Release v0.2.43

.github/workflows/e2e-tests.yaml

@@ -1,18 +1,6 @@
 name: E2E Tests
 
-on:
-  pull_request:
-    branches:
-      - main
-    paths:
-      - 'frontend/**'
-      - 'deploy/helm/**'
-      - 'tests/**'
-      - '.github/workflows/e2e-tests.yaml'
-  push:
-    branches:
-      - main
-  workflow_dispatch:
+on: [pull_request, push, workflow_dispatch]
 
 # MaaS configuration - can be overridden with repository secrets for different environments
 env:
@@ -754,3 +742,67 @@ jobs:
         run: |
           pkill -f "kubectl port-forward" || true
           # kind delete cluster --name rag-e2e-ui
+
+  # Single check to mark as required in branch protection (Settings → Branches →
+  # Require status checks): "PR tests gate". Fails if unit/integration failed, or
+  # if MaaS-backed jobs ran and failed. Fork PRs may skip secret-dependent jobs.
+  pr-required-checks:
+    name: PR tests gate
+    runs-on: ubuntu-latest
+    if: always() && github.event_name == 'pull_request'
+    needs:
+      - unit-tests
+      - integration-tests
+      - llamastack-integration-tests
+      - ui-e2e-tests
+    steps:
+      - name: Enforce test job outcomes
+        env:
+          UNIT_RESULT: ${{ needs.unit-tests.result }}
+          INTEG_RESULT: ${{ needs.integration-tests.result }}
+          LLAMA_RESULT: ${{ needs.llamastack-integration-tests.result }}
+          UI_RESULT: ${{ needs.ui-e2e-tests.result }}
+          HEAD_REPO: ${{ github.event.pull_request.head.repo.full_name }}
+          THIS_REPO: ${{ github.repository }}
+        run: |
+          set -euo pipefail
+          is_fork=false
+          if [ "${HEAD_REPO}" != "${THIS_REPO}" ]; then
+            is_fork=true
+          fi
+
+          echo "Job results: unit=${UNIT_RESULT} integration=${INTEG_RESULT} llamastack=${LLAMA_RESULT} ui-e2e=${UI_RESULT} fork=${is_fork}"
+
+          if [ "${UNIT_RESULT}" != "success" ]; then
+            echo "::error::Unit tests did not succeed (result=${UNIT_RESULT})"
+            exit 1
+          fi
+          if [ "${INTEG_RESULT}" != "success" ]; then
+            echo "::error::Integration tests (Streamlit) did not succeed (result=${INTEG_RESULT})"
+            exit 1
+          fi
+
+          allow_skipped_secret_jobs() {
+            local r="$1"
+            [ "${is_fork}" = true ] && [ "${r}" = "skipped" ]
+          }
+
+          if [ "${LLAMA_RESULT}" != "success" ]; then
+            if allow_skipped_secret_jobs "${LLAMA_RESULT}"; then
+              echo "LlamaStack integration tests were skipped (fork PR; repository secrets are not available)."
+            else
+              echo "::error::LlamaStack integration tests did not succeed (result=${LLAMA_RESULT})"
+              exit 1
+            fi
+          fi
+
+          if [ "${UI_RESULT}" != "success" ]; then
+            if allow_skipped_secret_jobs "${UI_RESULT}"; then
+              echo "UI E2E tests were skipped (fork PR; repository secrets are not available)."
+            else
+              echo "::error::UI E2E tests did not succeed (result=${UI_RESULT})"
+              exit 1
+            fi
+          fi
+
+          echo "All required test outcomes for this pull request passed."

deploy/helm/rag/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: rag
 description: A Helm chart for Kubernetes
 type: application
-version: 0.2.42
-appVersion: "0.2.42"
+version: 0.2.43
+appVersion: "0.2.43"
 
 dependencies:
   - name: pgvector

deploy/helm/rag/values.yaml

@@ -3,7 +3,7 @@ replicaCount: 1
 image:
   repository: quay.io/rh-ai-quickstart/llamastack-dist-ui
   pullPolicy: Always
-  tag: 0.2.42
+  tag: 0.2.43
 
 service:
   type: ClusterIP
@@ -164,19 +164,6 @@ pgvector:
     dbname: rag_blueprint
     host: pgvector
     port: "5432"
-
-  # Create a separate vector database for each ingestion pipeline
-  extraDatabases:
-    - name: hr_vector_db
-      vectordb: true
-    - name: legal_vector_db
-      vectordb: true
-    - name: sales_vector_db
-      vectordb: true
-    - name: procurement_vector_db
-      vectordb: true
-    - name: techsupport_vector_db
-      vectordb: true
 
     # Upload sample files to the minio bucket 
   sampleFileUpload:

tests/e2e_ui/requirements.txt

@@ -2,4 +2,6 @@ pytest==8.3.3
 pytest-playwright==0.5.2
 playwright==1.48.0
 requests==2.32.3
+python-docx
+openpyxl
 

tests/e2e_ui/test_chat_ui.py

@@ -183,12 +183,6 @@ def test_initial_greeting_message(self, page: Page):
         """Test that initial greeting message is displayed"""
         greeting = page.get_by_text("How can I help you?", exact=False)
         expect(greeting).to_be_visible(timeout=TEST_TIMEOUT)
-
-    def test_tool_debug_toggle(self, page: Page):
-        """Test that tool debug toggle is visible"""
-        debug_toggle = page.get_by_text("Show Tool/Debug Info", exact=False)
-        expect(debug_toggle).to_be_visible(timeout=TEST_TIMEOUT)
-
 
 class TestMaaSIntegration:
     """UI tests for MaaS (Model-as-a-Service) integration through the UI

tests/integration/llamastack/test_user_workflow.py

@@ -286,18 +286,41 @@ def test_complete_rag_workflow():
     print("🤖 Step 4: Checking for available models...")
     skip_inference = SKIP_MODEL_TESTS == "true"
     model_available = False
-
+    model_ids = []
+
     try:
-        # Llama-stack uses /v1/openai/v1/* paths for OpenAI-compatible API
-        openai_base_url = f"{LLAMA_STACK_ENDPOINT}/v1/openai/v1"
-        client = OpenAI(
-            api_key="not_needed",
-            base_url=openai_base_url,
-            timeout=30.0
-        )
-
-        print(f"   DEBUG: Calling {openai_base_url}/models endpoint...")
-        models = client.models.list()
+        # OpenAI-compatible base URL: current Llama Stack uses /v1 (see conftest.py).
+        # Older stacks used /v1/openai/v1 (client-examples-python/README.md).
+        endpoint = LLAMA_STACK_ENDPOINT.rstrip("/")
+        explicit_openai = os.getenv("LLAMA_STACK_OPENAI_BASE")
+        openai_base_candidates = []
+        if explicit_openai:
+            openai_base_candidates.append(explicit_openai.rstrip("/"))
+        openai_base_candidates.append(f"{endpoint}/v1")
+        openai_base_candidates.append(f"{endpoint}/v1/openai/v1")
+
+        client = None
+        openai_base_url = None
+        models = None
+        last_models_error: Exception | None = None
+        for base in openai_base_candidates:
+            candidate = OpenAI(
+                api_key="not_needed",
+                base_url=base,
+                timeout=30.0,
+            )
+            try:
+                print(f"   DEBUG: Calling {base}/models endpoint...")
+                models = candidate.models.list()
+                client = candidate
+                openai_base_url = base
+                break
+            except Exception as e:
+                last_models_error = e
+                continue
+
+        if client is None or models is None:
+            raise last_models_error or RuntimeError("Could not list models from any OpenAI base URL")
         print(f"   DEBUG: Raw response type: {type(models)}")
         print(f"   DEBUG: Number of models in response: {len(models.data)}")
 

tests/integration/test_upload_integration.py

@@ -64,7 +64,7 @@ def mock_uploaded_file():
 class TestDocumentUploadIntegration:
     """Integration tests for document upload workflow"""
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_single_file_upload_workflow(self, mock_api, mock_uploaded_file):
         """Test complete workflow for uploading a single file"""
         from llama_stack_client import RAGDocument
@@ -113,7 +113,7 @@ def test_single_file_upload_workflow(self, mock_api, mock_uploaded_file):
         mock_api.client.vector_dbs.register.assert_called_once()
         mock_api.client.tool_runtime.rag_tool.insert.assert_called_once()
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_multiple_files_upload_workflow(self, mock_api):
         """Test uploading multiple files at once"""
         from llama_stack_client import RAGDocument
@@ -182,7 +182,7 @@ def test_file_type_validation(self):
 class TestVectorDBCreation:
     """Integration tests for vector database creation"""
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_vector_db_registration_params(self, mock_api):
         """Test vector DB registration with correct parameters"""
         mock_api.client.providers.list.return_value = [
@@ -204,7 +204,7 @@ def test_vector_db_registration_params(self, mock_api):
         assert call_args[1]['embedding_model'] == "all-MiniLM-L6-v2"
         assert call_args[1]['provider_id'] == "pgvector"
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_vector_db_with_custom_name(self, mock_api):
         """Test creating vector DB with custom name"""
         mock_api.client.providers.list.return_value = [
@@ -227,7 +227,7 @@ def test_vector_db_with_custom_name(self, mock_api):
 class TestProviderDetection:
     """Integration tests for provider detection"""
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_vector_io_provider_detection(self, mock_api):
         """Test that vector_io provider is correctly detected"""
         mock_api.client.providers.list.return_value = [
@@ -244,7 +244,7 @@ def test_vector_io_provider_detection(self, mock_api):
 
         assert vector_io_provider == "pgvector"
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_no_vector_io_provider(self, mock_api):
         """Test handling when no vector_io provider is available"""
         mock_api.client.providers.list.return_value = [
@@ -264,7 +264,7 @@ def test_no_vector_io_provider(self, mock_api):
 class TestDocumentInsertion:
     """Integration tests for document insertion into vector DB"""
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_document_insertion_with_chunks(self, mock_api):
         """Test document insertion with chunking"""
         from llama_stack_client import RAGDocument
@@ -288,7 +288,7 @@ def test_document_insertion_with_chunks(self, mock_api):
         call_args = mock_api.client.tool_runtime.rag_tool.insert.call_args
         assert call_args[1]['chunk_size_in_tokens'] == 512
 
-    @patch('llama_stack_ui.distribution.ui.page.upload.upload.llama_stack_api')
+    @patch('llama_stack_ui.distribution.ui.modules.api.llama_stack_api')
     def test_empty_document_list(self, mock_api):
         """Test handling of empty document list"""
         documents = []

tests/unit/requirements.txt

@@ -4,4 +4,5 @@ pytest-cov==5.0.0
 llama-stack-client>=0.2.9,<0.2.13
 llama-stack
 streamlit>=1.31.0
-
+python-docx
+openpyxl