3, 0.7])
+    with col1:
+        display_topic_info()
+    with col2:
+        plot_topic_over_time()
 
-    # Get unique sources
-    sources = representative_df[URL_COLUMN].apply(get_website_name).unique()
+    col1, col2 = st.columns(2)
+    with col1:
+        # Number of articles to display
+        top_n_docs = st.number_input(
+            "Number of articles to display",
+            min_value=1,
+            max_value=st.session_state["topics_info"].iloc[
+                st.session_state["selected_topic_number"]
+            ]["Count"],
+            value=st.session_state["topics_info"].iloc[
+                st.session_state["selected_topic_number"]
+            ]["Count"],
+            step=1,
+        )
+    with col2:
+        representative_df = get_representative_documents(top_n_docs)
+        representative_df = representative_df.sort_values(
+            by="timestamp", ascending=False
+        )
 
-    # Multi-select for sources
-    selected_sources = st.multiselect(
-        "Select the sources to display",
-        options=["All"] + list(sources),
-        default=["All"],
-    )
+        # Get unique sources
+        sources = representative_df[URL_COLUMN].apply(get_website_name).unique()
 
-    """
-    # Filter the dataframe based on selected sources
-    if "All" not in selected_sources:
-        filtered_df = representative_df[
-            representative_df[URL_COLUMN].apply(get_website_name).isin(selected_sources)
-        ]
-    else:
-        filtered_df = representative_df
-    """
+        # Multi-select for sources
+        selected_sources = st.multiselect(
+            "Select the sources to display",
+            options=["All"] + list(sources),
+            default=["All"],
+        )
 
     # Create two columns
-    col1, col2 = st.columns([0.5, 0.5])
+    col1, col2 = st.columns([0.3, 0.7])
 
     with col1:
         # Pass the full representative_df to display_source_distribution
@@ -353,26 +371,8 @@ def main():
 
     display_new_documents()
 
-    st.divider()
-
-    # GPT description button
-    if st.button(
-        "Generate a short description of the topic",
-        type="primary",
-        use_container_width=True,
-    ):
-        with st.spinner("Génération de la description en cours..."):
-            language_code = (
-                "fr" if SessionStateManager.get("language") == "French" else "en"
-            )
-            gpt_description = generate_topic_description(
-                st.session_state["topic_model"],
-                st.session_state["selected_topic_number"],
-                filtered_df,
-                language_code=language_code,
-            )
-        with st.container(border=True):
-            st.markdown(gpt_description)
+    # GPT generated topic description
+    _display_topic_description(filtered_df)
 
     st.divider()
 

bertrend/demos/topic_analysis/demo_pages/temporal_visualizations.py

@@ -23,6 +23,7 @@
     register_widget,
     save_widget_state,
     restore_widget_state,
+    register_multiple_widget,
 )
 from bertrend.demos.weak_signals.visualizations_utils import PLOTLY_BUTTON_SAVE_CONFIG
 from bertrend.utils.data_loading import TIMESTAMP_COLUMN, TEXT_COLUMN
@@ -411,30 +412,48 @@ def format_timedelta(td):
 
 def select_time_granularity(max_granularity):
     """Allow user to select custom time granularity within limits."""
-    st.write("Select custom time granularity:")
+    st.write("Select custom time granularity")
     col1, col2, col3, col4 = st.columns(4)
 
     max_days = max_granularity.days
-
+    register_multiple_widget(
+        "granularity_days", "granularity_hours", "granularity_minutes"
+    )
     with col1:
-        days = st.number_input(
+        days = st.slider(
             "Days",
             min_value=0,
             max_value=max_days,
             value=min(1, max_days),
             key="granularity_days",
+            on_change=save_widget_state,
         )
     with col2:
-        hours = st.number_input(
-            "Hours", min_value=0, max_value=23, value=0, key="granularity_hours"
+        hours = st.slider(
+            "Hours",
+            min_value=0,
+            max_value=23,
+            value=0,
+            key="granularity_hours",
+            on_change=save_widget_state,
         )
     with col3:
-        minutes = st.number_input(
-            "Minutes", min_value=0, max_value=59, value=0, key="granularity_minutes"
+        minutes = st.slider(
+            "Minutes",
+            min_value=0,
+            max_value=59,
+            value=0,
+            key="granularity_minutes",
+            on_change=save_widget_state,
         )
     with col4:
-        seconds = st.number_input(
-            "Seconds", min_value=0, max_value=59, value=0, key="granularity_seconds"
+        seconds = st.slider(
+            "Seconds",
+            min_value=0,
+            max_value=59,
+            value=0,
+            key="granularity_seconds",
+            on_change=save_widget_state,
         )
 
     selected_granularity = timedelta(
@@ -743,6 +762,8 @@ def main():
     # Check if model is trained
     check_model_trained()
 
+    st.title("Temporal visualizations of topics")
+
     # Display sidebar
     display_sidebar()
 

bertrend/demos/topic_analysis/demo_pages/training_page.py

@@ -36,7 +36,7 @@
     display_bertopic_hyperparameters,
 )
 from bertrend.demos.demos_utils.session_state_manager import SessionStateManager
-from bertrend.demos.topic_analysis.data_utils import data_overview
+from bertrend.demos.topic_analysis.data_utils import data_distribution
 from bertrend.metrics.topic_metrics import compute_cluster_metrics
 from bertrend.parameters import BERTOPIC_SERIALIZATION
 from bertrend.topic_model import TopicModel
@@ -172,7 +172,7 @@ def main():
     # Data overview
     if "time_filtered_df" not in st.session_state:
         st.stop()
-    data_overview(st.session_state["time_filtered_df"])
+    data_distribution(st.session_state["time_filtered_df"])
     SessionStateManager.set("split_type", st.session_state["split_by_paragraph"])
 
     # Embed documents