Create cache_embeddings.py

cache_embeddings.py

@@ -0,0 +1,16 @@
+from sentence_transformers import SentenceTransformer
+from messages import keyword_groups, krishna_blessings
+import joblib
+import os
+
+# Initialize model
+model = SentenceTransformer('all-MiniLM-L6-v2')
+
+# Compute embeddings
+embeddings_cache = {}
+for group, keywords in keyword_groups.items():
+    keyword_texts = keywords + [krishna_blessings.get(k, "") for k in keywords if k in krishna_blessings]
+    embeddings_cache[group] = model.encode(keyword_texts, convert_to_tensor=True)
+
+# Save to file
+joblib.dump(embeddings_cache, 'embeddings_cache.joblib')