fix: throw errors if azure deployments are no

.changeset/smooth-points-float.md

@@ -2,4 +2,4 @@
 "create-llama": patch
 ---
 
-Add azure model provider
+Add Azure OpenAI as model provider

templates/components/settings/python/settings.py

@@ -75,15 +75,15 @@ def init_azure_openai():
     from llama_index.embeddings.azure_openai import AzureOpenAIEmbedding
     from llama_index.llms.azure_openai import AzureOpenAI
 
-    llm_deployment = os.getenv("AZURE_OPENAI_LLM_DEPLOYMENT")
-    embedding_deployment = os.getenv("AZURE_OPENAI_EMBEDDING_DEPLOYMENT")
+    llm_deployment = os.environ["AZURE_OPENAI_LLM_DEPLOYMENT"]
+    embedding_deployment = os.environ["AZURE_OPENAI_EMBEDDING_DEPLOYMENT"]
     max_tokens = os.getenv("LLM_MAX_TOKENS")
     temperature = os.getenv("LLM_TEMPERATURE", DEFAULT_TEMPERATURE)
     dimensions = os.getenv("EMBEDDING_DIM")
 
     azure_config = {
-        "api_key": os.getenv("AZURE_OPENAI_KEY"),
-        "azure_endpoint": os.getenv("AZURE_OPENAI_ENDPOINT"),
+        "api_key": os.environ["AZURE_OPENAI_KEY"],
+        "azure_endpoint": os.environ["AZURE_OPENAI_ENDPOINT"],
         "api_version": os.getenv("AZURE_OPENAI_API_VERSION")
         or os.getenv("OPENAI_API_VERSION"),
     }