auto-append /v1 to per-model Ollama base_url, fix flaky tests

haiku_rag_slim/haiku/rag/embeddings/__init__.py

@@ -167,7 +167,9 @@ def get_embedder(config: AppConfig = Config) -> EmbedderWrapper:
 
     if provider == "ollama":
         # Use model-level base_url if set, otherwise fall back to providers config
-        base_url = embedding_model.base_url or f"{config.providers.ollama.base_url}/v1"
+        base_url = embedding_model.base_url or config.providers.ollama.base_url
+        if not base_url.rstrip("/").endswith("/v1"):
+            base_url = base_url.rstrip("/") + "/v1"
         model = OpenAIEmbeddingModel(
             model_name,
             provider=OllamaProvider(base_url=base_url),

haiku_rag_slim/haiku/rag/utils.py

@@ -158,8 +158,11 @@ def get_model(
             model_settings, OpenAIChatModelSettings, model_config
         )
 
-        # Use model-level base_url if set, otherwise fall back to providers config
-        base_url = model_config.base_url or f"{app_config.providers.ollama.base_url}/v1"
+        # Ollama's OpenAI-compatible API lives under /v1. Append it if the
+        # configured base_url doesn't already include it.
+        base_url = model_config.base_url or app_config.providers.ollama.base_url
+        if not base_url.rstrip("/").endswith("/v1"):
+            base_url = base_url.rstrip("/") + "/v1"
 
         return OpenAIChatModel(
             model_name=model,

tests/store/test_document_items.py

@@ -5,6 +5,7 @@ from haiku.rag.client.documents import (
     _store_document_with_chunks,
     _update_document_with_chunks,
 )
+from haiku.rag.config import AppConfig
 from haiku.rag.store.engine import Store
 from haiku.rag.store.models.document_item import (
     DocumentItem,
@@ -646,9 +647,11 @@ class TestPictureDataPreservedThroughRoundTrip:
 
         docling_doc = _docling_doc_with_picture()
 
-        async with HaikuRAG(temp_db_path, create=True) as rag:
-            # Preservation only kicks in under modes that retain picture bytes.
-            rag._config.processing.pictures = "image"
+        config = AppConfig()
+        # Preservation only kicks in under modes that retain picture bytes.
+        config.processing.pictures = "image"
+
+        async with HaikuRAG(temp_db_path, config=config, create=True) as rag:
             document = Document(content="Hello world", uri="test://doc")
             document.set_docling(docling_doc)
             created = await _store_document_with_chunks(rag, document, [], docling_doc)

tests/test_config.py

@@ -1,3 +1,5 @@
+import logging
+
 import pytest
 import yaml
 
@@ -7,6 +9,16 @@ from haiku.rag.config.loader import (
     generate_default_config,
     load_yaml_config,
 )
+from haiku.rag.config.loader import logger as loader_logger
+
+
+class _ListHandler(logging.Handler):
+    def __init__(self) -> None:
+        super().__init__(level=logging.WARNING)
+        self.records: list[logging.LogRecord] = []
+
+    def emit(self, record: logging.LogRecord) -> None:
+        self.records.append(record)
 
 
 def test_load_yaml_config(tmp_path):
@@ -237,7 +249,7 @@ def _write(tmp_path, body: str):
     return p
 
 
-def test_load_yaml_legacy_picture_description_maps_to_description(tmp_path, caplog):
+def test_load_yaml_legacy_picture_description_maps_to_description(tmp_path):
     """`picture_description.enabled=true` (with or without the image flag)
     maps to `processing.pictures: description`."""
     config_file = _write(
@@ -251,15 +263,21 @@ processing:
       timeout: 120
 """,
     )
-    with caplog.at_level("WARNING", logger="haiku.rag.config.loader"):
+    handler = _ListHandler()
+    loader_logger.addHandler(handler)
+    try:
         data = load_yaml_config(config_file)
+    finally:
+        loader_logger.removeHandler(handler)
     cfg = AppConfig.model_validate(data)
     assert cfg.processing.pictures == "description"
     assert cfg.processing.conversion_options.picture_description.timeout == 120
-    assert any("picture_description.enabled=true" in m.message for m in caplog.records)
+    assert any(
+        "picture_description.enabled=true" in r.getMessage() for r in handler.records
+    )
 
 
-def test_load_yaml_legacy_generate_picture_images_maps_to_image(tmp_path, caplog):
+def test_load_yaml_legacy_generate_picture_images_maps_to_image(tmp_path):
     """`generate_picture_images=true` alone maps to `pictures: image`."""
     config_file = _write(
         tmp_path,
@@ -269,14 +287,20 @@ processing:
     generate_picture_images: true
 """,
     )
-    with caplog.at_level("WARNING", logger="haiku.rag.config.loader"):
+    handler = _ListHandler()
+    loader_logger.addHandler(handler)
+    try:
         data = load_yaml_config(config_file)
+    finally:
+        loader_logger.removeHandler(handler)
     cfg = AppConfig.model_validate(data)
     assert cfg.processing.pictures == "image"
-    assert any("generate_picture_images=true" in m.message for m in caplog.records)
+    assert any(
+        "generate_picture_images=true" in r.getMessage() for r in handler.records
+    )
 
 
-def test_load_yaml_no_legacy_fields_keeps_default_none(tmp_path, caplog):
+def test_load_yaml_no_legacy_fields_keeps_default_none(tmp_path):
     """Empty processing block leaves the default `none` mode untouched and
     does not warn."""
     config_file = _write(
@@ -286,11 +310,15 @@ processing:
   chunk_size: 256
 """,
     )
-    with caplog.at_level("WARNING", logger="haiku.rag.config.loader"):
+    handler = _ListHandler()
+    loader_logger.addHandler(handler)
+    try:
         data = load_yaml_config(config_file)
+    finally:
+        loader_logger.removeHandler(handler)
     cfg = AppConfig.model_validate(data)
     assert cfg.processing.pictures == "none"
-    assert not caplog.records
+    assert not handler.records
 
 
 def test_load_yaml_explicit_pictures_wins_over_legacy(tmp_path):

tests/test_embedder_config.py

@@ -75,3 +75,39 @@ def test_unsupported_provider_raises():
 
     with pytest.raises(ValueError, match="Unsupported embedding provider"):
         get_embedder(custom_config)
+
+
+def test_ollama_embedder_appends_v1_when_missing():
+    """Per-model base_url without /v1 should get it appended for Ollama."""
+    config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            model=EmbeddingModelConfig(
+                provider="ollama",
+                name="qwen3-embedding:4b",
+                vector_dim=2560,
+                base_url="http://my-ollama:11434",
+            ),
+        ),
+    )
+    embedder = get_embedder(config)
+    pa_model = embedder._embedder._model  # type: ignore[union-attr]  # ty: ignore[unresolved-attribute]
+    assert str(pa_model.base_url).rstrip("/").endswith("/v1")  # type: ignore[union-attr]  # ty: ignore[unresolved-attribute]
+
+
+def test_ollama_embedder_does_not_double_append_v1():
+    """If the user already includes /v1 we leave it alone."""
+    config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            model=EmbeddingModelConfig(
+                provider="ollama",
+                name="qwen3-embedding:4b",
+                vector_dim=2560,
+                base_url="http://my-ollama:11434/v1",
+            ),
+        ),
+    )
+    embedder = get_embedder(config)
+    pa_model = embedder._embedder._model  # type: ignore[union-attr]  # ty: ignore[unresolved-attribute]
+    url = str(pa_model.base_url).rstrip("/")  # type: ignore[union-attr]  # ty: ignore[unresolved-attribute]
+    assert url.endswith("/v1")
+    assert not url.endswith("/v1/v1")

tests/test_picture_in_context.py

@@ -12,7 +12,7 @@ from pydantic_ai.usage import RunUsage
 
 from haiku.rag.client import HaikuRAG
 from haiku.rag.client.search import _populate_image_data
-from haiku.rag.config import Config
+from haiku.rag.config import AppConfig, Config
 from haiku.rag.context import expand_with_items
 from haiku.rag.store.models.chunk import SearchResult
 from haiku.rag.store.models.document_item import DocumentItem
@@ -167,6 +167,7 @@ async def test_expand_context_preserves_picture_refs_with_empty_text(temp_db_pat
         assert "picture" in out.labels
 
 
+@pytest.mark.vcr()
 @pytest.mark.asyncio
 async def test_rechunk_preserves_picture_data(temp_db_path):
     """``rebuild --rechunk`` keeps ``picture_data`` for every picture row."""
@@ -177,8 +178,10 @@ async def test_rechunk_preserves_picture_data(temp_db_path):
 
     docling_doc = _docling_doc_with_picture()
 
-    async with HaikuRAG(temp_db_path, create=True) as rag:
-        rag._config.processing.pictures = "image"
+    config = AppConfig()
+    config.processing.pictures = "image"
+
+    async with HaikuRAG(temp_db_path, config=config, create=True) as rag:
         document = Document(content="x", uri="test://doc")
         document.set_docling(docling_doc)
         created = await _store_document_with_chunks(rag, document, [], docling_doc)
@@ -193,6 +196,7 @@ async def test_rechunk_preserves_picture_data(temp_db_path):
         assert after.get("#/pictures/0") == before.get("#/pictures/0")
 
 
+@pytest.mark.vcr()
 @pytest.mark.asyncio
 async def test_update_clears_picture_data_when_mode_none(temp_db_path):
     """Switching to ``pictures="none"`` and re-running update_document
@@ -207,9 +211,11 @@ async def test_update_clears_picture_data_when_mode_none(temp_db_path):
 
     docling_doc = _docling_doc_with_picture()
 
-    async with HaikuRAG(temp_db_path, create=True) as rag:
+    config = AppConfig()
+    config.processing.pictures = "image"
+
+    async with HaikuRAG(temp_db_path, config=config, create=True) as rag:
         # Ingest under "image" so picture bytes land in document_items.
-        rag._config.processing.pictures = "image"
         document = Document(content="x", uri="test://doc")
         document.set_docling(docling_doc)
         created = await _store_document_with_chunks(rag, document, [], docling_doc)

tests/test_utils.py

@@ -162,6 +162,27 @@ def test_get_model_ollama_with_settings():
     assert isinstance(result, OpenAIChatModel)
 
 
+def test_get_model_ollama_appends_v1_to_per_model_base_url():
+    """Per-model base_url without /v1 should get it appended."""
+    model_config = ModelConfig(
+        provider="ollama", name="qwen3.6", base_url="http://my-ollama:11434"
+    )
+    result = get_model(model_config)
+    assert isinstance(result, OpenAIChatModel)
+    assert str(result.client.base_url).rstrip("/").endswith("/v1")
+
+
+def test_get_model_ollama_does_not_double_append_v1():
+    """If the per-model base_url already ends with /v1, leave it alone."""
+    model_config = ModelConfig(
+        provider="ollama", name="qwen3.6", base_url="http://my-ollama:11434/v1"
+    )
+    result = get_model(model_config)
+    url = str(result.client.base_url).rstrip("/")
+    assert url.endswith("/v1")
+    assert not url.endswith("/v1/v1")
+
+
 def test_get_model_openai():
     """Test get_model returns OpenAIChatModel for OpenAI."""
     model_config = ModelConfig(provider="openai", name="gpt-4o")