Update app.py

app.py

@@ -8,6 +8,7 @@ st.set_page_config(layout="wide")
 DATA_FILE = "data/aclanthology2016-23_specter2_base.json"
 THEMES = {"cluster": "fall", "year": "mint", "source": "phase"}
 
+
 def load_df(data_file: os.PathLike):
     df = pd.read_json(data_file, orient="records")
     df["x"] = df["point2d"].apply(lambda x: x[0])
@@ -26,6 +27,8 @@ def load_dataframe():
 
 
 DF = load_dataframe()
+DF["opacity"] = 0.04
+min_year, max_year = DF["year"].min(), DF["year"].max()
 
 with st.sidebar:
     venues = st.multiselect(
@@ -35,7 +38,9 @@ with st.sidebar:
     )
 
     start_year, end_year = st.select_slider(
-        "Publication year", options=("2020", "2021", "2022", "2023"), value=("2020", "2023")
+        "Publication year",
+        options=[str(y) for y in range(min_year, max_year + 1)],
+        value=(str(min_year), str(max_year)),
     )
     author_names = st.text_input("Author names (separated by comma)")
 
@@ -43,32 +48,34 @@ with st.sidebar:
 
     start_year = int(start_year)
     end_year = int(end_year)
-    df = DF[(DF["year"] >= start_year) & (DF["year"] <= end_year)]
+    df_mask = (DF["year"] >= start_year) & (DF["year"] <= end_year)
     if 0 < len(venues) < 4:
         selected_venues = [v.lower() for v in venues]
-        df = df[df["source"].isin(selected_venues)]
+        df_mask = df_mask & DF["source"].isin(selected_venues)
     elif not venues:
         st.write(":red[Please select a venue]")
 
     if author_names:
         authors = [a.strip() for a in author_names.split(",")]
-        author_mask = df.authors.apply(
+        author_mask = DF.authors.apply(
             lambda row: all(any(re.match(rf".*{a}.*", x, re.IGNORECASE) for x in row) for a in authors)
         )
-        df = df[author_mask]
+        df_mask = df_mask & author_mask
 
     if title:
-        df = df[df.title.apply(lambda x: title.lower() in x.lower())]
+        df_mask = df_mask & DF.title.apply(lambda x: title.lower() in x.lower())
 
-    st.write(f"Number of points: {df.shape[0]}")
+    DF.loc[df_mask, "opacity"] = 1.0
+    st.write(f"Number of points: {DF[df_mask].shape[0]}")
 
     color = st.selectbox("Color", ("cluster", "year", "source"))
 
 
 fig = px.scatter(
-    df,
+    DF,
     x="x",
     y="y",
+    opacity=DF["opacity"],
     color=color,
     width=1000,
     height=800,
@@ -86,4 +93,4 @@ fig.update_layout(
 fig.update_xaxes(title="")
 fig.update_yaxes(title="")
 
-st.plotly_chart(fig, use_container_width=True)
+st.plotly_chart(fig, use_container_width=True)