Create app.py

app.py

@@ -0,0 +1,31 @@
+import gradio as gr
+import numpy as np
+import voyageai
+import pickle
+from sklearn.metrics.pairwise import cosine_similarity
+
+def my_pipe(text):
+    # To load the numpy array
+    with open('embeddings.pkl', 'rb') as f:
+        embeddings = pickle.load(f)
+    with open('all_quotes.pkl', 'rb') as f:
+        all_quotes = pickle.load(f)
+    
+    text = [text]
+
+    vo = voyageai.Client(api_key="pa-pxjjMtiZrbP6e2gDl4AWHwsPzSTK00Uww0CnTUTW79U")
+    emb = np.array(vo.embed(text, model="voyage-2", input_type="document").embeddings)
+
+    similarities = cosine_similarity(embeddings, emb)
+    most_similar_index = np.argmax(similarities)
+    return all_quotes[most_similar_index]
+
+def launch(input):
+    out = my_pipe(input)
+    return out
+
+iface = gr.Interface(launch,
+                     inputs="text",
+                     outputs="text")
+
+iface.launch()