Create build_retriever.py

build_retriever.py

@@ -0,0 +1,21 @@
+# build_retriever.py
+
+import faiss
+import json
+from sentence_transformers import SentenceTransformer
+
+# Load the extracted text
+with open("apexcustoms.json", "r") as f:
+    documents = json.load(f)
+
+# Load a pre-trained sentence transformer model
+model = SentenceTransformer('paraphrase-MiniLM-L6-v2')
+document_embeddings = model.encode(documents)
+
+# Create a FAISS index
+index = faiss.IndexFlatL2(document_embeddings.shape[1])
+index.add(document_embeddings)
+
+# Save the FAISS index and the embeddings
+faiss.write_index(index, "apexcustoms_index.faiss")
+model.save("sentence_transformer_model")