Model name is configurable for the cloud function (#40)

explore-assistant-cloud-function/README.md

@@ -55,3 +55,15 @@ To set up and run the function locally, follow these steps:
    ```
 
 This setup allows developers to test and modify the function in a local environment before deploying it to a cloud function service.
+
+## Model configuration
+
+By default, the cloud function will use a default model. However, you may want to test out different Gemini models are they are released. We have made the model name configurable via an environment variable. 
+
+In development, you can run the main script with a new MODEL_NAME variable:
+
+```bash
+PROJECT=XXXX LOCATION=us-central-1 VERTEX_CF_AUTH_TOKEN=$(cat ../.vertex_cf_auth_token) MODEL_NAME=XXXXX python main.py
+```
+
+In production, on the cloud function, you can manually set a variable in the GCP UI. Updating the variable will re-deploy the cloud function.

explore-assistant-cloud-function/main.py

@@ -37,6 +37,8 @@
 project = os.environ.get("PROJECT")
 location = os.environ.get("REGION")
 vertex_cf_auth_token = os.environ.get("VERTEX_CF_AUTH_TOKEN")
+model_name = os.environ.get("MODEL_NAME", "gemini-1.0-pro-001")
+
 vertexai.init(project=project, location=location)
 
 def get_response_headers(request):
@@ -61,7 +63,7 @@ def has_valid_signature(request):
 
     return hmac.compare_digest(signature, expected_signature)
 
-def generate_looker_query(contents, parameters=None, model_name="gemini-1.0-pro-001"):
+def generate_looker_query(contents, parameters=None):
 
    # Define default parameters
     default_parameters = {