Create upload_to_supabase.py

upload_to_supabase.py

@@ -0,0 +1,36 @@
+import os
+import pandas as pd
+from tqdm import tqdm
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.vectorstores import SupabaseVectorStore
+from langchain.schema.document import Document
+from supabase import create_client, Client
+
+# --- Load Environment Variables ---
+SUPABASE_URL = os.getenv("SUPABASE_URL")
+SUPABASE_KEY = os.getenv("SUPABASE_KEY")
+
+# --- Init Supabase & Embeddings ---
+supabase: Client = create_client(SUPABASE_URL, SUPABASE_KEY)
+embedding_model = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")  # Or OpenAIEmbeddings if you use Groq
+
+# --- Read CSV File ---
+df = pd.read_csv("supabase_docs.csv")  # Assuming columns: 'content', 'metadata' or just 'content'
+
+# --- Convert rows to LangChain Document objects ---
+documents = []
+for _, row in tqdm(df.iterrows(), total=len(df)):
+    content = str(row["content"])
+    metadata = row.drop("content").to_dict() if "content" in row else {}
+    documents.append(Document(page_content=content, metadata=metadata))
+
+# --- Create Supabase Vector Store and Upload ---
+vectorstore = SupabaseVectorStore.from_documents(
+    documents=documents,
+    embedding=embedding_model,
+    client=supabase,
+    table_name="documents",
+    query_name="match_documents_langchain"
+)
+
+print("✅ Upload complete.")