From d909b499d8f8556ac54982021f595b26e15698c6 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Thu, 9 Jan 2025 22:39:15 +0000
Subject: [PATCH 01/19] enable all xbins for plotly histogram

---
 code/pages/0_Data inventory.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index cd07753..66a726e 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -57,6 +57,14 @@
     "VAST_raw_data_on_VAST",
 ] + [query["alias"] for query in QUERY_PRESET]
 
+X_BIN_SIZE_MAPPER = {  # For plotly histogram xbins
+    "Daily": 1000*3600*24,  # Milliseconds
+    "Weekly": 1000*3600*24*7, # Milliseconds
+    "Monthly": "M1",
+    "Quarterly": "M4",
+}
+
+
 @st.cache_data(ttl=3600*12)
 def merge_queried_dfs(dfs, queries_to_merge):
     # Combine queried dfs using df_unique_mouse_date (on index "subject_id", "session_date" only)
@@ -270,7 +278,7 @@ def count_true_values(df, time_period, column):
         for i, column in enumerate(columns):
             fig.add_trace(go.Histogram( 
                 x=df[df[column]==True]["session_date"],
-                xbins=dict(size="M1"), # Only monthly bins look good
+                xbins=dict(size=X_BIN_SIZE_MAPPER[time_period]), # Only monthly bins look good
                 name=column,
                 marker_color=colors[i],
                 opacity=0.75
@@ -430,7 +438,6 @@ def app():
                 "Bin size",
                 ["Daily", "Weekly", "Monthly", "Quarterly"],
                 index=1,
-                disabled=not if_separate_plots,
             )
 
         for i_venn, venn_preset in enumerate(VENN_PRESET):

From 77d758a3ad5b4772efc926f29030569a59d586dc Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Thu, 9 Jan 2025 22:41:31 +0000
Subject: [PATCH 02/19] caching presets

---
 code/pages/0_Data inventory.py | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index 66a726e..a14e201 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -48,8 +48,16 @@
 )
 
 # Load QUERY_PRESET from json
-with open("data_inventory_QUERY_PRESET.json", "r") as f:
-    QUERY_PRESET = json.load(f)
+@st.cache_data()
+def load_presets():
+    with open("data_inventory_QUERY_PRESET.json", "r") as f:
+        QUERY_PRESET = json.load(f)
+
+    with open("data_inventory_VENN_PRESET.json", "r") as f:
+        VENN_PRESET = json.load(f)
+    return QUERY_PRESET, VENN_PRESET
+
+QUERY_PRESET, VENN_PRESET = load_presets()
 
 META_COLUMNS = [
     "Han_temp_pipeline (bpod)",
@@ -421,9 +429,6 @@ def app():
     )
 
     # --- Venn diagram from presets ---
-    with open("data_inventory_VENN_PRESET.json", "r") as f:
-        VENN_PRESET = json.load(f)
-
     if VENN_PRESET:
 
         cols = st.columns([2, 1])

From b24040c89b22e8ec68ecdf42060251ed93d8b3fa Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Thu, 9 Jan 2025 22:48:33 +0000
Subject: [PATCH 03/19] change use_container_width

---
 code/util/aws_s3.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/code/util/aws_s3.py b/code/util/aws_s3.py
index 300dffc..9c69e32 100644
--- a/code/util/aws_s3.py
+++ b/code/util/aws_s3.py
@@ -94,7 +94,7 @@ def show_session_level_img_by_key_and_prefix(key, prefix, column=None, other_pat
     _f.image(img if img is not None else "https://cdn-icons-png.flaticon.com/512/3585/3585596.png", 
                 output_format='PNG', 
                 caption=f_name.split('/')[-1] if caption and f_name else '',
-                use_column_width='always',
+                use_container_width='always',
                 **kwargs)
 
     return img

From 6263024e29baf8ef2b69b9bc66b8e17a3d344efc Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:01:06 +0000
Subject: [PATCH 04/19] bump versions

---
 requirements.txt | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/requirements.txt b/requirements.txt
index bcb790a..a617adb 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,11 +1,11 @@
-streamlit==1.31.0
+streamlit==1.41.1
 streamlit-aggrid==0.3.5
 streamlit-bokeh3-events==0.1.4
 streamlit_dynamic_filters==0.1.9
 streamlit-nested-layout==0.1.1
 streamlit-plotly-events==0.0.6
-pygwalker==0.4.7
-extra-streamlit-components==0.1.56
+pygwalker==0.4.9.13
+extra-streamlit-components==0.1.71
 numpy==1.26.4
 pandas==2.2.2
 matplotlib==3.9.2

From 9e2e161408b9b2c46d2ca847b3d602b51f4a4f0c Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:15:59 +0000
Subject: [PATCH 05/19] Use fragment in the main page

---
 code/Home.py | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index f4c7646..8db30d1 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -585,9 +585,12 @@ def app():
     if len(st.session_state.df_session_filtered) == 0:
         st.markdown('## No filtered results!')
         return
-    
-    aggrid_outputs = aggrid_interactive_table_session(df=st.session_state.df_session_filtered, table_height=table_height)
-    
+
+    aggrid_outputs = aggrid_interactive_table_session(
+        df=st.session_state.df_session_filtered,
+        table_height=table_height,
+    )
+
     if len(aggrid_outputs['selected_rows']) and not set(pd.DataFrame(aggrid_outputs['selected_rows']
                                                                  ).set_index(['h2o', 'session']).index
                                                         ) == set(st.session_state.df_selected_from_dataframe.set_index(['h2o', 'session']).index):
@@ -596,6 +599,10 @@ def app():
         # if st.session_state.tab_id == "tab_session_x_y":
         st.rerun()
 
+    add_tabs()
+
+@st.fragment
+def add_tabs():
     chosen_id = stx.tab_bar(data=[
         stx.TabBarItemData(id="tab_auto_train_history", title="🎓 Automatic Training History", description="Track progress"),
         stx.TabBarItemData(id="tab_session_inspector", title="👀 Session Inspector", description="Select sessions from the table and show plots"),
@@ -658,7 +665,7 @@ def app():
                     spec="./gw_config.json",
                     )
                             
-            pygwalker_renderer.render_explore(height=1010, scrolling=False)
+            pygwalker_renderer.render_explore()
         
     elif chosen_id == "tab_session_inspector":
         with placeholder:

From f883477a179c0ff5c2e526c42e73834ce056b4a0 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:20:00 +0000
Subject: [PATCH 06/19] use fragment in Data inventory

---
 code/pages/0_Data inventory.py | 142 +++++++++++++++++----------------
 1 file changed, 73 insertions(+), 69 deletions(-)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index a14e201..71c5722 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -430,83 +430,87 @@ def app():
 
     # --- Venn diagram from presets ---
     if VENN_PRESET:
+        add_venn_diagrms(df_merged)
+        
+@st.fragment
+def add_venn_diagrms(df_merged):
+
+    cols = st.columns([2, 1])
+    cols[0].markdown("## Venn diagrams from presets")
+    with cols[1].expander("Time view settings", expanded=True):
+        cols_1 = st.columns([1, 1])                    
+        if_separate_plots = cols_1[0].checkbox("Separate in subplots", value=True)
+        if_sync_y_limits = cols_1[0].checkbox(
+            "Sync Y limits", value=True, disabled=not if_separate_plots
+        )
+        time_period = cols_1[1].selectbox(
+            "Bin size",
+            ["Daily", "Weekly", "Monthly", "Quarterly"],
+            index=1,
+        )
 
-        cols = st.columns([2, 1])
-        cols[0].markdown("## Venn diagrams from presets")
-        with cols[1].expander("Time view settings", expanded=True):
-            cols_1 = st.columns([1, 1])                    
-            if_separate_plots = cols_1[0].checkbox("Separate in subplots", value=True)
-            if_sync_y_limits = cols_1[0].checkbox(
-                "Sync Y limits", value=True, disabled=not if_separate_plots
+    for i_venn, venn_preset in enumerate(VENN_PRESET):
+        # -- Venn diagrams --
+        st.markdown(f"### ({i_venn+1}). {venn_preset['name']}")
+        fig, notes = generate_venn(
+                df_merged,
+                venn_preset
             )
-            time_period = cols_1[1].selectbox(
-                "Bin size",
-                ["Daily", "Weekly", "Monthly", "Quarterly"],
-                index=1,
+        for note in notes:
+            st.markdown(note)
+
+        cols = st.columns([1, 1])
+        with cols[0]:
+            st.pyplot(fig, use_container_width=True)
+
+        # -- Show and download df for this Venn --
+        circle_columns = [c_s["column"] for c_s in venn_preset["circle_settings"]]
+        # Show histogram over time for the columns and patches in preset
+        df_this_preset = df_merged[circle_columns]
+        # Filter out rows that have at least one True in this Venn
+        df_this_preset = df_this_preset[df_this_preset.any(axis=1)]
+
+        # Create a new column to indicate sessions in patches specified by patch_ids like ["100", "101", "110", "111"]
+        for patch_setting in venn_preset.get("patch_settings", []):
+            idx = _filter_df_by_patch_ids(
+                df_this_preset[circle_columns],
+                patch_setting["patch_ids"]
             )
+            df_this_preset.loc[idx, str(patch_setting["patch_ids"])] = True 
 
-        for i_venn, venn_preset in enumerate(VENN_PRESET):
-            # -- Venn diagrams --
-            st.markdown(f"### ({i_venn+1}). {venn_preset['name']}")
-            fig, notes = generate_venn(
-                    df_merged,
-                    venn_preset
-                )
-            for note in notes:
-                st.markdown(note)
-
-            cols = st.columns([1, 1])
-            with cols[0]:
-                st.pyplot(fig, use_container_width=True)
-
-            # -- Show and download df for this Venn --
-            circle_columns = [c_s["column"] for c_s in venn_preset["circle_settings"]]
-            # Show histogram over time for the columns and patches in preset
-            df_this_preset = df_merged[circle_columns]
-            # Filter out rows that have at least one True in this Venn
-            df_this_preset = df_this_preset[df_this_preset.any(axis=1)]
-
-            # Create a new column to indicate sessions in patches specified by patch_ids like ["100", "101", "110", "111"]
-            for patch_setting in venn_preset.get("patch_settings", []):
-                idx = _filter_df_by_patch_ids(
-                    df_this_preset[circle_columns],
-                    patch_setting["patch_ids"]
-                )
-                df_this_preset.loc[idx, str(patch_setting["patch_ids"])] = True 
+        # Join in other extra columns
+        df_this_preset = df_this_preset.join(
+            df_merged[[col for col in df_merged.columns if col not in META_COLUMNS]], how="left"
+        )
 
-            # Join in other extra columns
-            df_this_preset = df_this_preset.join(
-                df_merged[[col for col in df_merged.columns if col not in META_COLUMNS]], how="left"
+        with cols[0]:
+            download_df(
+                df_this_preset,
+                label="Download as CSV for this Venn diagram",
+                file_name=f"df_{venn_preset['name']}.csv",
             )
+            with st.expander(f"Show dataframe, n = {len(df_this_preset)}"):
+                st.write(df_this_preset)
 
-            with cols[0]:
-                download_df(
-                    df_this_preset,
-                    label="Download as CSV for this Venn diagram",
-                    file_name=f"df_{venn_preset['name']}.csv",
-                )
-                with st.expander(f"Show dataframe, n = {len(df_this_preset)}"):
-                    st.write(df_this_preset)
-
-            with cols[1]:
-                # -- Show histogram over time --
-                fig = plot_histogram_over_time(
-                    df=df_this_preset.reset_index(),
-                    venn_preset=venn_preset,
-                    time_period=time_period,
-                    if_sync_y_limits=if_sync_y_limits,
-                    if_separate_plots=if_separate_plots,
-                )
-                plotly_events(
-                    fig,
-                    click_event=False,
-                    hover_event=False,
-                    select_event=False,
-                    override_height=fig.layout.height * 1.1,
-                    override_width=fig.layout.width,
-                )
+        with cols[1]:
+            # -- Show histogram over time --
+            fig = plot_histogram_over_time(
+                df=df_this_preset.reset_index(),
+                venn_preset=venn_preset,
+                time_period=time_period,
+                if_sync_y_limits=if_sync_y_limits,
+                if_separate_plots=if_separate_plots,
+            )
+            plotly_events(
+                fig,
+                click_event=False,
+                hover_event=False,
+                select_event=False,
+                override_height=fig.layout.height * 1.1,
+                override_width=fig.layout.width,
+            )
 
-            st.markdown("---")
+        st.markdown("---")
 
     # --- User-defined Venn diagram ---
     # Multiselect for selecting queries up to three

From b7e25175e1c3111cd2840354a52195bf6442f17f Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:32:02 +0000
Subject: [PATCH 07/19] currucilum page: improve sorting

---
 code/Home.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/code/Home.py b/code/Home.py
index 8db30d1..b7ca704 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -688,7 +688,9 @@ def add_tabs():
 
     elif chosen_id == "tab_auto_train_curriculum":  # Automatic training curriculums
         df_curriculums = st.session_state.curriculum_manager.df_curriculums().sort_values(
-            by=['curriculum_schema_version', 'curriculum_name', 'curriculum_version']).reset_index().drop(columns='index')
+            by=['curriculum_version', 'curriculum_schema_version', 'curriculum_name'],
+            ascending=[False, True, False],
+            ).reset_index().drop(columns='index')
         with placeholder:
             # Show curriculum manager dataframe
             st.markdown("#### Select auto training curriculums")

From 30cbccd85b6245f04139a784e19a2a0fa35a58fe Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:48:15 +0000
Subject: [PATCH 08/19] improve selectbox_wrapper_for_url_query

---
 code/util/url_query_helper.py | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/code/util/url_query_helper.py b/code/util/url_query_helper.py
index e8fd43f..47f3396 100644
--- a/code/util/url_query_helper.py
+++ b/code/util/url_query_helper.py
@@ -95,17 +95,21 @@ def checkbox_wrapper_for_url_query(st_prefix, label, key, default, **kwargs):
         **kwargs,
     )
 
-def selectbox_wrapper_for_url_query(st_prefix, label, options, key, default, **kwargs):
+def selectbox_wrapper_for_url_query(st_prefix, label, options, key, default, default_override=True, **kwargs):
+    # If default_override, use default. Otherwise, session_state or query_params has higher priority
+    if not default_override:
+        default = (
+            st.session_state[key]
+            if key in st.session_state and st.session_state[key] in options
+            else st.query_params[key]
+            if key in st.query_params and st.query_params[key] in options
+            else default
+        )
+    
     return st_prefix.selectbox(
         label,
         options=options,
-        index=(
-            options.index(st.session_state[key])
-            if key in st.session_state
-            else options.index(st.query_params[key]) 
-            if key in st.query_params 
-            else options.index(default)
-        ),
+        index=options.index(default),
         key=key,
         **kwargs,
     )

From 202f791fbd00ed55e30ebad118db99ebcc42ccf0 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:49:02 +0000
Subject: [PATCH 09/19] set default_override=False by default

---
 code/util/url_query_helper.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/code/util/url_query_helper.py b/code/util/url_query_helper.py
index 47f3396..06702e5 100644
--- a/code/util/url_query_helper.py
+++ b/code/util/url_query_helper.py
@@ -95,7 +95,7 @@ def checkbox_wrapper_for_url_query(st_prefix, label, key, default, **kwargs):
         **kwargs,
     )
 
-def selectbox_wrapper_for_url_query(st_prefix, label, options, key, default, default_override=True, **kwargs):
+def selectbox_wrapper_for_url_query(st_prefix, label, options, key, default, default_override=False, **kwargs):
     # If default_override, use default. Otherwise, session_state or query_params has higher priority
     if not default_override:
         default = (

From 84397743146e46452b222cb959d18056ad11cfca Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 00:55:01 +0000
Subject: [PATCH 10/19] improve curriculum tab

---
 code/Home.py | 64 ++++++++++++++++++++++++----------------------------
 1 file changed, 29 insertions(+), 35 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index b7ca704..cae002c 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -38,6 +38,7 @@
                             add_footnote)
 from util.url_query_helper import (checkbox_wrapper_for_url_query,
                                    multiselect_wrapper_for_url_query,
+                                   selectbox_wrapper_for_url_query,
                                    number_input_wrapper_for_url_query,
                                    slider_wrapper_for_url_query,
                                    sync_session_state_to_URL,
@@ -689,8 +690,9 @@ def add_tabs():
     elif chosen_id == "tab_auto_train_curriculum":  # Automatic training curriculums
         df_curriculums = st.session_state.curriculum_manager.df_curriculums().sort_values(
             by=['curriculum_version', 'curriculum_schema_version', 'curriculum_name'],
-            ascending=[False, True, False],
-            ).reset_index().drop(columns='index')
+            ascending=[False, True, False], 
+            ).reset_index().drop(columns='index').query("curriculum_name != 'Dummy task'")
+        
         with placeholder:
             # Show curriculum manager dataframe
             st.markdown("#### Select auto training curriculums")
@@ -698,51 +700,43 @@ def add_tabs():
             # Curriculum drop down selector
             cols = st.columns([0.8, 0.5, 0.8, 4])
             cols[3].markdown(f"(aind_auto_train lib version = {auto_train_version})")
+            
             options = list(df_curriculums['curriculum_name'].unique())
-            selected_curriculum_name = cols[0].selectbox(
-                'Curriculum name', 
+            selected_curriculum_name = selectbox_wrapper_for_url_query(
+                st_prefix=cols[0],
+                label='Curriculum name',
                 options=options,
-                index=options.index(st.session_state['auto_training_curriculum_name'])
-                    if ('auto_training_curriculum_name' in st.session_state) and (st.session_state['auto_training_curriculum_name'] != '') else 
-                    options.index(st.query_params['auto_training_curriculum_name'])
-                    if 'auto_training_curriculum_name' in st.query_params and st.query_params['auto_training_curriculum_name'] != ''
-                    else 0, 
-                key='auto_training_curriculum_name'
-                )
-            
+                default=options[0],
+                default_override=True,
+                key='auto_training_curriculum_name',
+            )
+                       
             options = list(df_curriculums[
                 df_curriculums['curriculum_name'] == selected_curriculum_name
                 ]['curriculum_version'].unique())
-            if ('auto_training_curriculum_version' in st.session_state) and (st.session_state['auto_training_curriculum_version'] in options):
-                default = options.index(st.session_state['auto_training_curriculum_version'])
-            elif 'auto_training_curriculum_version' in st.query_params and st.query_params['auto_training_curriculum_version'] in options:
-                default = options.index(st.query_params['auto_training_curriculum_version'])
-            else:
-                default = 0
-            selected_curriculum_version = cols[1].selectbox(
-                'Curriculum version', 
-                options=options, 
-                index=default, 
-                key='auto_training_curriculum_version'
+            selected_curriculum_version = selectbox_wrapper_for_url_query(
+                st_prefix=cols[1],
+                label='Curriculum version',
+                options=options,
+                default=options[0],
+                default_override=True,
+                key='auto_training_curriculum_version',
             )
             
             options = list(df_curriculums[
                 (df_curriculums['curriculum_name'] == selected_curriculum_name) 
                 & (df_curriculums['curriculum_version'] == selected_curriculum_version)
                 ]['curriculum_schema_version'].unique())
-            if ('auto_training_curriculum_schema_version' in st.session_state) and (st.session_state['auto_training_curriculum_schema_version'] in options):
-                default = options.index(st.session_state['auto_training_curriculum_schema_version'])
-            elif 'auto_training_curriculum_schema_version' in st.query_params and st.query_params['auto_training_curriculum_schema_version'] in options:
-                default = options.index(st.query_params['auto_training_curriculum_schema_version'])
-            else:
-                default = 0
-            selected_curriculum_schema_version = cols[2].selectbox(
-                'Curriculum schema version', 
+            
+            selected_curriculum_schema_version = selectbox_wrapper_for_url_query(
+                st_prefix=cols[2],
+                label='Curriculum schema version',
                 options=options,
-                index=default,
-                key='auto_training_curriculum_schema_version'
-                )
-                        
+                default=options[0],
+                default_override=True,
+                key='auto_training_curriculum_schema_version',
+            )
+                                   
             selected_curriculum = st.session_state.curriculum_manager.get_curriculum(
                 curriculum_name=selected_curriculum_name,
                 curriculum_schema_version=selected_curriculum_schema_version,

From 74ff41cd41b20f99e04d7c269d380891cf9f92a5 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 01:54:24 +0000
Subject: [PATCH 11/19] minor refactor

---
 code/Home.py | 43 +++++++++++++++++++++----------------------
 1 file changed, 21 insertions(+), 22 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index cae002c..86ca1e5 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -72,7 +72,7 @@ def _user_name_mapper(user_name):
                 return name
     else:
         return user_name
-    
+
 
 @st.cache_resource(ttl=24*3600)
 def get_pyg_renderer(df, spec="./gw_config.json", **kwargs) -> "StreamlitRenderer":
@@ -132,22 +132,20 @@ def session_plot_settings(need_click=True):
     with st.form(key='session_plot_settings'):
         st.markdown('##### Show plots for individual sessions ')
         cols = st.columns([2, 6, 1])
-        
+
         session_plot_modes = [f'sessions selected from table or plot', f'all sessions filtered from sidebar']
-        st.session_state.selected_draw_sessions = cols[0].selectbox(f'Which session(s) to draw?', 
-                                                                    session_plot_modes,
-                                                                    index=session_plot_modes.index(st.session_state['session_plot_mode'])
-                                                                        if 'session_plot_mode' in st.session_state else 
-                                                                        session_plot_modes.index(st.query_params['session_plot_mode'])
-                                                                        if 'session_plot_mode' in st.query_params 
-                                                                        else 0, 
-                                                                    key='session_plot_mode',
-                                                                )
-        
+        st.session_state.selected_draw_sessions = selectbox_wrapper_for_url_query(
+            cols[0],
+            label='Which session(s) to draw?',
+            options=session_plot_modes,
+            default=session_plot_modes[0],
+            key='session_plot_mode',
+        )
+
         n_session_to_draw = len(st.session_state.df_selected_from_plotly) \
             if 'selected from table or plot' in st.session_state.selected_draw_sessions \
             else len(st.session_state.df_session_filtered) 
-                
+
         _ = number_input_wrapper_for_url_query(
             st_prefix=cols[2],
             label='number of columns',
@@ -156,7 +154,7 @@ def session_plot_settings(need_click=True):
             default=3,
             key='session_plot_number_cols',
         )
-        
+
         st.markdown(
         """
         <style>
@@ -173,18 +171,19 @@ def session_plot_settings(need_click=True):
             default=draw_type_mapper_session_level.keys(),
             key='session_plot_selected_draw_types',
         )
-                
+
         cols[0].markdown(f'{n_session_to_draw} sessions to draw')
         draw_it_now_override = cols[2].checkbox('Auto show', value=not need_click, disabled=not need_click)
-        submitted = cols[0].form_submit_button("Update settings", type='primary')
-        
-        
+        submitted = cols[0].form_submit_button(
+            "Update settings", type="primary"
+        )
+
     if not need_click:
         return True
-        
+
     if draw_it_now_override:
         return True
-    
+
     draw_it = st.button(f'Show {n_session_to_draw} sessions!', use_container_width=False, type="primary")
     return draw_it
 
@@ -600,10 +599,10 @@ def app():
         # if st.session_state.tab_id == "tab_session_x_y":
         st.rerun()
 
-    add_tabs()
+    add_main_tabs()
 
 @st.fragment
-def add_tabs():
+def add_main_tabs():
     chosen_id = stx.tab_bar(data=[
         stx.TabBarItemData(id="tab_auto_train_history", title="🎓 Automatic Training History", description="Track progress"),
         stx.TabBarItemData(id="tab_session_inspector", title="👀 Session Inspector", description="Select sessions from the table and show plots"),

From 4690416a65ba19bc84c4bd75991b0db5ffd65898 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 05:18:01 +0000
Subject: [PATCH 12/19] fix styling

---
 code/Home.py           | 39 ++++++++++++++++++++++++---------------
 code/util/settings.py  | 24 ++++++++++++++++++++++++
 code/util/streamlit.py | 12 ++++--------
 3 files changed, 52 insertions(+), 23 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index 86ca1e5..ebb964a 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -192,7 +192,7 @@ def plot_x_y_session():
     with st.expander("X-Y plot settings", expanded=True):            
         with st.form(key='x_y_plot_settings', border=False):
             cols = st.columns([1, 1, 1])
-            
+
             with cols[0]:
                 x_name, y_name, group_by = add_xy_selector(if_bonsai=True)
 
@@ -201,28 +201,28 @@ def plot_x_y_session():
                 if_aggr_all, aggr_method_all, if_use_x_quantile_all, q_quantiles_all, smooth_factor, if_show_diagonal,
                 dot_size, dot_opacity, line_width, x_y_plot_figure_width, x_y_plot_figure_height, 
                 font_size_scale, color_map) = add_xy_setting()
-            
+
             if st.session_state.x_y_plot_if_show_dots:
                 with cols[2]:
                     size_mapper, size_mapper_range, size_mapper_gamma = add_dot_property_mapper()
             else:
                 size_mapper = 'None'
                 size_mapper_range, size_mapper_gamma = None, None
-            
+
             submitted = st.form_submit_button("👉 Update X-Y settings 👈", type='primary')
-    
+
     # If no sessions are selected, use all filtered entries
     # df_x_y_session = st.session_state.df_selected_from_dataframe if if_plot_only_selected_from_dataframe else st.session_state.df_session_filtered
     df_x_y_session = st.session_state.df_session_filtered
-    
+
     names = {('session', 'foraging_eff'): 'Foraging efficiency',
              ('session', 'finished'):   'Finished trials', 
              }
 
     df_selected_from_plotly = pd.DataFrame()
     # for i, (title, (x_name, y_name)) in enumerate(names.items()):
-        # with cols[i]:
-    
+    # with cols[i]:
+
     if hasattr(st.session_state, 'x_y_plot_figure_width'):
         _x_y_plot_scale = st.session_state.x_y_plot_figure_width / 1300
         cols = st.columns([1 * _x_y_plot_scale, 0.7])
@@ -256,19 +256,28 @@ def plot_x_y_session():
                                     font_size_scale=font_size_scale,
                                     color_map=color_map,
                                     )
-        
-        # st.plotly_chart(fig)
-        selected = plotly_events(fig, click_event=True, hover_event=False, select_event=True, 
-                                 override_height=fig.layout.height * 1.1, override_width=fig.layout.width)
-    
+
+        selected = st.plotly_chart(fig, 
+                                   on_select="rerun",
+                                   use_container_width=True,
+                                   theme=None,  # full controlled by plotly chart itself
+                        )
+        # selected = plotly_events(fig, click_event=True, hover_event=False, select_event=True,
+        #                          override_height=fig.layout.height * 1.1, override_width=fig.layout.width)
+
     with cols[1]:
         st.markdown('#### 👀 Quick preview')
         st.markdown('###### Click on one session to preview here, or Box/Lasso select multiple sessions to draw them in the section below')
         st.markdown('(sometimes you have to click twice...)')
-      
+
     if len(selected):
-        df_selected_from_plotly = df_x_y_session.merge(pd.DataFrame(selected).rename({'x': x_name, 'y': y_name}, axis=1), 
-                                                    on=[x_name, y_name], how='inner')
+        df_key_selected = pd.DataFrame(
+            [data["customdata"][:2] for data in selected.selection.points],
+            columns=["subject_id", "session_date"],
+        )
+        df_key_selected["session_date"] = pd.to_datetime(df_key_selected["session_date"])
+        df_selected_from_plotly = df_x_y_session.merge(df_key_selected, on=["subject_id", "session_date"], how='inner')
+        
     if len(st.session_state.df_selected_from_plotly) == 1:
         with cols[1]:
             draw_session_plots_quick_preview(st.session_state.df_selected_from_plotly)
diff --git a/code/util/settings.py b/code/util/settings.py
index 44b05ab..10cca02 100644
--- a/code/util/settings.py
+++ b/code/util/settings.py
@@ -1,3 +1,5 @@
+import plotly.io as pio
+
 # Setting up layout for each session
 draw_type_layout_definition = [
     [1],  # columns in the first row
@@ -33,3 +35,25 @@
 draw_types_quick_preview = [
     '1. Choice history',
     '3. Logistic regression (Su2022)']
+
+
+def override_plotly_theme(fig, theme):
+    """
+    Fix the problem that simply using fig.update_layout(template=theme) doesn't work with st.plotly_chart.
+    I have to use update_layout to explicitly set the theme.
+    """
+    
+    dict_plotly_template = pio.templates[theme].layout.to_plotly_json()
+    fig.update_layout(**dict_plotly_template)  # First apply the plotly official theme
+    
+    fig.update_layout(font_family="Arial")  # Add user-defined styling
+    
+    fig.update_xaxes(showline=True, linewidth=2, linecolor='black', showgrid=True, zeroline=True,
+                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ')
+
+    fig.update_yaxes(showline=True, linewidth=2, linecolor='black', showgrid=True, zeroline=True,
+                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ',
+                     title_standoff=40,
+                     )
+
+    return
\ No newline at end of file
diff --git a/code/util/streamlit.py b/code/util/streamlit.py
index 941a3f4..dc1cdec 100644
--- a/code/util/streamlit.py
+++ b/code/util/streamlit.py
@@ -27,6 +27,7 @@
                                multiselect_wrapper_for_url_query,
                                selectbox_wrapper_for_url_query,
                                slider_wrapper_for_url_query)
+from.settings import override_plotly_theme
 
 custom_css = {
 ".ag-root.ag-unselectable.ag-layout-normal": {"font-size": "15px !important",
@@ -1125,6 +1126,8 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
     n_mice = len(df['h2o'].unique())
     n_sessions = len(df.groupby(['h2o', 'session']).count())
 
+    override_plotly_theme(fig, theme="simple_white")
+
     fig.update_layout(
         width=x_y_plot_figure_width,
         height=x_y_plot_figure_height,
@@ -1142,14 +1145,7 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
                     b=130 * font_size_scale, 
                     t=100 * font_size_scale,
                     ),
-    )
-    fig.update_xaxes(showline=True, linewidth=2, linecolor='black', 
-                    #  range=[1, min(100, df[x_name].max())],
-                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ')
-
-    fig.update_yaxes(showline=True, linewidth=2, linecolor='black',
-                     title_standoff=40,
-                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ')
+    ) 
     return fig
 
 

From 30f3f6dd6c2f4f1a6e4755d6429e65d99ddd4630 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Fri, 10 Jan 2025 06:30:44 +0000
Subject: [PATCH 13/19] improve performance by fixing plotly_chart selection
 logic

---
 code/Home.py                            | 26 +++++++++------------
 code/pages/1_Basic behavior analysis.py |  1 -
 code/util/streamlit.py                  | 31 ++++++++++++++-----------
 3 files changed, 29 insertions(+), 29 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index ebb964a..a38ff4f 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -23,7 +23,6 @@
 from aind_auto_train.auto_train_manager import DynamicForagingAutoTrainManager
 from aind_auto_train.curriculum_manager import CurriculumManager
 from pygwalker.api.streamlit import StreamlitRenderer, init_streamlit_comm
-from streamlit_plotly_events import plotly_events
 from util.aws_s3 import (draw_session_plots_quick_preview, load_data,
                          show_debug_info,
                          show_session_level_img_by_key_and_prefix)
@@ -128,7 +127,7 @@ def draw_session_plots(df_to_draw_session):
                 my_bar.progress(int((i + 1) / len(df_to_draw_session) * 100))
 
 
-def session_plot_settings(need_click=True):
+def session_plot_settings(df_selected, need_click=True):
     with st.form(key='session_plot_settings'):
         st.markdown('##### Show plots for individual sessions ')
         cols = st.columns([2, 6, 1])
@@ -142,7 +141,7 @@ def session_plot_settings(need_click=True):
             key='session_plot_mode',
         )
 
-        n_session_to_draw = len(st.session_state.df_selected_from_plotly) \
+        n_session_to_draw = len(df_selected) \
             if 'selected from table or plot' in st.session_state.selected_draw_sessions \
             else len(st.session_state.df_session_filtered) 
 
@@ -243,7 +242,6 @@ def plot_x_y_session():
                                     if_use_x_quantile_all=if_use_x_quantile_all,
                                     q_quantiles_all=q_quantiles_all,
                                     title=names[(x_name, y_name)] if (x_name, y_name) in names else y_name,
-                                    states = st.session_state.df_selected_from_plotly,
                                     if_show_diagonal=if_show_diagonal,
                                     dot_size_base=dot_size,
                                     dot_size_mapping_name=size_mapper,
@@ -278,9 +276,9 @@ def plot_x_y_session():
         df_key_selected["session_date"] = pd.to_datetime(df_key_selected["session_date"])
         df_selected_from_plotly = df_x_y_session.merge(df_key_selected, on=["subject_id", "session_date"], how='inner')
         
-    if len(st.session_state.df_selected_from_plotly) == 1:
+    if len(df_selected_from_plotly) == 1:
         with cols[1]:
-            draw_session_plots_quick_preview(st.session_state.df_selected_from_plotly)
+            draw_session_plots_quick_preview(df_selected_from_plotly)
 
     return df_selected_from_plotly, cols
 
@@ -632,18 +630,16 @@ def add_main_tabs():
             # Add session_plot_setting
             with st.columns([1, 0.5])[0]:
                 st.markdown("***")
-                if_draw_all_sessions = session_plot_settings()
+                if_draw_all_sessions = session_plot_settings(df_selected_from_plotly)
 
-            df_to_draw_sessions = st.session_state.df_selected_from_plotly if 'selected' in st.session_state.selected_draw_sessions else st.session_state.df_session_filtered
+            df_to_draw_sessions = df_selected_from_plotly if 'selected' in st.session_state.selected_draw_sessions else st.session_state.df_session_filtered
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
                 draw_session_plots(df_to_draw_sessions)
-                
-            if len(df_selected_from_plotly) and not set(df_selected_from_plotly.set_index(['h2o', 'session']).index) == set(
-                                                st.session_state.df_selected_from_plotly.set_index(['h2o', 'session']).index):
-                st.session_state.df_selected_from_plotly = df_selected_from_plotly
-                st.session_state.df_selected_from_dataframe = df_selected_from_plotly  # Sync selected on dataframe
-                st.rerun()
+            
+            st.session_state.df_selected_from_plotly = df_selected_from_plotly
+            st.session_state.df_selected_from_dataframe = df_selected_from_plotly  # Sync selected on dataframe
+              
                 
     elif chosen_id == "tab_pygwalker":
         with placeholder:
@@ -680,8 +676,8 @@ def add_main_tabs():
         with placeholder:
             cols = st.columns([1, 0.5])
             with cols[0]:
-                if_draw_all_sessions = session_plot_settings(need_click=False)
                 df_to_draw_sessions = st.session_state.df_selected_from_plotly if 'selected' in st.session_state.selected_draw_sessions else st.session_state.df_session_filtered
+                if_draw_all_sessions = session_plot_settings(df_to_draw_sessions, need_click=False)
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
                 draw_session_plots(df_to_draw_sessions)
diff --git a/code/pages/1_Basic behavior analysis.py b/code/pages/1_Basic behavior analysis.py
index 1ecd2c8..d35cc59 100644
--- a/code/pages/1_Basic behavior analysis.py	
+++ b/code/pages/1_Basic behavior analysis.py	
@@ -8,7 +8,6 @@
 from plotly.subplots import make_subplots
 from sklearn.decomposition import PCA
 from sklearn.preprocessing import StandardScaler
-from streamlit_plotly_events import plotly_events
 from util.aws_s3 import load_data
 from util.streamlit import add_session_filter, data_selector, add_footnote
 from scipy.stats import gaussian_kde
diff --git a/code/util/streamlit.py b/code/util/streamlit.py
index dc1cdec..02715ee 100644
--- a/code/util/streamlit.py
+++ b/code/util/streamlit.py
@@ -713,15 +713,20 @@ def data_selector():
         #     st.session_state.df_selected_from_dataframe = pd.DataFrame()
         #     st.rerun()
 
-        cols = st.columns([5, 1, 1])
-        with cols[0].expander(f"Selected: {len(st.session_state.df_selected_from_plotly)} sessions, "
+        with st.expander(f"Selected: {len(st.session_state.df_selected_from_plotly)} sessions, "
                               f"{len(st.session_state.df_selected_from_plotly.h2o.unique())} mice", expanded=False):
             st.dataframe(st.session_state.df_selected_from_plotly)
-        if cols[1].button('all'):
+        cols = st.columns([1, 1, 1])
+            
+        if cols[0].button('sync'):
+            # simply rerun to update session states that are changed in fragment
+            st.rerun()
+            
+        if cols[1].button('select all'):
             st.session_state.df_selected_from_plotly = st.session_state.df_session_filtered
             st.rerun()
         
-        if cols[2].button('❌ '):
+        if cols[2].button('clear all'):
             st.session_state.df_selected_from_plotly = pd.DataFrame(columns=['h2o', 'session'])
             st.session_state.df_selected_from_dataframe = pd.DataFrame(columns=['h2o', 'session'])
             st.rerun()
@@ -1063,14 +1068,14 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
         col = col_map[i%len(col_map)]
 
         if if_show_dots:
-            if not len(st.session_state.df_selected_from_plotly):   
-                this_session['colors'] = col  # all use normal colors
-            else:
-                merged = pd.merge(this_session, st.session_state.df_selected_from_plotly, on=['h2o', 'session'], how='left')
-                merged['colors'] = 'lightgrey'  # default, grey
-                merged.loc[merged.subject_id_y.notna(), 'colors'] = col   # only use normal colors for the selected dots 
-                this_session['colors'] = merged.colors.values
-                this_session = pd.concat([this_session.query('colors != "lightgrey"'), this_session.query('colors == "lightgrey"')])  # make sure the real color goes first
+            # if not len(st.session_state.df_selected_from_plotly):   
+            this_session['colors'] = col  # all use normal colors
+            # else:
+            #     merged = pd.merge(this_session, st.session_state.df_selected_from_plotly, on=['h2o', 'session'], how='left')
+            #     merged['colors'] = 'lightgrey'  # default, grey
+            #     merged.loc[merged.subject_id_y.notna(), 'colors'] = col   # only use normal colors for the selected dots 
+            #     this_session['colors'] = merged.colors.values
+            #     this_session = pd.concat([this_session.query('colors != "lightgrey"'), this_session.query('colors == "lightgrey"')])  # make sure the real color goes first
 
             fig.add_trace(go.Scattergl(
                             x=this_session[x_name], 
@@ -1081,7 +1086,7 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
                             mode="markers",
                             line_width=line_width,
                             marker_size=this_session['dot_size'],
-                            marker_color=this_session['colors'],
+                            # marker_color=this_session['colors'],
                             opacity=dot_opacity,
                             hovertemplate =  '<b>%{customdata[0]}, %{customdata[1]}, Session %{customdata[2]}'
                                              '<br>%{customdata[4]} @ %{customdata[9]}'

From 5dca7e20867516e706facf0d61082f8588a6b0f9 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Sat, 11 Jan 2025 00:39:13 +0000
Subject: [PATCH 14/19] use orjson to improve plotly performance

---
 code/pages/0_Data inventory.py      | 3 +++
 code/util/foraging_plotly.py        | 2 ++
 code/util/plot_autotrain_manager.py | 3 +++
 code/util/settings.py               | 1 +
 requirements.txt                    | 1 +
 5 files changed, 10 insertions(+)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index 71c5722..99ce188 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -10,6 +10,9 @@
 import numpy as np
 import plotly.graph_objects as go
 from plotly.subplots import make_subplots
+import plotly.io as pio
+pio.json.config.default_engine = "orjson"
+
 from streamlit_plotly_events import plotly_events
 
 import time
diff --git a/code/util/foraging_plotly.py b/code/util/foraging_plotly.py
index 4a8119f..b2f7b67 100644
--- a/code/util/foraging_plotly.py
+++ b/code/util/foraging_plotly.py
@@ -1,6 +1,8 @@
 import numpy as np
 import plotly.express as px
 import plotly.graph_objs as go
+import plotly.io as pio
+pio.json.config.default_engine = "orjson"
 
 
 def moving_average(a, n=3) :
diff --git a/code/util/plot_autotrain_manager.py b/code/util/plot_autotrain_manager.py
index f5e0209..3d3d28d 100644
--- a/code/util/plot_autotrain_manager.py
+++ b/code/util/plot_autotrain_manager.py
@@ -3,6 +3,9 @@
 import numpy as np
 import pandas as pd
 import plotly.graph_objects as go
+import plotly.io as pio
+pio.json.config.default_engine = "orjson"
+
 import streamlit as st
 from aind_auto_train.plot.curriculum import get_stage_color_mapper
 from aind_auto_train.schema.curriculum import TrainingStage
diff --git a/code/util/settings.py b/code/util/settings.py
index 10cca02..7fcc985 100644
--- a/code/util/settings.py
+++ b/code/util/settings.py
@@ -1,4 +1,5 @@
 import plotly.io as pio
+pio.json.config.default_engine = "orjson"
 
 # Setting up layout for each session
 draw_type_layout_definition = [
diff --git a/requirements.txt b/requirements.txt
index a617adb..5bafda3 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -21,6 +21,7 @@ pillow==10.4.0
 tabulate==0.9.0
 aind-data-access-api[docdb]==0.13.0
 matplotlib_venn==1.1.1
+orjson==3.10.14
 git+https://github.com/AllenNeuralDynamics/aind-foraging-behavior-bonsai-automatic-training.git@main
 git+https://github.com/AllenNeuralDynamics/aind-dynamic-foraging-models.git@develop
 git+https://github.com/AllenNeuralDynamics/aind-behavior-gym.git@develop
\ No newline at end of file

From 7cc12d432e075c0ac18cef593a49afd8fc92b6cb Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Sat, 11 Jan 2025 02:45:21 +0000
Subject: [PATCH 15/19] fix selection logic!! finallllly...

---
 code/Home.py           | 44 +++++++++++++++++------------------------
 code/util/streamlit.py | 45 +++++++++++++++++++++++++-----------------
 2 files changed, 45 insertions(+), 44 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index a38ff4f..5fe8ad1 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -211,7 +211,6 @@ def plot_x_y_session():
             submitted = st.form_submit_button("👉 Update X-Y settings 👈", type='primary')
 
     # If no sessions are selected, use all filtered entries
-    # df_x_y_session = st.session_state.df_selected_from_dataframe if if_plot_only_selected_from_dataframe else st.session_state.df_session_filtered
     df_x_y_session = st.session_state.df_session_filtered
 
     names = {('session', 'foraging_eff'): 'Foraging efficiency',
@@ -256,19 +255,18 @@ def plot_x_y_session():
                                     )
 
         selected = st.plotly_chart(fig, 
+                                   key='x_y_plot',
                                    on_select="rerun",
                                    use_container_width=True,
                                    theme=None,  # full controlled by plotly chart itself
                         )
-        # selected = plotly_events(fig, click_event=True, hover_event=False, select_event=True,
-        #                          override_height=fig.layout.height * 1.1, override_width=fig.layout.width)
 
     with cols[1]:
         st.markdown('#### 👀 Quick preview')
         st.markdown('###### Click on one session to preview here, or Box/Lasso select multiple sessions to draw them in the section below')
         st.markdown('(sometimes you have to click twice...)')
 
-    if len(selected):
+    if len(selected.selection.points):  # Selected this time
         df_key_selected = pd.DataFrame(
             [data["customdata"][:2] for data in selected.selection.points],
             columns=["subject_id", "session_date"],
@@ -276,6 +274,9 @@ def plot_x_y_session():
         df_key_selected["session_date"] = pd.to_datetime(df_key_selected["session_date"])
         df_selected_from_plotly = df_x_y_session.merge(df_key_selected, on=["subject_id", "session_date"], how='inner')
         
+        # Update session state
+        st.session_state.df_selected_from_plotly = df_selected_from_plotly
+        
     if len(df_selected_from_plotly) == 1:
         with cols[1]:
             draw_session_plots_quick_preview(df_selected_from_plotly)
@@ -319,8 +320,8 @@ def init(if_load_bpod_data_override=None, if_load_docDB_override=None):
         df['sessions_bonsai'] = pd.concat([df['sessions_bonsai'], df_bpod['sessions_bonsai']], axis=0)
         
     st.session_state.df = df
-    st.session_state.df_selected_from_plotly = pd.DataFrame(columns=['h2o', 'session'])
-    st.session_state.df_selected_from_dataframe = pd.DataFrame(columns=['h2o', 'session'])
+    for source in ["dataframe", "plotly"]:
+        st.session_state[f'df_selected_from_{source}'] = pd.DataFrame(columns=['h2o', 'session'])
             
     # Init auto training database
     st.session_state.curriculum_manager = CurriculumManager(
@@ -537,11 +538,6 @@ def app():
         data_selector()
         add_footnote()
         
-        with st.expander('Debug', expanded=False):
-            if st.button('Clear session state and reload data'):
-                st.cache_data.clear()
-                init()
-                st.rerun()
         
     with st.container():
         # col1, col2 = st.columns([1.5, 1], gap='small')
@@ -582,9 +578,6 @@ def app():
                                                     step=50,
                                                     key='table_height',
         )
-            
-        # aggrid_outputs = aggrid_interactive_table_units(df=df['ephys_units'])
-        # st.session_state.df_session_filtered = aggrid_outputs['data']
         
         container_filtered_frame = st.container()
 
@@ -598,13 +591,14 @@ def app():
         table_height=table_height,
     )
 
-    if len(aggrid_outputs['selected_rows']) and not set(pd.DataFrame(aggrid_outputs['selected_rows']
-                                                                 ).set_index(['h2o', 'session']).index
-                                                        ) == set(st.session_state.df_selected_from_dataframe.set_index(['h2o', 'session']).index):
-        st.session_state.df_selected_from_dataframe = pd.DataFrame(aggrid_outputs['selected_rows'])
-        st.session_state.df_selected_from_plotly = st.session_state.df_selected_from_dataframe  # Sync selected on plotly
-        # if st.session_state.tab_id == "tab_session_x_y":
+    if len(aggrid_outputs['selected_rows']) \
+        and not set(pd.DataFrame(aggrid_outputs['selected_rows']).set_index(['h2o', 'session']).index
+            ) == set(st.session_state.df_selected_from_dataframe.set_index(['h2o', 'session']).index) \
+        and not st.session_state.get("df_selected_from_dataframe_just_overriden", False):  # so that if the user just overriden the df_selected_from_dataframe by pressing sidebar button, it won't sync selected rows in the table to session state
+        st.session_state.df_selected_from_dataframe = pd.DataFrame(aggrid_outputs['selected_rows'])  # Use selected in dataframe to update "selected"
         st.rerun()
+        
+    st.session_state["df_selected_from_dataframe_just_overriden"] = False  # Reset the flag anyway
 
     add_main_tabs()
 
@@ -612,8 +606,8 @@ def app():
 def add_main_tabs():
     chosen_id = stx.tab_bar(data=[
         stx.TabBarItemData(id="tab_auto_train_history", title="🎓 Automatic Training History", description="Track progress"),
-        stx.TabBarItemData(id="tab_session_inspector", title="👀 Session Inspector", description="Select sessions from the table and show plots"),
-        stx.TabBarItemData(id="tab_session_x_y", title="📈 Session X-Y plot", description="Interactive session-wise scatter plot"),
+        stx.TabBarItemData(id="tab_session_inspector", title="👀 Session Inspector (table)", description="Select sessions from the table and show figures"),
+        stx.TabBarItemData(id="tab_session_x_y", title="📈 Session X-Y plot", description="Select sessions from x-y plot and show figures"),
         stx.TabBarItemData(id="tab_pygwalker", title="📊 PyGWalker (Tableau)", description="Interactive dataframe explorer"),
         stx.TabBarItemData(id="tab_auto_train_curriculum", title="📚 Automatic Training Curriculums", description="Collection of curriculums"),
         # stx.TabBarItemData(id="tab_mouse_inspector", title="🐭 Mouse Inspector", description="Mouse-level summary"),
@@ -636,9 +630,6 @@ def add_main_tabs():
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
                 draw_session_plots(df_to_draw_sessions)
-            
-            st.session_state.df_selected_from_plotly = df_selected_from_plotly
-            st.session_state.df_selected_from_dataframe = df_selected_from_plotly  # Sync selected on dataframe
               
                 
     elif chosen_id == "tab_pygwalker":
@@ -676,7 +667,7 @@ def add_main_tabs():
         with placeholder:
             cols = st.columns([1, 0.5])
             with cols[0]:
-                df_to_draw_sessions = st.session_state.df_selected_from_plotly if 'selected' in st.session_state.selected_draw_sessions else st.session_state.df_session_filtered
+                df_to_draw_sessions = st.session_state.df_selected_from_dataframe if 'selected' in st.session_state.get("selected_draw_sessions", "") else st.session_state.df_session_filtered
                 if_draw_all_sessions = session_plot_settings(df_to_draw_sessions, need_click=False)
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
@@ -797,3 +788,4 @@ def add_main_tabs():
 
     if ok:
         app()
+        pass
diff --git a/code/util/streamlit.py b/code/util/streamlit.py
index 02715ee..30e8991 100644
--- a/code/util/streamlit.py
+++ b/code/util/streamlit.py
@@ -10,6 +10,8 @@
 import plotly
 import plotly.express as px
 import plotly.graph_objects as go
+import plotly.io as pio
+pio.json.config.default_engine = "orjson"
 import statsmodels.api as sm
 import streamlit as st
 import streamlit.components.v1 as components
@@ -68,8 +70,8 @@ def aggrid_interactive_table_session(df: pd.DataFrame, table_height: int = 400):
         df = df.sort_values('session_date', ascending=False)
     
     # preselect
-    if (('df_selected_from_dataframe' in st.session_state and len(st.session_state.df_selected_from_dataframe)) 
-       and ('tab_id' in st.session_state and st.session_state.tab_id == "tab_session_x_y")):
+    if len(st.session_state.get("df_selected_from_dataframe", [])) \
+       and ('tab_id' in st.session_state) and (st.session_state.tab_id == "tab_session_x_y"):
         try:
             indexer = st.session_state.df_selected_from_dataframe.set_index(['h2o', 'session']
                                                                 ).index.get_indexer(df.set_index(['h2o', 'session']).index)
@@ -693,7 +695,7 @@ def add_dot_property_mapper():
         
     return size_mapper, size_mapper_range, size_mapper_gamma
 
-
+@st.fragment(run_every=5)
 def data_selector():
             
     with st.expander(f'Session selector', expanded=True):        
@@ -713,23 +715,30 @@ def data_selector():
         #     st.session_state.df_selected_from_dataframe = pd.DataFrame()
         #     st.rerun()
 
-        with st.expander(f"Selected: {len(st.session_state.df_selected_from_plotly)} sessions, "
-                              f"{len(st.session_state.df_selected_from_plotly.h2o.unique())} mice", expanded=False):
-            st.dataframe(st.session_state.df_selected_from_plotly)
-        cols = st.columns([1, 1, 1])
+        # Separate selection from table or streamlit
+        def _show_selected(source="dataframe"):
+            df_this = st.session_state['df_selected_from_' + source]
+            with st.expander(f"Selected from {source}: {len(df_this)} sessions, "
+                                f"{len(df_this.h2o.unique())} mice", expanded=False):
+                st.dataframe(df_this)
+            cols = st.columns([1, 1, 1])
+            
+            if source == "plotly": 
+                return  # Don't allow select all or clear all for plotly
             
-        if cols[0].button('sync'):
-            # simply rerun to update session states that are changed in fragment
-            st.rerun()
+            if cols[1].button('select all', key=f'select_all_from_{source}'):
+                st.session_state['df_selected_from_' + source] = st.session_state.df_session_filtered
+                st.session_state['df_selected_from_' + source + '_just_overriden'] = True
+                st.rerun()
             
-        if cols[1].button('select all'):
-            st.session_state.df_selected_from_plotly = st.session_state.df_session_filtered
-            st.rerun()
+            if cols[2].button('clear all', key=f'clear_all_from_{source}'):
+                st.session_state['df_selected_from_' + source] = pd.DataFrame(columns=['h2o', 'session'])
+                st.session_state['df_selected_from_' + source + '_just_overriden'] = True
+                st.rerun()
+
+        for source in ['dataframe', 'plotly']:
+            _show_selected(source)
         
-        if cols[2].button('clear all'):
-            st.session_state.df_selected_from_plotly = pd.DataFrame(columns=['h2o', 'session'])
-            st.session_state.df_selected_from_dataframe = pd.DataFrame(columns=['h2o', 'session'])
-            st.rerun()
 
 def _add_download_filtered_session():
     """Download the master table of the filtered session"""
@@ -1086,7 +1095,7 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
                             mode="markers",
                             line_width=line_width,
                             marker_size=this_session['dot_size'],
-                            # marker_color=this_session['colors'],
+                            marker_color=this_session['colors'],
                             opacity=dot_opacity,
                             hovertemplate =  '<b>%{customdata[0]}, %{customdata[1]}, Session %{customdata[2]}'
                                              '<br>%{customdata[4]} @ %{customdata[9]}'

From 018a7da2c19b429c3659cecca62aa1075d5fb906 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Mon, 13 Jan 2025 22:51:03 +0000
Subject: [PATCH 16/19] fix a bug where selected_draw_sessions is not properly
 initiated

---
 code/Home.py | 55 +++++++++++++++++++++++++++++-----------------------
 1 file changed, 31 insertions(+), 24 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index 5fe8ad1..5c0777a 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -620,18 +620,21 @@ def add_main_tabs():
     if chosen_id == "tab_session_x_y":
         with placeholder:
             df_selected_from_plotly, x_y_cols = plot_x_y_session()
-            
+
             # Add session_plot_setting
             with st.columns([1, 0.5])[0]:
                 st.markdown("***")
                 if_draw_all_sessions = session_plot_settings(df_selected_from_plotly)
 
-            df_to_draw_sessions = df_selected_from_plotly if 'selected' in st.session_state.selected_draw_sessions else st.session_state.df_session_filtered
+            df_to_draw_sessions = (
+                df_selected_from_plotly
+                if "selected" in st.session_state.get("selected_draw_sessions", "sessions selected from table or plot")
+                else st.session_state.df_session_filtered
+            )
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
                 draw_session_plots(df_to_draw_sessions)
-              
-                
+
     elif chosen_id == "tab_pygwalker":
         with placeholder:
             cols = st.columns([1, 4])
@@ -642,7 +645,7 @@ def add_main_tabs():
                     pyg_user_json = st.text_area("Export your plot settings to json by clicking `export_code` "
                                                  "button below and then paste your json here to reproduce your plots", 
                                                 key='pyg_walker', height=100)
-            
+
             # If pyg_user_json is not empty, use it; otherwise, use the default gw_config.json
             if pyg_user_json:
                 try:
@@ -660,24 +663,30 @@ def add_main_tabs():
                     df=st.session_state.df_session_filtered,
                     spec="./gw_config.json",
                     )
-                            
+
             pygwalker_renderer.render_explore()
-        
+
     elif chosen_id == "tab_session_inspector":
         with placeholder:
             cols = st.columns([1, 0.5])
             with cols[0]:
-                df_to_draw_sessions = st.session_state.df_selected_from_dataframe if 'selected' in st.session_state.get("selected_draw_sessions", "") else st.session_state.df_session_filtered
-                if_draw_all_sessions = session_plot_settings(df_to_draw_sessions, need_click=False)
+                df_to_draw_sessions = (
+                    st.session_state.df_selected_from_dataframe
+                    if "selected" in st.session_state.get("selected_draw_sessions", "sessions selected from table or plot")
+                    else st.session_state.df_session_filtered
+                )
+                if_draw_all_sessions = session_plot_settings(
+                    df_to_draw_sessions, need_click=False
+                )
 
             if if_draw_all_sessions and len(df_to_draw_sessions):
                 draw_session_plots(df_to_draw_sessions)
-                
+
     elif chosen_id == "tab_mouse_inspector":
         with placeholder:
             selected_subject_id = st.columns([1, 3])[0].selectbox('Select a mouse', options=st.session_state.df_session_filtered['subject_id'].unique())
             st.markdown(f"### [Go to WaterLog](http://eng-tools:8004/water_weight_log/?external_donor_name={selected_subject_id})")
-            
+
     elif chosen_id == "tab_auto_train_history":  # Automatic training history
         with placeholder:
             add_auto_train_manager()
@@ -687,7 +696,7 @@ def add_main_tabs():
             by=['curriculum_version', 'curriculum_schema_version', 'curriculum_name'],
             ascending=[False, True, False], 
             ).reset_index().drop(columns='index').query("curriculum_name != 'Dummy task'")
-        
+
         with placeholder:
             # Show curriculum manager dataframe
             st.markdown("#### Select auto training curriculums")
@@ -695,7 +704,7 @@ def add_main_tabs():
             # Curriculum drop down selector
             cols = st.columns([0.8, 0.5, 0.8, 4])
             cols[3].markdown(f"(aind_auto_train lib version = {auto_train_version})")
-            
+
             options = list(df_curriculums['curriculum_name'].unique())
             selected_curriculum_name = selectbox_wrapper_for_url_query(
                 st_prefix=cols[0],
@@ -705,7 +714,7 @@ def add_main_tabs():
                 default_override=True,
                 key='auto_training_curriculum_name',
             )
-                       
+
             options = list(df_curriculums[
                 df_curriculums['curriculum_name'] == selected_curriculum_name
                 ]['curriculum_version'].unique())
@@ -717,12 +726,12 @@ def add_main_tabs():
                 default_override=True,
                 key='auto_training_curriculum_version',
             )
-            
+
             options = list(df_curriculums[
                 (df_curriculums['curriculum_name'] == selected_curriculum_name) 
                 & (df_curriculums['curriculum_version'] == selected_curriculum_version)
                 ]['curriculum_schema_version'].unique())
-            
+
             selected_curriculum_schema_version = selectbox_wrapper_for_url_query(
                 st_prefix=cols[2],
                 label='Curriculum schema version',
@@ -731,31 +740,30 @@ def add_main_tabs():
                 default_override=True,
                 key='auto_training_curriculum_schema_version',
             )
-                                   
+
             selected_curriculum = st.session_state.curriculum_manager.get_curriculum(
                 curriculum_name=selected_curriculum_name,
                 curriculum_schema_version=selected_curriculum_schema_version,
                 curriculum_version=selected_curriculum_version,
                 )
-            
+
             # Get selected curriculum from previous selected or the URL
             if 'auto_training_curriculum_name' in st.session_state:
                 selected_row = {'curriculum_name': st.session_state['auto_training_curriculum_name'],
                                 'curriculum_schema_version': st.session_state['auto_training_curriculum_schema_version'],
                                 'curriculum_version': st.session_state['auto_training_curriculum_version']}
                 matched_curriculum = df_curriculums[(df_curriculums[list(selected_row)] == pd.Series(selected_row)).all(axis=1)]
-                
+
                 if len(matched_curriculum):
                     pre_selected_rows = matched_curriculum.index.to_list() 
                 else:
                     selected_row = None # Clear selected row if not found
                     pre_selected_rows = None
-            
-            # Show df_curriculum       
+
+            # Show df_curriculum
             aggrid_interactive_table_basic(df=df_curriculums,
                                                 pre_selected_rows=pre_selected_rows)        
 
-            
             if selected_curriculum is not None:
                 curriculum = selected_curriculum['curriculum']
                 # Show diagrams
@@ -775,10 +783,9 @@ def add_main_tabs():
         st.markdown('---\n##### Debug zone')
         show_debug_info()
 
-    
     # Update back to URL
     sync_session_state_to_URL()
-    
+
     # st.dataframe(st.session_state.df_session_filtered, use_container_width=True, height=1000)
 
 if __name__ == "__main__":

From 06ed37c813c7dd4ff03ab97891aed21ab5ddcf0a Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Tue, 14 Jan 2025 18:55:47 +0000
Subject: [PATCH 17/19] further improve plotly styling

---
 code/Home.py                   |  6 +--
 code/pages/0_Data inventory.py | 19 +++-----
 code/util/settings.py          | 82 +++++++++++++++++++++++++++-------
 code/util/streamlit.py         | 14 ++----
 4 files changed, 79 insertions(+), 42 deletions(-)

diff --git a/code/Home.py b/code/Home.py
index 5c0777a..7dfd715 100644
--- a/code/Home.py
+++ b/code/Home.py
@@ -257,7 +257,7 @@ def plot_x_y_session():
         selected = st.plotly_chart(fig, 
                                    key='x_y_plot',
                                    on_select="rerun",
-                                   use_container_width=True,
+                                   use_container_width=False,
                                    theme=None,  # full controlled by plotly chart itself
                         )
 
@@ -622,7 +622,7 @@ def add_main_tabs():
             df_selected_from_plotly, x_y_cols = plot_x_y_session()
 
             # Add session_plot_setting
-            with st.columns([1, 0.5])[0]:
+            with st.columns([1])[0]:
                 st.markdown("***")
                 if_draw_all_sessions = session_plot_settings(df_selected_from_plotly)
 
@@ -668,7 +668,7 @@ def add_main_tabs():
 
     elif chosen_id == "tab_session_inspector":
         with placeholder:
-            cols = st.columns([1, 0.5])
+            cols = st.columns([1])
             with cols[0]:
                 df_to_draw_sessions = (
                     st.session_state.df_selected_from_dataframe
diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index 99ce188..add9a21 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -13,8 +13,6 @@
 import plotly.io as pio
 pio.json.config.default_engine = "orjson"
 
-from streamlit_plotly_events import plotly_events
-
 import time
 import streamlit_nested_layout
 
@@ -25,6 +23,7 @@
 )
 from util.reformat import formatting_metadata_df
 from util.aws_s3 import load_raw_sessions_on_VAST
+from util.settings import override_plotly_theme
 from Home import init
 
 
@@ -254,7 +253,7 @@ def count_true_values(df, time_period, column):
             rows=len(columns),
             cols=1,
             shared_xaxes=True,
-            vertical_spacing=0.05,
+            vertical_spacing=0.1,
             subplot_titles=columns,
         )
 
@@ -280,7 +279,7 @@ def count_true_values(df, time_period, column):
 
         # Updating layout
         fig.update_layout(
-            height=200 * len(columns),
+            height=250 * len(columns),
             showlegend=False,
             title=f"{time_period} counts",
         )
@@ -434,7 +433,7 @@ def app():
     # --- Venn diagram from presets ---
     if VENN_PRESET:
         add_venn_diagrms(df_merged)
-        
+
 @st.fragment
 def add_venn_diagrms(df_merged):
 
@@ -504,14 +503,8 @@ def add_venn_diagrms(df_merged):
                 if_sync_y_limits=if_sync_y_limits,
                 if_separate_plots=if_separate_plots,
             )
-            plotly_events(
-                fig,
-                click_event=False,
-                hover_event=False,
-                select_event=False,
-                override_height=fig.layout.height * 1.1,
-                override_width=fig.layout.width,
-            )
+            override_plotly_theme(fig, font_size_scale=1.0)
+            st.plotly_chart(fig, use_container_width=True)
 
         st.markdown("---")
 
diff --git a/code/util/settings.py b/code/util/settings.py
index 7fcc985..5b1bd98 100644
--- a/code/util/settings.py
+++ b/code/util/settings.py
@@ -33,28 +33,80 @@
 }
 
 # For quick preview
-draw_types_quick_preview = [
-    '1. Choice history',
-    '3. Logistic regression (Su2022)']
+draw_types_quick_preview = ["1. Choice history", "3. Logistic regression (Su2022)"]
 
 
-def override_plotly_theme(fig, theme):
+# For plotly styling
+PLOTLY_FIG_DEFAULT = dict(
+        font_family="Arial",
+    )
+PLOTLY_AXIS_DEFAULT = dict(
+        showline=True,
+        linewidth=2,
+        linecolor="black",
+        showgrid=True,
+        gridcolor="lightgray",
+        griddash="solid",
+        minor_showgrid=False,
+        minor_gridcolor="lightgray",
+        minor_griddash="solid",
+        zeroline=True,
+        ticks="outside",
+        tickcolor="black",
+        ticklen=7,
+        tickwidth=2,
+        ticksuffix=" ",
+        tickfont=dict(
+            family="Arial",
+            color="black",
+        ),
+    )
+
+def override_plotly_theme(
+    fig,
+    theme="simple_white",
+    fig_specs=PLOTLY_FIG_DEFAULT,
+    axis_specs=PLOTLY_AXIS_DEFAULT,
+    font_size_scale=1.0,
+):
     """
     Fix the problem that simply using fig.update_layout(template=theme) doesn't work with st.plotly_chart.
     I have to use update_layout to explicitly set the theme.
     """
-    
+
     dict_plotly_template = pio.templates[theme].layout.to_plotly_json()
     fig.update_layout(**dict_plotly_template)  # First apply the plotly official theme
-    
-    fig.update_layout(font_family="Arial")  # Add user-defined styling
-    
-    fig.update_xaxes(showline=True, linewidth=2, linecolor='black', showgrid=True, zeroline=True,
-                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ')
 
-    fig.update_yaxes(showline=True, linewidth=2, linecolor='black', showgrid=True, zeroline=True,
-                     ticks = "outside", tickcolor='black', ticklen=10, tickwidth=2, ticksuffix=' ',
-                     title_standoff=40,
-                     )
+    # Apply settings to all x-axes
+    for axis in fig.layout:
+        if axis.startswith('xaxis') or axis.startswith('yaxis'):
+            fig.layout[axis].update(axis_specs)
+            fig.layout[axis].update(
+                tickfont_size=20 * font_size_scale, 
+                title_font_size=22 * font_size_scale,
+            )
+        if axis.startswith("yaxis"):
+            fig.layout[axis].update(title_standoff=10 * font_size_scale)
+
+    fig.update_layout(**fig_specs)  # Apply settings to the entire figure
+    
+    # Customize the font of subplot titles
+    for annotation in fig['layout']['annotations']:
+        annotation['font'] = dict(
+            family="Arial",  # Font family
+            size=20 * font_size_scale,    # Font size
+            color="black"     # Font color
+        )
 
-    return
\ No newline at end of file
+    fig.update_layout(
+        font_size=22 * font_size_scale,
+        hoverlabel_font_size=17 * font_size_scale,
+        legend_font_size=20 * font_size_scale,
+        margin=dict(
+            l=130 * font_size_scale,
+            r=50 * font_size_scale,
+            b=130 * font_size_scale,
+            t=100 * font_size_scale,
+        ),
+    )
+    return
diff --git a/code/util/streamlit.py b/code/util/streamlit.py
index 30e8991..9189167 100644
--- a/code/util/streamlit.py
+++ b/code/util/streamlit.py
@@ -1140,25 +1140,17 @@ def _add_agg(df_this, x_name, y_name, group, aggr_method, if_use_x_quantile, q_q
     n_mice = len(df['h2o'].unique())
     n_sessions = len(df.groupby(['h2o', 'session']).count())
 
-    override_plotly_theme(fig, theme="simple_white")
+    override_plotly_theme(fig, theme="simple_white", font_size_scale=font_size_scale)
 
     fig.update_layout(
         width=x_y_plot_figure_width,
         height=x_y_plot_figure_height,
-        xaxis_title=x_name,
-        yaxis_title=y_name,
-        font=dict(size=24 * font_size_scale),
+        xaxis_title_text=x_name,
+        yaxis_title_text=y_name,
         hovermode="closest",
-        hoverlabel=dict(font_size=17 * font_size_scale),
         legend={"traceorder": "reversed"},
-        legend_font_size=20 * font_size_scale,
         title=f"{title}, {n_mice} mice, {n_sessions} sessions",
         dragmode="zoom",  # 'select',
-        margin=dict(l=130 * font_size_scale, 
-                    r=50 * font_size_scale, 
-                    b=130 * font_size_scale, 
-                    t=100 * font_size_scale,
-                    ),
     ) 
     return fig
 

From 7c0654a4c032bb3b9c27aac96b9b5697c6b5aee7 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Tue, 14 Jan 2025 19:01:56 +0000
Subject: [PATCH 18/19] minor

---
 code/pages/0_Data inventory.py | 2 +-
 code/util/settings.py          | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index add9a21..d18d4a1 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -503,7 +503,7 @@ def add_venn_diagrms(df_merged):
                 if_sync_y_limits=if_sync_y_limits,
                 if_separate_plots=if_separate_plots,
             )
-            override_plotly_theme(fig, font_size_scale=1.0)
+            override_plotly_theme(fig, font_size_scale=0.9)
             st.plotly_chart(fig, use_container_width=True)
 
         st.markdown("---")
diff --git a/code/util/settings.py b/code/util/settings.py
index 5b1bd98..cb27d7a 100644
--- a/code/util/settings.py
+++ b/code/util/settings.py
@@ -82,7 +82,7 @@ def override_plotly_theme(
         if axis.startswith('xaxis') or axis.startswith('yaxis'):
             fig.layout[axis].update(axis_specs)
             fig.layout[axis].update(
-                tickfont_size=20 * font_size_scale, 
+                tickfont_size=22 * font_size_scale, 
                 title_font_size=22 * font_size_scale,
             )
         if axis.startswith("yaxis"):
@@ -101,7 +101,7 @@ def override_plotly_theme(
     fig.update_layout(
         font_size=22 * font_size_scale,
         hoverlabel_font_size=17 * font_size_scale,
-        legend_font_size=20 * font_size_scale,
+        legend_font_size=17 * font_size_scale,
         margin=dict(
             l=130 * font_size_scale,
             r=50 * font_size_scale,

From 24ca965e11f1e7af13c3d91e4c40f455eecf2df4 Mon Sep 17 00:00:00 2001
From: "houhan@gmail.com" <han.hou@alleninstitute.org>
Date: Tue, 14 Jan 2025 19:10:17 +0000
Subject: [PATCH 19/19] fix legend color

---
 code/pages/0_Data inventory.py | 8 ++++----
 code/util/settings.py          | 2 ++
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/code/pages/0_Data inventory.py b/code/pages/0_Data inventory.py
index d18d4a1..7e6c660 100644
--- a/code/pages/0_Data inventory.py	
+++ b/code/pages/0_Data inventory.py	
@@ -299,18 +299,18 @@ def count_true_values(df, time_period, column):
             height=500,
             bargap=0.05,  # Gap between bars of adjacent locations
             bargroupgap=0.1,  # Gap between bars of the same location
-            barmode='group',  # Grouped style
+            barmode="group",  # Grouped style
             showlegend=True,
+            title="Monthly counts",
             legend=dict(
                 orientation="h",  # Horizontal legend
                 y=-0.2,  # Position below the plot
                 x=0.5,  # Center the legend
                 xanchor="center",  # Anchor the legend's x position
-                yanchor="top"  # Anchor the legend's y position
+                yanchor="top",  # Anchor the legend's y position
             ),
-            title="Monthly counts"
         )
-        
+
     return fig
 
 def app():
diff --git a/code/util/settings.py b/code/util/settings.py
index cb27d7a..65d157a 100644
--- a/code/util/settings.py
+++ b/code/util/settings.py
@@ -39,6 +39,7 @@
 # For plotly styling
 PLOTLY_FIG_DEFAULT = dict(
         font_family="Arial",
+        legend_font_color='black',
     )
 PLOTLY_AXIS_DEFAULT = dict(
         showline=True,
@@ -98,6 +99,7 @@ def override_plotly_theme(
             color="black"     # Font color
         )
 
+    # Figure-level settings
     fig.update_layout(
         font_size=22 * font_size_scale,
         hoverlabel_font_size=17 * font_size_scale,