AAP-51301: Modify ansible-chatbot-stack build to include dependencies for gemini provider

Makefile

@@ -124,6 +124,7 @@ run: check-env-run
 	  --env VLLM_API_TOKEN=$(ANSIBLE_CHATBOT_VLLM_API_TOKEN) \
 	  --env INFERENCE_MODEL=$(ANSIBLE_CHATBOT_INFERENCE_MODEL) \
 	  --env INFERENCE_MODEL_FILTER=$(ANSIBLE_CHATBOT_INFERENCE_MODEL_FILTER) \
+	  --env GEMINI_API_KEY=$(GEMINI_API_KEY) \
 	  ansible-chatbot-stack:$(ANSIBLE_CHATBOT_VERSION)
 
 run-test:
@@ -162,6 +163,7 @@ run-local-db: check-env-run-local-db
 	  --env VLLM_API_TOKEN=$(ANSIBLE_CHATBOT_VLLM_API_TOKEN) \
 	  --env INFERENCE_MODEL=$(ANSIBLE_CHATBOT_INFERENCE_MODEL) \
 	  --env INFERENCE_MODEL_FILTER=$(ANSIBLE_CHATBOT_INFERENCE_MODEL_FILTER) \
+	  --env GEMINI_API_KEY=$(GEMINI_API_KEY) \
 	  ansible-chatbot-stack:$(ANSIBLE_CHATBOT_VERSION)
 
 clean:

README.md

@@ -154,6 +154,20 @@ Runs basic tests against the local container.
     kubectl apply -f my-chatbot-stack-deploy.yaml
 ```
 
+## Appendix - Google Gemini
+
+* Please set the environment variable `OPENAI_API_KEY=<YOUR_API_KEY>`
+* Example of a `v1/query` request:
+```json
+{
+    "query": "hello",
+    "system_prompt": "You are a helpful assistant.",
+    "model": "gemini/gemini-2.5-flash",
+    "provider": "gemini"
+}
+```
+
+
 ## Appendix - Host clean-up
 
 If you have the need for re-building images, apply the following clean-ups right before:

ansible-chatbot-mcp-run.yaml

@@ -18,6 +18,10 @@ providers:
       max_tokens: ${env.VLLM_MAX_TOKENS:=4096}
       api_token: ${env.VLLM_API_TOKEN:=fake}
       tls_verify: ${env.VLLM_TLS_VERIFY:=true}
+  - provider_id: gemini
+    provider_type: remote::gemini
+    config:
+      api_key: ${env.GEMINI_API_KEY:=fake}
   - provider_id: inline_sentence-transformer
     provider_type: inline::sentence-transformers
     config: {}
@@ -85,6 +89,11 @@ models:
   model_id: ${env.EMBEDDINGS_MODEL:=/.llama/data/distributions/ansible-chatbot/embeddings_model}
   provider_id: inline_sentence-transformer
   model_type: embedding
+- metadata: {}
+  model_id: ${env.GEMINI_INFERENCE_MODEL:=gemini/gemini-2.5-flash}
+  provider_id: gemini
+  provider_model_id: gemini/gemini-2.5-flash
+  model_type: llm
 shields: []
 vector_dbs:
 - metadata: {}

ansible-chatbot-run.yaml

@@ -18,6 +18,10 @@ providers:
       max_tokens: ${env.VLLM_MAX_TOKENS:=4096}
       api_token: ${env.VLLM_API_TOKEN:=fake}
       tls_verify: ${env.VLLM_TLS_VERIFY:=true}
+  - provider_id: gemini
+    provider_type: remote::gemini
+    config:
+      api_key: ${env.GEMINI_API_KEY:=fake}
   - provider_id: inline_sentence-transformer
     provider_type: inline::sentence-transformers
     config: {}
@@ -85,6 +89,11 @@ models:
   model_id: ${env.EMBEDDINGS_MODEL:=/.llama/data/distributions/ansible-chatbot/embeddings_model}
   provider_id: inline_sentence-transformer
   model_type: embedding
+- metadata: {}
+  model_id: ${env.GEMINI_INFERENCE_MODEL:=gemini/gemini-2.5-flash}
+  provider_id: gemini
+  provider_model_id: gemini/gemini-2.5-flash
+  model_type: llm
 shields: []
 vector_dbs:
 - metadata: {}

pyproject.toml

@@ -15,6 +15,7 @@ dependencies = [
     "opentelemetry-exporter-otlp~=1.34.1",
     "sentence-transformers>=5.0.0",
     "sqlalchemy~=2.0.41",
+    "litellm~=1.75.3",
 ]
 
 [dependency-groups]

requirements.txt

@@ -37,6 +37,7 @@ joblib==1.5.1
 jsonschema==4.24.0
 jsonschema-specifications==2025.4.1
 lightspeed-stack-providers==0.1.14
+litellm==1.75.5.post1
 llama-api-client==0.1.2
 llama-stack==0.2.16
 llama-stack-client==0.2.16
@@ -62,7 +63,7 @@ nvidia-cusparselt-cu12==0.6.3 ; platform_machine == 'x86_64' and sys_platform ==
 nvidia-nccl-cu12==2.26.2 ; platform_machine == 'x86_64' and sys_platform == 'linux'
 nvidia-nvjitlink-cu12==12.6.85 ; platform_machine == 'x86_64' and sys_platform == 'linux'
 nvidia-nvtx-cu12==12.6.77 ; platform_machine == 'x86_64' and sys_platform == 'linux'
-openai==1.91.0
+openai==1.99.9
 opentelemetry-api==1.34.1
 opentelemetry-exporter-otlp==1.34.1
 opentelemetry-exporter-otlp-proto-common==1.34.1