fix: Adding a copy of cleaned text for internal representation and keeping original text for visualization

app.py

@@ -115,11 +115,12 @@ def draw_interactive_scatter_plot(
 # Up to here
 def generate_plot(
     tws: List[str],
+    tws_cleaned: List[str],
     model: SentenceTransformer,
     tw_user: str
 ) -> Figure:
     with st.spinner(text=f"Trying to understand '{tw_user}' tweets... 🤔"):
-        embeddings = embed_text(tws, model)
+        embeddings = embed_text(tws_cleaned, model)
     # encoded_labels = encode_labels(labels)
     cluster = hdbscan.HDBSCAN(
         min_cluster_size=3,
@@ -173,9 +174,8 @@ if go_btn and tw_user != '':
         tweets_objs += tweets_response.data
     tweets_txt = [tweet.text for tweet in tweets_objs]
     tweets_txt = list(set(tweets_txt))
-    tweets_txt = preprocess(tweets_txt)
-    # plot = generate_plot(df, text_column, label_column, sample, dimensionality_reduction_function, model)
-    plot = generate_plot(tweets_txt, model, tw_user)
+    tweets_txt_cleaned = preprocess(tweets_txt)
+    plot = generate_plot(tweets_txt, tweets_txt_cleaned, model, tw_user)
     st.bokeh_chart(plot)
 elif go_btn and tw_user == '':
     st.warning('Twitter handler field is empty 🙄')