added visualizations

app.py

@@ -82,7 +82,58 @@ def similarity_top(descr_emb, disorder_embs):
 
   return results[:5]
 
+def vis_results_2d(input_embed):
 
+    # performing dimensionality reduction using PCA
+    pca = PCA(n_components=2)
+    disease_embeddings_2d = pca.fit_transform(icd_embeddings)
+
+    # creating a DataFrame for disease embeddings plot
+    disease_data_df = pd.DataFrame(disease_embeddings_2d, columns=['PC1', 'PC2'])
+    disease_data_df['Type'] = 'Disease'
+    disease_data_df['Name'] = disease_names
+
+    input_embed_2d = input_embed.reshape(1, -1)
+    input_embed_2d = pca.transform(input_embed_2d)
+
+    # creating a DataFrame for character embedding plot
+    pca_2d = pd.DataFrame(input_embed_2d, columns=['PC1', 'PC2'])
+    pca_2d['Type'] = 'Character'
+    pca_2d['Your character'] = 'Your character'
+
+    # concatenating the two DataFrames
+    combined_2d = pd.concat([disease_data_df, pca_2d], ignore_index=True)
+
+    # creating an interactive 3D scatter plot
+    fig = px.scatter(combined_2d, x='PC1', y='PC2', text='Name', color='Type', symbol='Type', width=800, height=800)
+    fig.show()
+
+    
+def vis_results_3d(input_embed):
+
+    # performing dimensionality reduction using PCA
+    pca = PCA(n_components=3)
+    disease_embeddings_3d = pca.fit_transform(icd_embeddings)
+
+    # creating a DataFrame for disease embeddings plot
+    disease_data_df = pd.DataFrame(disease_embeddings_3d, columns=['PC1', 'PC2', 'PC3'])
+    disease_data_df['Type'] = 'Disease'
+    disease_data_df['Name'] = disease_names
+
+    input_embed_2d = input_embed.reshape(1, -1)
+    input_embed_3d = pca.transform(input_embed_2d)
+    
+    # creating a DataFrame for character embedding plot
+    pca_3d = pd.DataFrame(input_embed_3d, columns=['PC1', 'PC2', 'PC3'])
+    pca_3d['Type'] = 'Character'
+    pca_3d['Your character'] = 'Your character'
+
+    # concatenating the two DataFrames
+    combined_3d = pd.concat([disease_data_df, pca_3d], ignore_index=True)
+
+    # creating an interactive 3D scatter plot
+    fig = px.scatter_3d(combined_3d, x='PC1', y='PC2', z='PC3', text='Name', color='Type', symbol='Type', width=800, height=800)
+    fig.show()
 # with text_spinner_placeholder:
 #     with st.spinner("Please wait while your Tweet is being generated..."):
 #         mood_prompt = f"{mood} " if mood else ""
@@ -95,24 +146,21 @@ def similarity_top(descr_emb, disorder_embs):
 #                 f"and in the style of the following Tweets:\n\n{tweets_prompt}\n\n"
                     
 # Configure Streamlit page and state
-st.title("Detect the disorder")
+st.title("Detect your character's mental disorder!")
 st.markdown(
-    "This mini-app predicts a mental disorder based on your description."
+    "This mini-app predicts top-5 most likely mental disorder based on your description. The more information you provide, the more informative the results will be. \
+    Note that this app can't be used for diagnostic purposes."
 )
 
-input = st.text_input(label="Your description)", placeholder="Insert a description of a character")
+input = st.text_input(label="Your description", placeholder="Insert a description of your character")
 if input:
     input_embed = model.encode(input)
     sim_score = similarity_top(input_embed, icd_embeddings)
     st.write(sim_score)
+
+    vis_results_2d(input_embed)
+    vis_results_3d(input_embed)
+    
     
-# mood = st.text_input(
-#     label="Mood (e.g. inspirational, funny, serious) (optional)",
-#     placeholder="inspirational",
-# )
-# style = st.text_input(
-#     label="Twitter account handle to style-copy recent Tweets (optional, limited by Twitter's API)",
-#     placeholder="elonmusk",
-# )
 
 text_spinner_placeholder = st.empty()

requirements.txt

@@ -1 +1,7 @@
+nltk==3.8.1
+numpy==1.23.5
+pandas==1.5.3
+plotly==5.15.0
+scikit_learn==1.3.1
+sentence_transformers==2.2.2
 streamlit==1.26.0