haiku.rag 0.13.1__tar.gz → 0.13.3__tar.gz

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haiku.rag
-Version: 0.13.1
+Version: 0.13.3
 Summary: Agentic Retrieval Augmented Generation (RAG) with LanceDB
 Author-email: Yiorgis Gozadinos <ggozadinos@gmail.com>
 License: MIT
@@ -17,25 +17,30 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Typing :: Typed
 Requires-Python: >=3.12
-Requires-Dist: docling>=2.56.1
-Requires-Dist: fastmcp>=2.12.4
+Requires-Dist: docling>=2.58.0
+Requires-Dist: fastmcp>=2.13.0.2
 Requires-Dist: httpx>=0.28.1
 Requires-Dist: lancedb>=0.25.2
-Requires-Dist: pydantic-ai>=1.0.18
-Requires-Dist: pydantic-graph>=1.0.18
-Requires-Dist: pydantic>=2.12.2
-Requires-Dist: python-dotenv>=1.1.1
-Requires-Dist: pyyaml>=6.0.1
+Requires-Dist: pathspec>=0.12.1
+Requires-Dist: pydantic-ai>=1.7.0
+Requires-Dist: pydantic-graph>=1.7.0
+Requires-Dist: pydantic>=2.12.3
+Requires-Dist: python-dotenv>=1.2.1
+Requires-Dist: pyyaml>=6.0.3
 Requires-Dist: rich>=14.2.0
 Requires-Dist: tiktoken>=0.12.0
-Requires-Dist: typer>=0.19.2
-Requires-Dist: watchfiles>=1.1.0
+Requires-Dist: typer<0.20.0,>=0.19.2
+Requires-Dist: watchfiles>=1.1.1
 Provides-Extra: a2a
 Requires-Dist: fasta2a>=0.1.0; extra == 'a2a'
+Provides-Extra: cohere
+Requires-Dist: cohere>=5.0.0; extra == 'cohere'
 Provides-Extra: mxbai
 Requires-Dist: mxbai-rerank>=0.1.6; extra == 'mxbai'
 Provides-Extra: voyageai
 Requires-Dist: voyageai>=0.3.5; extra == 'voyageai'
+Provides-Extra: zeroentropy
+Requires-Dist: zeroentropy>=0.1.0a6; extra == 'zeroentropy'
 Description-Content-Type: text/markdown
 
 # Haiku RAG
@@ -55,7 +60,7 @@ Retrieval-Augmented Generation (RAG) library built on LanceDB.
 - **Multiple QA providers**: Any provider/model supported by Pydantic AI
 - **Research graph (multi‑agent)**: Plan → Search → Evaluate → Synthesize with agentic AI
 - **Native hybrid search**: Vector + full-text search with native LanceDB RRF reranking
-- **Reranking**: Default search result reranking with MixedBread AI, Cohere, or vLLM
+- **Reranking**: Default search result reranking with MixedBread AI, Cohere, Zero Entropy, or vLLM
 - **Question answering**: Built-in QA agents on your documents
 - **File monitoring**: Auto-index files when run as server
 - **40+ file formats**: PDF, DOCX, HTML, Markdown, code files, URLs
@@ -78,6 +83,9 @@ haiku-rag add-src document.pdf --meta source=manual
 # Search
 haiku-rag search "query"
 
+# Search with filters
+haiku-rag search "query" --filter "uri LIKE '%.pdf' AND title LIKE '%paper%'"
+
 # Ask questions
 haiku-rag ask "Who is the author of haiku.rag?"
 

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/README.md

@@ -15,7 +15,7 @@ Retrieval-Augmented Generation (RAG) library built on LanceDB.
 - **Multiple QA providers**: Any provider/model supported by Pydantic AI
 - **Research graph (multi‑agent)**: Plan → Search → Evaluate → Synthesize with agentic AI
 - **Native hybrid search**: Vector + full-text search with native LanceDB RRF reranking
-- **Reranking**: Default search result reranking with MixedBread AI, Cohere, or vLLM
+- **Reranking**: Default search result reranking with MixedBread AI, Cohere, Zero Entropy, or vLLM
 - **Question answering**: Built-in QA agents on your documents
 - **File monitoring**: Auto-index files when run as server
 - **40+ file formats**: PDF, DOCX, HTML, Markdown, code files, URLs
@@ -38,6 +38,9 @@ haiku-rag add-src document.pdf --meta source=manual
 # Search
 haiku-rag search "query"
 
+# Search with filters
+haiku-rag search "query" --filter "uri LIKE '%.pdf' AND title LIKE '%paper%'"
+
 # Ask questions
 haiku-rag ask "Who is the author of haiku.rag?"
 

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/pyproject.toml

@@ -2,7 +2,7 @@
 
 name = "haiku.rag"
 description = "Agentic Retrieval Augmented Generation (RAG) with LanceDB"
-version = "0.13.1"
+version = "0.13.3"
 authors = [{ name = "Yiorgis Gozadinos", email = "ggozadinos@gmail.com" }]
 license = { text = "MIT" }
 readme = { file = "README.md", content-type = "text/markdown" }
@@ -22,25 +22,28 @@ classifiers = [
 ]
 
 dependencies = [
-    "docling>=2.56.1",
-    "fastmcp>=2.12.4",
+    "docling>=2.58.0",
+    "fastmcp>=2.13.0.2",
     "httpx>=0.28.1",
     "lancedb>=0.25.2",
-    "pydantic>=2.12.2",
-    "pydantic-ai>=1.0.18",
-    "pydantic-graph>=1.0.18",
-    "python-dotenv>=1.1.1",
-    "pyyaml>=6.0.1",
+    "pathspec>=0.12.1",
+    "pydantic>=2.12.3",
+    "pydantic-ai>=1.7.0",
+    "pydantic-graph>=1.7.0",
+    "python-dotenv>=1.2.1",
+    "pyyaml>=6.0.3",
     "rich>=14.2.0",
     "tiktoken>=0.12.0",
-    "typer>=0.19.2",
-    "watchfiles>=1.1.0",
+    "typer>=0.19.2,<0.20.0",
+    "watchfiles>=1.1.1",
 ]
 
 [project.optional-dependencies]
 voyageai = ["voyageai>=0.3.5"]
 mxbai = ["mxbai-rerank>=0.1.6"]
 a2a = ["fasta2a>=0.1.0"]
+cohere = ["cohere>=5.0.0"]
+zeroentropy = ["zeroentropy>=0.1.0a6"]
 
 [project.scripts]
 haiku-rag = "haiku.rag.cli:cli"
@@ -57,17 +60,17 @@ packages = ["src/haiku"]
 
 [dependency-groups]
 dev = [
-    "datasets>=4.1.0",
-    "logfire>=4.7.0",
+    "datasets>=4.3.0",
+    "logfire>=4.14.2",
     "mkdocs>=1.6.1",
-    "mkdocs-material>=9.6.14",
-    "pydantic-evals>=1.0.8",
-    "pre-commit>=4.2.0",
-    "pyright>=1.1.406",
+    "mkdocs-material>=9.6.22",
+    "pydantic-evals>=1.7.0",
+    "pre-commit>=4.3.0",
+    "pyright>=1.1.407",
     "pytest>=8.4.2",
     "pytest-asyncio>=1.2.0",
     "pytest-cov>=7.0.0",
-    "ruff>=0.13.0",
+    "ruff>=0.14.2",
 ]
 
 [tool.ruff]

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/evaluations/benchmark.py

@@ -1,5 +1,6 @@
 import asyncio
 from collections.abc import Mapping
+from pathlib import Path
 from typing import Any, cast
 
 import logfire
@@ -12,13 +13,12 @@ from pydantic_evals.reporting import ReportCaseFailure
 from rich.console import Console
 from rich.progress import Progress
 
-from evaluations.config import DatasetSpec, RetrievalSample
+from evaluations.config import DatasetSpec
 from evaluations.datasets import DATASETS
 from evaluations.llm_judge import ANSWER_EQUIVALENCE_RUBRIC
 from evaluations.prompts import WIX_SUPPORT_PROMPT
-from haiku.rag import logging  # noqa: F401
 from haiku.rag.client import HaikuRAG
-from haiku.rag.config import Config
+from haiku.rag.config import AppConfig, find_config_file, load_yaml_config
 from haiku.rag.logging import configure_cli_logging
 from haiku.rag.qa import get_qa_agent
 
@@ -30,7 +30,7 @@ configure_cli_logging()
 console = Console()
 
 
-async def populate_db(spec: DatasetSpec) -> None:
+async def populate_db(spec: DatasetSpec, config: AppConfig) -> None:
     spec.db_path.parent.mkdir(parents=True, exist_ok=True)
     corpus = spec.document_loader()
     if spec.document_limit is not None:
@@ -38,7 +38,7 @@ async def populate_db(spec: DatasetSpec) -> None:
 
     with Progress() as progress:
         task = progress.add_task("[green]Populating database...", total=len(corpus))
-        async with HaikuRAG(spec.db_path) as rag:
+        async with HaikuRAG(spec.db_path, config=config) as rag:
             for doc in corpus:
                 doc_mapping = cast(Mapping[str, Any], doc)
                 payload = spec.document_mapper(doc_mapping)
@@ -64,11 +64,9 @@ async def populate_db(spec: DatasetSpec) -> None:
                 progress.advance(task)
 
 
-def _is_relevant_match(retrieved_uri: str | None, sample: RetrievalSample) -> bool:
-    return retrieved_uri is not None and retrieved_uri in sample.expected_uris
-
-
-async def run_retrieval_benchmark(spec: DatasetSpec) -> dict[str, float] | None:
+async def run_retrieval_benchmark(
+    spec: DatasetSpec, config: AppConfig
+) -> dict[str, float] | None:
     if spec.retrieval_loader is None or spec.retrieval_mapper is None:
         console.print("Skipping retrieval benchmark; no retrieval config.")
         return None
@@ -91,7 +89,7 @@ async def run_retrieval_benchmark(spec: DatasetSpec) -> dict[str, float] | None:
         task = progress.add_task(
             "[blue]Running retrieval benchmark...", total=len(corpus)
         )
-        async with HaikuRAG(spec.db_path) as rag:
+        async with HaikuRAG(spec.db_path, config=config) as rag:
             for doc in corpus:
                 doc_mapping = cast(Mapping[str, Any], doc)
                 sample = spec.retrieval_mapper(doc_mapping)
@@ -161,7 +159,7 @@ async def run_retrieval_benchmark(spec: DatasetSpec) -> dict[str, float] | None:
 
 
 async def run_qa_benchmark(
-    spec: DatasetSpec, qa_limit: int | None = None
+    spec: DatasetSpec, config: AppConfig, qa_limit: int | None = None
 ) -> ReportCaseFailure[str, str, dict[str, str]] | None:
     corpus = spec.qa_loader()
     if qa_limit is not None:
@@ -174,7 +172,7 @@ async def run_qa_benchmark(
 
     judge_model = OpenAIChatModel(
         model_name=QA_JUDGE_MODEL,
-        provider=OllamaProvider(base_url=f"{Config.providers.ollama.base_url}/v1"),
+        provider=OllamaProvider(base_url=f"{config.providers.ollama.base_url}/v1"),
     )
 
     evaluation_dataset = EvalDataset[str, str, dict[str, str]](
@@ -204,7 +202,7 @@ async def run_qa_benchmark(
             total=len(evaluation_dataset.cases),
         )
 
-        async with HaikuRAG(spec.db_path) as rag:
+        async with HaikuRAG(spec.db_path, config=config) as rag:
             system_prompt = WIX_SUPPORT_PROMPT if spec.key == "wix" else None
             qa = get_qa_agent(rag, system_prompt=system_prompt)
 
@@ -272,6 +270,7 @@ async def run_qa_benchmark(
 
 async def evaluate_dataset(
     spec: DatasetSpec,
+    config: AppConfig,
     skip_db: bool,
     skip_retrieval: bool,
     skip_qa: bool,
@@ -279,15 +278,15 @@ async def evaluate_dataset(
 ) -> None:
     if not skip_db:
         console.print(f"Using dataset: {spec.key}", style="bold magenta")
-        await populate_db(spec)
+        await populate_db(spec, config)
 
     if not skip_retrieval:
         console.print("Running retrieval benchmarks...", style="bold blue")
-        await run_retrieval_benchmark(spec)
+        await run_retrieval_benchmark(spec, config)
 
     if not skip_qa:
         console.print("\nRunning QA benchmarks...", style="bold yellow")
-        await run_qa_benchmark(spec, qa_limit=qa_limit)
+        await run_qa_benchmark(spec, config, qa_limit=qa_limit)
 
 
 app = typer.Typer(help="Run retrieval and QA benchmarks for configured datasets.")
@@ -296,6 +295,9 @@ app = typer.Typer(help="Run retrieval and QA benchmarks for configured datasets.
 @app.command()
 def run(
     dataset: str = typer.Argument(..., help="Dataset key to evaluate."),
+    config: Path | None = typer.Option(
+        None, "--config", help="Path to haiku.rag YAML config file."
+    ),
     skip_db: bool = typer.Option(
         False, "--skip-db", help="Skip updateing the evaluation db."
     ),
@@ -314,9 +316,28 @@ def run(
             f"Unknown dataset '{dataset}'. Choose from: {valid_datasets}"
         )
 
+    # Load config from file or use defaults
+    if config:
+        if not config.exists():
+            raise typer.BadParameter(f"Config file not found: {config}")
+        console.print(f"Loading config from: {config}", style="dim")
+        yaml_data = load_yaml_config(config)
+        app_config = AppConfig.model_validate(yaml_data)
+    else:
+        # Try to find config file using standard search path
+        config_path = find_config_file(None)
+        if config_path:
+            console.print(f"Loading config from: {config_path}", style="dim")
+            yaml_data = load_yaml_config(config_path)
+            app_config = AppConfig.model_validate(yaml_data)
+        else:
+            console.print("No config file found, using defaults", style="dim")
+            app_config = AppConfig()
+
     asyncio.run(
         evaluate_dataset(
             spec=spec,
+            config=app_config,
             skip_db=skip_db,
             skip_retrieval=skip_retrieval,
             skip_qa=skip_qa,

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/app.py

@@ -81,13 +81,10 @@ class HaikuRAGApp:
                 raw = rows[0].get("settings") or "{}"
                 data = json.loads(raw) if isinstance(raw, str) else (raw or {})
                 stored_version = str(data.get("version", stored_version))
-                embed_provider = data.get("EMBEDDINGS_PROVIDER")
-                embed_model = data.get("EMBEDDINGS_MODEL")
-                vector_dim = (
-                    int(data.get("EMBEDDINGS_VECTOR_DIM"))  # pyright: ignore[reportArgumentType]
-                    if data.get("EMBEDDINGS_VECTOR_DIM") is not None
-                    else None
-                )
+                embeddings = data.get("embeddings", {})
+                embed_provider = embeddings.get("provider")
+                embed_model = embeddings.get("model")
+                vector_dim = embeddings.get("vector_dim")
 
         num_docs = 0
         if "documents" in table_names:
@@ -195,9 +192,9 @@ class HaikuRAGApp:
                     f"[yellow]Document with id {doc_id} not found.[/yellow]"
                 )
 
-    async def search(self, query: str, limit: int = 5):
+    async def search(self, query: str, limit: int = 5, filter: str | None = None):
         async with HaikuRAG(db_path=self.db_path) as self.client:
-            results = await self.client.search(query, limit=limit)
+            results = await self.client.search(query, limit=limit, filter=filter)
             if not results:
                 self.console.print("[yellow]No results found.[/yellow]")
                 return
@@ -474,9 +471,7 @@ class HaikuRAGApp:
 
             # Start file monitor if enabled
             if enable_monitor:
-                monitor = FileWatcher(
-                    paths=Config.storage.monitor_directories, client=client
-                )
+                monitor = FileWatcher(client=client)
                 monitor_task = asyncio.create_task(monitor.observe())
                 tasks.append(monitor_task)
 

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/cli.py

@@ -221,6 +221,12 @@ def search(
         "-l",
         help="Maximum number of results to return",
     ),
+    filter: str | None = typer.Option(
+        None,
+        "--filter",
+        "-f",
+        help="SQL WHERE clause to filter documents (e.g., \"uri LIKE '%arxiv%'\")",
+    ),
     db: Path = typer.Option(
         Config.storage.data_dir / "haiku.rag.lancedb",
         "--db",
@@ -230,7 +236,7 @@ def search(
     from haiku.rag.app import HaikuRAGApp
 
     app = HaikuRAGApp(db_path=db)
-    asyncio.run(app.search(query=query, limit=limit))
+    asyncio.run(app.search(query=query, limit=limit, filter=filter))
 
 
 @cli.command("ask", help="Ask a question using the QA agent")

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/client.py

@@ -135,9 +135,6 @@ class HaikuRAG:
             ValueError: If the file/URL cannot be parsed or doesn't exist
             httpx.RequestError: If URL request fails
         """
-        # Lazy import to avoid loading docling
-        from haiku.rag.reader import FileReader
-
         # Normalize metadata
         metadata = metadata or {}
 
@@ -157,15 +154,17 @@ class HaikuRAG:
 
         # Handle directories
         if source_path.is_dir():
+            from haiku.rag.monitor import FileFilter
+
             documents = []
-            supported_extensions = set(FileReader.extensions)
-            for file_path in source_path.rglob("*"):
-                if (
-                    file_path.is_file()
-                    and file_path.suffix.lower() in supported_extensions
-                ):
+            filter = FileFilter(
+                ignore_patterns=self._config.monitor.ignore_patterns or None,
+                include_patterns=self._config.monitor.include_patterns or None,
+            )
+            for path in source_path.rglob("*"):
+                if path.is_file() and filter.include_file(str(path)):
                     doc = await self._create_document_from_file(
-                        file_path, title=None, metadata=metadata
+                        path, title=None, metadata=metadata
                     )
                     documents.append(doc)
             return documents
@@ -424,7 +423,11 @@ class HaikuRAG:
         return await self.document_repository.list_all(limit=limit, offset=offset)
 
     async def search(
-        self, query: str, limit: int = 5, search_type: str = "hybrid"
+        self,
+        query: str,
+        limit: int = 5,
+        search_type: str = "hybrid",
+        filter: str | None = None,
     ) -> list[tuple[Chunk, float]]:
         """Search for relevant chunks using the specified search method with optional reranking.
 
@@ -432,6 +435,7 @@ class HaikuRAG:
             query: The search query string.
             limit: Maximum number of results to return.
             search_type: Type of search - "vector", "fts", or "hybrid" (default).
+            filter: Optional SQL WHERE clause to filter documents before searching chunks.
 
         Returns:
             List of (chunk, score) tuples ordered by relevance.
@@ -441,12 +445,12 @@ class HaikuRAG:
 
         if reranker is None:
             # No reranking - return direct search results
-            return await self.chunk_repository.search(query, limit, search_type)
+            return await self.chunk_repository.search(query, limit, search_type, filter)
 
         # Get more initial results (3X) for reranking
         search_limit = limit * 3
         search_results = await self.chunk_repository.search(
-            query, search_limit, search_type
+            query, search_limit, search_type, filter
         )
 
         # Apply reranking

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/config/__init__.py

@@ -11,6 +11,7 @@ from haiku.rag.config.models import (
     AppConfig,
     EmbeddingsConfig,
     LanceDBConfig,
+    MonitorConfig,
     OllamaConfig,
     ProcessingConfig,
     ProvidersConfig,
@@ -25,6 +26,7 @@ __all__ = [
     "Config",
     "AppConfig",
     "StorageConfig",
+    "MonitorConfig",
     "LanceDBConfig",
     "EmbeddingsConfig",
     "RerankingConfig",

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/config/loader.py

@@ -10,7 +10,7 @@ def find_config_file(cli_path: Path | None = None) -> Path | None:
     Search order:
     1. CLI-provided path (via HAIKU_RAG_CONFIG_PATH env var or parameter)
     2. ./haiku.rag.yaml (current directory)
-    3. ~/.config/haiku.rag/config.yaml (user config)
+    3. Platform-specific user config directory
 
     Returns None if no config file is found.
     """
@@ -29,8 +29,10 @@ def find_config_file(cli_path: Path | None = None) -> Path | None:
     if cwd_config.exists():
         return cwd_config
 
-    user_config_dir = Path.home() / ".config" / "haiku.rag"
-    user_config = user_config_dir / "config.yaml"
+    # Use same directory as data storage for config
+    from haiku.rag.utils import get_default_data_dir
+
+    user_config = get_default_data_dir() / "config.yaml"
     if user_config.exists():
         return user_config
 
@@ -50,10 +52,14 @@ def generate_default_config() -> dict:
         "environment": "production",
         "storage": {
             "data_dir": "",
-            "monitor_directories": [],
             "disable_autocreate": False,
             "vacuum_retention_seconds": 60,
         },
+        "monitor": {
+            "directories": [],
+            "ignore_patterns": [],
+            "include_patterns": [],
+        },
         "lancedb": {"uri": "", "api_key": "", "region": ""},
         "embeddings": {
             "provider": "ollama",
@@ -88,7 +94,7 @@ def load_config_from_env() -> dict:
     env_mappings = {
         "ENV": "environment",
         "DEFAULT_DATA_DIR": ("storage", "data_dir"),
-        "MONITOR_DIRECTORIES": ("storage", "monitor_directories"),
+        "MONITOR_DIRECTORIES": ("monitor", "directories"),
         "DISABLE_DB_AUTOCREATE": ("storage", "disable_autocreate"),
         "VACUUM_RETENTION_SECONDS": ("storage", "vacuum_retention_seconds"),
         "LANCEDB_URI": ("lancedb", "uri"),

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/config/models.py

@@ -7,11 +7,16 @@ from haiku.rag.utils import get_default_data_dir
 
 class StorageConfig(BaseModel):
     data_dir: Path = Field(default_factory=get_default_data_dir)
-    monitor_directories: list[Path] = []
     disable_autocreate: bool = False
     vacuum_retention_seconds: int = 60
 
 
+class MonitorConfig(BaseModel):
+    directories: list[Path] = []
+    ignore_patterns: list[str] = []
+    include_patterns: list[str] = []
+
+
 class LanceDBConfig(BaseModel):
     uri: str = ""
     api_key: str = ""
@@ -72,6 +77,7 @@ class A2AConfig(BaseModel):
 class AppConfig(BaseModel):
     environment: str = "production"
     storage: StorageConfig = Field(default_factory=StorageConfig)
+    monitor: MonitorConfig = Field(default_factory=MonitorConfig)
     lancedb: LanceDBConfig = Field(default_factory=LanceDBConfig)
     embeddings: EmbeddingsConfig = Field(default_factory=EmbeddingsConfig)
     reranking: RerankingConfig = Field(default_factory=RerankingConfig)

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/monitor.py

@@ -2,9 +2,12 @@ import logging
 from pathlib import Path
 from typing import TYPE_CHECKING
 
+import pathspec
+from pathspec.patterns.gitwildmatch import GitWildMatchPattern
 from watchfiles import Change, DefaultFilter, awatch
 
 from haiku.rag.client import HaikuRAG
+from haiku.rag.config import AppConfig, Config
 from haiku.rag.store.models.document import Document
 
 if TYPE_CHECKING:
@@ -14,25 +17,70 @@ logger = logging.getLogger(__name__)
 
 
 class FileFilter(DefaultFilter):
-    def __init__(self, *, ignore_paths: list[Path] | None = None) -> None:
+    def __init__(
+        self,
+        *,
+        ignore_patterns: list[str] | None = None,
+        include_patterns: list[str] | None = None,
+    ) -> None:
         # Lazy import to avoid loading docling
         from haiku.rag.reader import FileReader
 
         self.extensions = tuple(FileReader.extensions)
-        super().__init__(ignore_paths=ignore_paths)
+        self.ignore_spec = (
+            pathspec.PathSpec.from_lines(GitWildMatchPattern, ignore_patterns)
+            if ignore_patterns
+            else None
+        )
+        self.include_spec = (
+            pathspec.PathSpec.from_lines(GitWildMatchPattern, include_patterns)
+            if include_patterns
+            else None
+        )
+        super().__init__()
 
     def __call__(self, change: Change, path: str) -> bool:
-        return path.endswith(self.extensions) and super().__call__(change, path)
+        if not self.include_file(path):
+            return False
+
+        # Apply default watchfiles filter
+        return super().__call__(change, path)
+
+    def include_file(self, path: str) -> bool:
+        """Check if a file should be included based on filters."""
+        # Check extension filter
+        if not path.endswith(self.extensions):
+            return False
+
+        # Apply include patterns if specified (whitelist mode)
+        if self.include_spec:
+            if not self.include_spec.match_file(path):
+                return False
+
+        # Apply ignore patterns (blacklist mode)
+        if self.ignore_spec:
+            if self.ignore_spec.match_file(path):
+                return False
+
+        return True
 
 
 class FileWatcher:
-    def __init__(self, paths: list[Path], client: HaikuRAG):
-        self.paths = paths
+    def __init__(
+        self,
+        client: HaikuRAG,
+        config: AppConfig = Config,
+    ):
+        self.paths = config.monitor.directories
         self.client = client
+        self.ignore_patterns = config.monitor.ignore_patterns or None
+        self.include_patterns = config.monitor.include_patterns or None
 
     async def observe(self):
         logger.info(f"Watching files in {self.paths}")
-        filter = FileFilter()
+        filter = FileFilter(
+            ignore_patterns=self.ignore_patterns, include_patterns=self.include_patterns
+        )
         await self.refresh()
 
         async for changes in awatch(*self.paths, watch_filter=filter):
@@ -49,10 +97,17 @@ class FileWatcher:
         # Lazy import to avoid loading docling
         from haiku.rag.reader import FileReader
 
+        # Create filter to apply same logic as observe()
+        filter = FileFilter(
+            ignore_patterns=self.ignore_patterns, include_patterns=self.include_patterns
+        )
+
         for path in self.paths:
             for f in Path(path).rglob("**/*"):
                 if f.is_file() and f.suffix in FileReader.extensions:
-                    await self._upsert_document(f)
+                    # Apply pattern filters
+                    if filter(Change.added, str(f)):
+                        await self._upsert_document(f)
 
     async def _upsert_document(self, file: Path) -> Document | None:
         try:

{haiku_rag-0.13.1 → haiku_rag-0.13.3}/src/haiku/rag/reranking/__init__.py

@@ -41,5 +41,23 @@ def get_reranker(config: AppConfig = Config) -> RerankerBase | None:
         except ImportError:
             reranker = None
 
+    elif config.reranking.provider == "vllm":
+        try:
+            from haiku.rag.reranking.vllm import VLLMReranker
+
+            reranker = VLLMReranker(config.reranking.model)
+        except ImportError:
+            reranker = None
+
+    elif config.reranking.provider == "zeroentropy":
+        try:
+            from haiku.rag.reranking.zeroentropy import ZeroEntropyReranker
+
+            # Use configured model or default to zerank-1
+            model = config.reranking.model or "zerank-1"
+            reranker = ZeroEntropyReranker(model)
+        except ImportError:
+            reranker = None
+
     _reranker_cache[config_id] = reranker
     return reranker