Merge pull request #120 from ggozad/fix/embedder-config

Pass custom config to embedders if provided.

docs/configuration.md

@@ -129,9 +129,11 @@ This is useful for:
 - Testing with different configurations
 - Applications that need multiple clients with different configurations
 
-## API Keys
+## Environment Variables
 
-API keys are configured through **environment variables**, not in the YAML file.
+API keys and some provider settings are configured through **environment variables**, not in the YAML file.
+
+### API Keys
 
 ```bash
 # OpenAI
@@ -147,6 +149,13 @@ export VOYAGE_API_KEY=your-key-here
 export CO_API_KEY=your-key-here
 ```
 
+### Provider Configuration
+
+```bash
+# Ollama base URL (defaults to http://localhost:11434)
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
 ## File Monitoring
 
 Set directories to monitor for automatic indexing:
@@ -171,6 +180,23 @@ embeddings:
   vector_dim: 1024
 ```
 
+The Ollama base URL can be configured via environment variable or config file:
+
+```bash
+# Via environment variable (recommended)
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
+Or in your config file:
+
+```yaml
+providers:
+  ollama:
+    base_url: http://localhost:11434
+```
+
+If neither is set, it defaults to `http://localhost:11434`.
+
 ### VoyageAI
 
 If you want to use VoyageAI embeddings you will need to install `haiku.rag` with the VoyageAI extras:
@@ -236,7 +262,17 @@ Configure which LLM provider to use for question answering. Any provider and mod
 qa:
   provider: ollama
   model: gpt-oss
+```
 
+The Ollama base URL can be configured via the `OLLAMA_BASE_URL` environment variable, config file, or defaults to `http://localhost:11434`:
+
+```bash
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
+Or in your config file:
+
+```yaml
 providers:
   ollama:
     base_url: http://localhost:11434

src/haiku/rag/config/models.py

@@ -46,7 +46,11 @@ class ProcessingConfig(BaseModel):
 
 
 class OllamaConfig(BaseModel):
-    base_url: str = "http://localhost:11434"
+    base_url: str = Field(
+        default_factory=lambda: __import__("os").environ.get(
+            "OLLAMA_BASE_URL", "http://localhost:11434"
+        )
+    )
 
 
 class VLLMConfig(BaseModel):

src/haiku/rag/embeddings/__init__.py

@@ -15,7 +15,9 @@ def get_embedder(config: AppConfig = Config) -> EmbedderBase:
     """
 
     if config.embeddings.provider == "ollama":
-        return OllamaEmbedder(config.embeddings.model, config.embeddings.vector_dim)
+        return OllamaEmbedder(
+            config.embeddings.model, config.embeddings.vector_dim, config
+        )
 
     if config.embeddings.provider == "voyageai":
         try:
@@ -26,16 +28,22 @@ def get_embedder(config: AppConfig = Config) -> EmbedderBase:
                 "Please install haiku.rag with the 'voyageai' extra: "
                 "uv pip install haiku.rag[voyageai]"
             )
-        return VoyageAIEmbedder(config.embeddings.model, config.embeddings.vector_dim)
+        return VoyageAIEmbedder(
+            config.embeddings.model, config.embeddings.vector_dim, config
+        )
 
     if config.embeddings.provider == "openai":
         from haiku.rag.embeddings.openai import Embedder as OpenAIEmbedder
 
-        return OpenAIEmbedder(config.embeddings.model, config.embeddings.vector_dim)
+        return OpenAIEmbedder(
+            config.embeddings.model, config.embeddings.vector_dim, config
+        )
 
     if config.embeddings.provider == "vllm":
         from haiku.rag.embeddings.vllm import Embedder as VllmEmbedder
 
-        return VllmEmbedder(config.embeddings.model, config.embeddings.vector_dim)
+        return VllmEmbedder(
+            config.embeddings.model, config.embeddings.vector_dim, config
+        )
 
     raise ValueError(f"Unsupported embedding provider: {config.embeddings.provider}")

src/haiku/rag/embeddings/base.py

@@ -1,15 +1,17 @@
 from typing import overload
 
-from haiku.rag.config import Config
+from haiku.rag.config import AppConfig, Config
 
 
 class EmbedderBase:
     _model: str = Config.embeddings.model
     _vector_dim: int = Config.embeddings.vector_dim
+    _config: AppConfig = Config
 
-    def __init__(self, model: str, vector_dim: int):
+    def __init__(self, model: str, vector_dim: int, config: AppConfig = Config):
         self._model = model
         self._vector_dim = vector_dim
+        self._config = config
 
     @overload
     async def embed(self, text: str) -> list[float]: ...

src/haiku/rag/embeddings/ollama.py

@@ -2,7 +2,6 @@ from typing import overload
 
 from openai import AsyncOpenAI
 
-from haiku.rag.config import Config
 from haiku.rag.embeddings.base import EmbedderBase
 
 
@@ -15,7 +14,7 @@ class Embedder(EmbedderBase):
 
     async def embed(self, text: str | list[str]) -> list[float] | list[list[float]]:
         client = AsyncOpenAI(
-            base_url=f"{Config.providers.ollama.base_url}/v1", api_key="dummy"
+            base_url=f"{self._config.providers.ollama.base_url}/v1", api_key="dummy"
         )
         if not text:
             return []

src/haiku/rag/embeddings/vllm.py

@@ -2,7 +2,6 @@ from typing import overload
 
 from openai import AsyncOpenAI
 
-from haiku.rag.config import Config
 from haiku.rag.embeddings.base import EmbedderBase
 
 
@@ -15,7 +14,8 @@ class Embedder(EmbedderBase):
 
     async def embed(self, text: str | list[str]) -> list[float] | list[list[float]]:
         client = AsyncOpenAI(
-            base_url=f"{Config.providers.vllm.embeddings_base_url}/v1", api_key="dummy"
+            base_url=f"{self._config.providers.vllm.embeddings_base_url}/v1",
+            api_key="dummy",
         )
         if not text:
             return []

tests/test_embedder_config.py

@@ -0,0 +1,90 @@
+import pytest
+
+from haiku.rag.config import (
+    AppConfig,
+    EmbeddingsConfig,
+    OllamaConfig,
+    ProvidersConfig,
+    VLLMConfig,
+)
+from haiku.rag.embeddings import get_embedder
+
+
+def test_embedder_uses_config_from_get_embedder():
+    """Test that embedders use the config passed to get_embedder."""
+    custom_config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            provider="ollama",
+            model="custom-model",
+            vector_dim=512,
+        ),
+        providers=ProvidersConfig(
+            ollama=OllamaConfig(base_url="http://custom-ollama:8080"),
+            vllm=VLLMConfig(embeddings_base_url="http://custom-vllm:9000"),
+        ),
+    )
+
+    embedder = get_embedder(custom_config)
+
+    assert embedder._model == "custom-model"
+    assert embedder._vector_dim == 512
+    assert embedder._config.providers.ollama.base_url == "http://custom-ollama:8080"
+
+
+def test_vllm_embedder_uses_config():
+    """Test that vllm embedder uses the config passed to get_embedder."""
+    custom_config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            provider="vllm",
+            model="custom-vllm-model",
+            vector_dim=768,
+        ),
+        providers=ProvidersConfig(
+            vllm=VLLMConfig(embeddings_base_url="http://custom-vllm:9001"),
+        ),
+    )
+
+    embedder = get_embedder(custom_config)
+
+    assert embedder._model == "custom-vllm-model"
+    assert embedder._vector_dim == 768
+    assert (
+        embedder._config.providers.vllm.embeddings_base_url == "http://custom-vllm:9001"
+    )
+
+
+def test_openai_embedder_uses_config():
+    """Test that openai embedder uses the config passed to get_embedder."""
+    custom_config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            provider="openai",
+            model="text-embedding-3-large",
+            vector_dim=3072,
+        ),
+    )
+
+    embedder = get_embedder(custom_config)
+
+    assert embedder._model == "text-embedding-3-large"
+    assert embedder._vector_dim == 3072
+    assert embedder._config == custom_config
+
+
+@pytest.mark.skipif(
+    True, reason="VoyageAI is an optional dependency, may not be installed"
+)
+def test_voyageai_embedder_uses_config():
+    """Test that voyageai embedder uses the config passed to get_embedder."""
+    custom_config = AppConfig(
+        embeddings=EmbeddingsConfig(
+            provider="voyageai",
+            model="voyage-large-2",
+            vector_dim=1536,
+        ),
+    )
+
+    embedder = get_embedder(custom_config)
+
+    assert embedder._model == "voyage-large-2"
+    assert embedder._vector_dim == 1536
+    assert embedder._config == custom_config