Add sidebar and fix distances

app.py

@@ -21,6 +21,17 @@ def load_embeddings():
 embeddings = load_embeddings()
 vectors_2d, reducer = reduce_embeddings(embeddings)
 
+def sidebar():
+    st.sidebar.title("About this app")
+    st.sidebar.markdown(
+        "This app is intended to give a more intuitive and interactive understanding of sequence embeddings (e.g. sentence), \n"
+        "through interactive plots and operations with these embeddings, with a focus on embedding inversion.\n"
+        "We explore both sequence embedding inversion using the method described in [Morris et al., 2023](https://arxiv.org/abs/2310.06816), as well as"
+        " dimensionality rediction transforms and inverse transforms, and its effect on embedding inversion."
+    )
+
+sidebar()
+
 tab1, tab2 = st.tabs(["plot", "diffs"])
 
 with tab1:

views.py

@@ -8,6 +8,7 @@ from streamlit_plotly_events import plotly_events
 from resources import reduce_embeddings
 import utils
 import pandas as pd
+from scipy.spatial import distance
 
 
 def diffs(embeddings: np.ndarray, corrector):
@@ -37,7 +38,7 @@ def plot(df: pd.DataFrame, embeddings: np.ndarray, vectors_2d, reducer, correcto
 
     x, y = 0.0, 0.0
     vec = np.array([x, y]).astype("float32")
-
+    inferred_embedding = None
     # Add a card container to the right of the content with Streamlit columns
     col1, col2 = st.columns([0.6, 0.4])  # Adjusting ratio to allocate space for the card container
 
@@ -48,8 +49,8 @@ def plot(df: pd.DataFrame, embeddings: np.ndarray, vectors_2d, reducer, correcto
         with st.form(key="form1_main"):
             if selected_points:
                 clicked_point = selected_points[0]
-                x_coord = x = clicked_point['x']
-                y_coord = y = clicked_point['y']
+                x = clicked_point['x']
+                y = clicked_point['y']
 
             x = st.number_input("X Coordinate", value=x, format="%.10f")
             y = st.number_input("Y Coordinate", value=y, format="%.10f")
@@ -75,6 +76,17 @@ def plot(df: pd.DataFrame, embeddings: np.ndarray, vectors_2d, reducer, correcto
 
     with col2:
         closest_sentence_index = utils.find_exact_match(vectors_2d, vec, decimals=3)
+        selected_sentence = df.title.iloc[closest_sentence_index] if closest_sentence_index > -1 else None
+        selected_sentence_embedding = embeddings[closest_sentence_index] if closest_sentence_index > -1 else None
+
         st.markdown(
-            f"### Selected text:\n```console\n{df.title.iloc[closest_sentence_index] if closest_sentence_index > -1 else '[no selected text]'}\n```"
+            f"### Selected text:\n```console\n{selected_sentence}\n```"
         )
+
+        if inferred_embedding is not None and (closest_sentence_index != -1):
+            couple = selected_sentence_embedding.squeeze(), inferred_embedding.squeeze()
+            st.markdown(f"### Inferred embedding distance:")
+            st.number_input("Euclidean", value=distance.euclidean(
+                *couple
+            ), disabled=True)
+            st.number_input("Cosine", value=distance.cosine(*couple), disabled=True)