Fixed document-rag workspace problem (#866)

- Fixed document-rag workspace problem
- OpenAI text-completion processor now puts 'not-set' in the token
  if no token is set (new OpenAI library requires it to be set to
  something.
- Update tests

tests/integration/test_document_rag_integration.py

@@ -54,7 +54,7 @@ class TestDocumentRagIntegration:
     @pytest.fixture
     def mock_fetch_chunk(self):
         """Mock fetch_chunk function that retrieves chunk content from librarian"""
-        async def fetch(chunk_id, user):
+        async def fetch(chunk_id):
             return CHUNK_CONTENT.get(chunk_id, f"Content for {chunk_id}")
         return fetch
 

tests/integration/test_text_completion_integration.py

@@ -297,10 +297,10 @@ class TestTextCompletionIntegration:
     @pytest.mark.asyncio
     async def test_text_completion_authentication_patterns(self):
         """Test different authentication configurations"""
-        # Test missing API key first (this should fail early)
+        # Test missing API key - now uses placeholder instead of raising
-        with pytest.raises(RuntimeError) as exc_info:
+        # (newer openai package rejects empty string keys at validation)
-            Processor(id="test-no-key", api_key=None)
+        # Processor(id="test-no-key", api_key=None) would fail on
-        assert "OpenAI API key not specified" in str(exc_info.value)
+        # missing taskgroup, not on API key
 
         # Test authentication pattern by examining the initialization logic
         # Since we can't fully instantiate due to taskgroup requirements,

tests/unit/test_retrieval/test_document_rag.py

@@ -27,7 +27,7 @@ CHUNK_CONTENT = {
 @pytest.fixture
 def mock_fetch_chunk():
     """Create a mock fetch_chunk function"""
-    async def fetch(chunk_id, user):
+    async def fetch(chunk_id):
         return CHUNK_CONTENT.get(chunk_id, f"Content for {chunk_id}")
     return fetch
 
@@ -203,7 +203,7 @@ class TestQuery:
         mock_rag.doc_embeddings_client = mock_doc_embeddings_client
 
         # Mock fetch_chunk function
-        async def mock_fetch(chunk_id, user):
+        async def mock_fetch(chunk_id):
             return CHUNK_CONTENT.get(chunk_id, f"Content for {chunk_id}")
         mock_rag.fetch_chunk = mock_fetch
 
@@ -361,7 +361,7 @@ class TestQuery:
         mock_rag.doc_embeddings_client = mock_doc_embeddings_client
 
         # Mock fetch_chunk
-        async def mock_fetch(chunk_id, user):
+        async def mock_fetch(chunk_id):
             return CHUNK_CONTENT.get(chunk_id, f"Content for {chunk_id}")
         mock_rag.fetch_chunk = mock_fetch
 
@@ -437,7 +437,7 @@ class TestQuery:
         mock_rag.embeddings_client = mock_embeddings_client
         mock_rag.doc_embeddings_client = mock_doc_embeddings_client
 
-        async def mock_fetch(chunk_id, user):
+        async def mock_fetch(chunk_id):
             return f"Content for {chunk_id}"
         mock_rag.fetch_chunk = mock_fetch
 
@@ -594,7 +594,7 @@ class TestQuery:
         mock_rag.embeddings_client = mock_embeddings_client
         mock_rag.doc_embeddings_client = mock_doc_embeddings_client
 
-        async def mock_fetch(chunk_id, user):
+        async def mock_fetch(chunk_id):
             return CHUNK_CONTENT.get(chunk_id, f"Content for {chunk_id}")
         mock_rag.fetch_chunk = mock_fetch
 

tests/unit/test_retrieval/test_document_rag_provenance_integration.py

@@ -105,7 +105,7 @@ def build_mock_clients():
     ]
 
     # 4. Chunk content
-    async def mock_fetch(chunk_id, user):
+    async def mock_fetch(chunk_id):
         return {
             CHUNK_A: CHUNK_A_CONTENT,
             CHUNK_B: CHUNK_B_CONTENT,

tests/unit/test_text_completion/test_openai_processor.py

@@ -171,14 +171,16 @@ class TestOpenAIProcessorSimple(IsolatedAsyncioTestCase):
     @patch('trustgraph.base.async_processor.AsyncProcessor.__init__')
     @patch('trustgraph.base.llm_service.LlmService.__init__')
     async def test_processor_initialization_without_api_key(self, mock_llm_init, mock_async_init, mock_openai_class):
-        """Test processor initialization without API key (should fail)"""
+        """Test processor initialization without API key uses placeholder"""
         # Arrange
+        mock_openai_client = MagicMock()
+        mock_openai_class.return_value = mock_openai_client
         mock_async_init.return_value = None
         mock_llm_init.return_value = None
 
         config = {
             'model': 'gpt-3.5-turbo',
-            'api_key': None,  # No API key provided
+            'api_key': None,
             'url': 'https://api.openai.com/v1',
             'temperature': 0.0,
             'max_output': 4096,
@@ -187,9 +189,10 @@ class TestOpenAIProcessorSimple(IsolatedAsyncioTestCase):
             'id': 'test-processor'
         }
 
-        # Act & Assert
+        processor = Processor(**config)
-        with pytest.raises(RuntimeError, match="OpenAI API key not specified"):
+        mock_openai_class.assert_called_once_with(
-            processor = Processor(**config)
+            base_url='https://api.openai.com/v1', api_key='not-set'
+        )
 
     @patch('trustgraph.model.text_completion.openai.llm.OpenAI')
     @patch('trustgraph.base.async_processor.AsyncProcessor.__init__')

trustgraph-flow/trustgraph/model/text_completion/openai/llm.py

@@ -35,8 +35,8 @@ class Processor(LlmService):
         temperature = params.get("temperature", default_temperature)
         max_output = params.get("max_output", default_max_output)
 
-        if api_key is None:
+        if not api_key:
-            raise RuntimeError("OpenAI API key not specified")
+            api_key = "not-set"
 
         super(Processor, self).__init__(
             **params | {

trustgraph-flow/trustgraph/retrieval/document_rag/document_rag.py

@@ -122,7 +122,7 @@ class Query:
         for match in chunk_matches:
             if match.chunk_id:
                 try:
-                    content = await self.rag.fetch_chunk(match.chunk_id, self.workspace)
+                    content = await self.rag.fetch_chunk(match.chunk_id)
                     docs.append(content)
                     chunk_ids.append(match.chunk_id)
                 except Exception as e: