cj-dev-code/semantic_search

1import json2from azure.search.documents import SearchClient3from azure.core.credentials import AzureKeyCredential4 5from dotenv import load_dotenv6import os7 8if __name__ == "__main__":9    load_dotenv()10 11    #     🔐 Replace with your actual info12    endpoint = os.getenv('AZURE_AAIS_ENDPT')13    key = os.getenv('AZURE_AAIS_KEY')14    index_name = "main"15 16    client = SearchClient(17        endpoint=endpoint,18        index_name=index_name,19        credential=AzureKeyCredential(key)20    )21 22    # 📄 Load your documents23    docs = []24    with open("data/chunks_indexed.jsonl", "r", encoding="utf-8") as f:25        for line in f:26            doc = json.loads(line)27            doc["id"] = str(doc["id"])  # force id to string bc error idw pay more for28            docs.append(doc)29 30    # 🚀 Upload in batches (Azure supports up to 1000 per batch)31    batch_size = 100032    for i in range(0, len(docs), batch_size):33        batch = docs[i:i + batch_size]34        result = client.upload_documents(documents=batch)35        print(f"Batch {i // batch_size + 1}:")36