agent-coderag 1.3.2__tar.gz → 1.3.4__tar.gz

{agent_coderag-1.3.2/agent_coderag.egg-info → agent_coderag-1.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-coderag
-Version: 1.3.2
+Version: 1.3.4
 Summary: Lightweight semantic code search and distillation utility for AI coding agents. It solves the API knowledge gap via real-time local signature extraction and intent analysis without PyTorch. Optimized for token efficiency, it compresses codebase context into compact semantic summaries stored in a local DuckDB vector similarity index.
 Author-email: Igor Boloban <naranor@gmail.com>
 License: MIT
@@ -29,6 +29,7 @@ Requires-Dist: onnxruntime
 Requires-Dist: tokenizers
 Requires-Dist: pydantic
 Requires-Dist: httpx
+Requires-Dist: requests
 Requires-Dist: aiofiles
 Requires-Dist: pathspec
 Requires-Dist: tree-sitter
@@ -118,6 +119,18 @@ If you don't configure an LLM provider, agent-coderag works in **100% Offline Mo
 - **Search:** Remains fast and accurate.
 - **Distillation:** Instead of AI-generated summaries, the system uses code signatures and entity names as fallback metadata. No data ever leaves your machine.
 
+Remote embeddings are **not** 100% Offline Mode. If you set `embedding_base` + `embedding_model`, `sync` / `search` / `rebuild` need the network. Embedder choice is process-global (`config.json`); after a remote model or dimension change, run `agent-coderag rebuild` (or delete that `--db`) for **each** project index.
+
+```bash
+agent-coderag config \
+  --embedding-url "http://localhost:8081/v1" \
+  --embedding-model "text-embedding-3-small" \
+  --embedding-key "your-api-key" \
+  --embedding-provider "openai"
+
+agent-coderag config --clear-embedding
+```
+
 ### First Sync & Search
 ```bash
 # Index your entire project (respects .gitignore automatically)
@@ -148,6 +161,21 @@ agent-coderag api fmt
 
 ---
 
+## Library Usage
+
+```python
+from code_rag import CodeRAG
+
+async def main():
+    async with CodeRAG(db="code_rag.db") as rag:
+        await rag.setup()
+        await rag.sync(index_all=True)
+        hits = await rag.search("authentication middleware", limit=5)
+        report = await rag.api("pydantic", lang="python")
+```
+
+---
+
 ## Supported Ecosystems (Discovery)
 
 | Language | Method | Discovery Source |

{agent_coderag-1.3.2 → agent_coderag-1.3.4}/README.md

@@ -73,6 +73,18 @@ If you don't configure an LLM provider, agent-coderag works in **100% Offline Mo
 - **Search:** Remains fast and accurate.
 - **Distillation:** Instead of AI-generated summaries, the system uses code signatures and entity names as fallback metadata. No data ever leaves your machine.
 
+Remote embeddings are **not** 100% Offline Mode. If you set `embedding_base` + `embedding_model`, `sync` / `search` / `rebuild` need the network. Embedder choice is process-global (`config.json`); after a remote model or dimension change, run `agent-coderag rebuild` (or delete that `--db`) for **each** project index.
+
+```bash
+agent-coderag config \
+  --embedding-url "http://localhost:8081/v1" \
+  --embedding-model "text-embedding-3-small" \
+  --embedding-key "your-api-key" \
+  --embedding-provider "openai"
+
+agent-coderag config --clear-embedding
+```
+
 ### First Sync & Search
 ```bash
 # Index your entire project (respects .gitignore automatically)
@@ -103,6 +115,21 @@ agent-coderag api fmt
 
 ---
 
+## Library Usage
+
+```python
+from code_rag import CodeRAG
+
+async def main():
+    async with CodeRAG(db="code_rag.db") as rag:
+        await rag.setup()
+        await rag.sync(index_all=True)
+        hits = await rag.search("authentication middleware", limit=5)
+        report = await rag.api("pydantic", lang="python")
+```
+
+---
+
 ## Supported Ecosystems (Discovery)
 
 | Language | Method | Discovery Source |

{agent_coderag-1.3.2 → agent_coderag-1.3.4/agent_coderag.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-coderag
-Version: 1.3.2
+Version: 1.3.4
 Summary: Lightweight semantic code search and distillation utility for AI coding agents. It solves the API knowledge gap via real-time local signature extraction and intent analysis without PyTorch. Optimized for token efficiency, it compresses codebase context into compact semantic summaries stored in a local DuckDB vector similarity index.
 Author-email: Igor Boloban <naranor@gmail.com>
 License: MIT
@@ -29,6 +29,7 @@ Requires-Dist: onnxruntime
 Requires-Dist: tokenizers
 Requires-Dist: pydantic
 Requires-Dist: httpx
+Requires-Dist: requests
 Requires-Dist: aiofiles
 Requires-Dist: pathspec
 Requires-Dist: tree-sitter
@@ -118,6 +119,18 @@ If you don't configure an LLM provider, agent-coderag works in **100% Offline Mo
 - **Search:** Remains fast and accurate.
 - **Distillation:** Instead of AI-generated summaries, the system uses code signatures and entity names as fallback metadata. No data ever leaves your machine.
 
+Remote embeddings are **not** 100% Offline Mode. If you set `embedding_base` + `embedding_model`, `sync` / `search` / `rebuild` need the network. Embedder choice is process-global (`config.json`); after a remote model or dimension change, run `agent-coderag rebuild` (or delete that `--db`) for **each** project index.
+
+```bash
+agent-coderag config \
+  --embedding-url "http://localhost:8081/v1" \
+  --embedding-model "text-embedding-3-small" \
+  --embedding-key "your-api-key" \
+  --embedding-provider "openai"
+
+agent-coderag config --clear-embedding
+```
+
 ### First Sync & Search
 ```bash
 # Index your entire project (respects .gitignore automatically)
@@ -148,6 +161,21 @@ agent-coderag api fmt
 
 ---
 
+## Library Usage
+
+```python
+from code_rag import CodeRAG
+
+async def main():
+    async with CodeRAG(db="code_rag.db") as rag:
+        await rag.setup()
+        await rag.sync(index_all=True)
+        hits = await rag.search("authentication middleware", limit=5)
+        report = await rag.api("pydantic", lang="python")
+```
+
+---
+
 ## Supported Ecosystems (Discovery)
 
 | Language | Method | Discovery Source |

{agent_coderag-1.3.2 → agent_coderag-1.3.4}/agent_coderag.egg-info/SOURCES.txt

@@ -9,6 +9,9 @@ agent_coderag.egg-info/entry_points.txt
 agent_coderag.egg-info/requires.txt
 agent_coderag.egg-info/top_level.txt
 code_rag/__init__.py
+code_rag/api/__init__.py
+code_rag/api/client.py
+code_rag/api/models.py
 code_rag/core/__init__.py
 code_rag/core/constants.py
 code_rag/core/exceptions.py
@@ -33,14 +36,28 @@ code_rag/entry/cli.py
 code_rag/intelligence/__init__.py
 code_rag/intelligence/distiller.py
 code_rag/intelligence/embedder.py
+code_rag/intelligence/factory.py
+code_rag/intelligence/openai_embedder.py
 code_rag/parsers/__init__.py
 code_rag/parsers/languages.py
 code_rag/parsers/multi_parser.py
 code_rag/parsers/tree_sitter.py
+code_rag/services/__init__.py
+code_rag/services/config.py
+code_rag/services/discovery_api.py
+code_rag/services/factory.py
+code_rag/services/search.py
+code_rag/services/setup.py
+code_rag/services/sync.py
 code_rag/storage/__init__.py
 code_rag/storage/duckdb_impl.py
+tests/test_api_client.py
+tests/test_api_models.py
+tests/test_api_rebuild_wipe.py
 tests/test_cli.py
 tests/test_cli_detailed.py
+tests/test_cli_embeddings.py
+tests/test_cli_json_parity.py
 tests/test_code_rag_simple.py
 tests/test_csharp_discovery_detailed.py
 tests/test_dependency_discovery.py
@@ -48,17 +65,29 @@ tests/test_discovery_manager_detailed.py
 tests/test_discovery_providers_extra.py
 tests/test_discovery_python.py
 tests/test_distiller.py
+tests/test_distiller_config_embedding.py
 tests/test_distiller_extra.py
 tests/test_embedder.py
 tests/test_embedder_detailed.py
+tests/test_embedder_factory.py
+tests/test_embedder_interface.py
+tests/test_embedder_stubs.py
+tests/test_factory_async.py
 tests/test_interfaces.py
 tests/test_js_discovery_detailed.py
 tests/test_languages.py
+tests/test_local_onnx_embedder.py
 tests/test_manager_detailed.py
+tests/test_manager_embeddings.py
 tests/test_manager_extra.py
 tests/test_models.py
 tests/test_multi_parser.py
+tests/test_openai_embedder.py
+tests/test_readme_library_usage.py
+tests/test_readme_offline_embeddings.py
 tests/test_rust_discovery_detailed.py
 tests/test_storage_detailed.py
+tests/test_storage_embeddings.py
+tests/test_sync_worker_errors.py
 tests/test_tree_sitter_parser.py
 tests/test_utils_detailed.py

{agent_coderag-1.3.2 → agent_coderag-1.3.4}/agent_coderag.egg-info/requires.txt

@@ -5,6 +5,7 @@ onnxruntime
 tokenizers
 pydantic
 httpx
+requests
 aiofiles
 pathspec
 tree-sitter

agent_coderag-1.3.4/code_rag/__init__.py

@@ -0,0 +1,16 @@
+from code_rag.api.client import CodeRAG
+from code_rag.api.models import SyncResult, ApiReport, SetupResult
+from code_rag.core.models import KnowledgeUnit, UnitKind
+from code_rag.intelligence.distiller import DistillerConfig
+from code_rag.core.exceptions import CodeRAGError
+
+__all__ = [
+    "CodeRAG",
+    "SyncResult",
+    "ApiReport",
+    "SetupResult",
+    "KnowledgeUnit",
+    "UnitKind",
+    "DistillerConfig",
+    "CodeRAGError",
+]

agent_coderag-1.3.4/code_rag/api/__init__.py

@@ -0,0 +1,6 @@
+"""Public API package for CodeRAG.
+
+`CodeRAG` lives in `code_rag.api.client` and is re-exported from top-level
+`code_rag`. Import models from `code_rag.api.models` directly; this package
+init stays empty to avoid circular imports with `code_rag.services.*`.
+"""

agent_coderag-1.3.4/code_rag/api/client.py

@@ -0,0 +1,110 @@
+from pathlib import Path
+from typing import Optional, Union
+
+from code_rag.api.models import ApiReport, SetupResult, SyncResult
+from code_rag.core.manager import CodeRAGManager
+from code_rag.core.models import KnowledgeUnit
+from code_rag.intelligence.distiller import DistillerConfig
+from code_rag.services.config import load_or_update_config
+from code_rag.services.discovery_api import run_api
+from code_rag.services.factory import create_manager
+from code_rag.services.search import run_search
+from code_rag.services.setup import run_setup
+from code_rag.services.sync import run_rebuild, run_sync
+
+
+class CodeRAG:
+    """Public async facade over CodeRAG services and the storage/search manager."""
+
+    def __init__(
+        self,
+        db: str = "code_rag.db",
+        onnx: Optional[str] = None,
+        *,
+        root: Optional[Union[str, Path]] = None,
+        allow_build_execution: bool = False,
+    ):
+        self._db = db
+        self._onnx = onnx
+        self._root = Path(root) if root is not None else Path.cwd()
+        self._allow_build_execution = allow_build_execution
+        self._manager: Optional[CodeRAGManager] = None
+
+    async def __aenter__(self) -> "CodeRAG":
+        return self
+
+    async def __aexit__(self, *_exc) -> None:
+        await self.close()
+
+    async def close(self) -> None:
+        if self._manager is not None:
+            await self._manager.close()
+            self._manager = None
+
+    async def _ensure_manager(self, wipe: bool = False):
+        if wipe and self._manager is not None:
+            await self.close()
+        if self._manager is None:
+            self._manager = await create_manager(
+                self._db,
+                self._onnx,
+                allow_build_execution=self._allow_build_execution,
+                wipe=wipe,
+            )
+        return self._manager
+
+    async def config(  # pylint: disable=too-many-arguments
+        self,
+        *,
+        url: Optional[str] = None,
+        key: Optional[str] = None,
+        model: Optional[str] = None,
+        provider: Optional[str] = None,
+        embedding_url: Optional[str] = None,
+        embedding_key: Optional[str] = None,
+        embedding_model: Optional[str] = None,
+        embedding_provider: Optional[str] = None,
+        clear_embedding: bool = False,
+    ) -> DistillerConfig:
+        return load_or_update_config(
+            url=url,
+            key=key,
+            model=model,
+            provider=provider,
+            embedding_url=embedding_url,
+            embedding_key=embedding_key,
+            embedding_model=embedding_model,
+            embedding_provider=embedding_provider,
+            clear_embedding=clear_embedding,
+        )
+
+    async def setup(self, *, force: bool = False) -> SetupResult:
+        return await run_setup(force=force)
+
+    async def sync(
+        self,
+        path: Optional[str] = None,
+        *,
+        index_all: bool = False,
+        force: bool = False,
+    ) -> SyncResult:
+        if path is None and not index_all:
+            return SyncResult(status="success", indexed_files=0)
+        return await run_sync(
+            await self._ensure_manager(),
+            root=self._root,
+            path=path,
+            index_all=index_all,
+            force=force,
+        )
+
+    async def search(self, query: str, *, limit: int = 5) -> list[KnowledgeUnit]:
+        return await run_search(await self._ensure_manager(), query, limit=limit)
+
+    async def api(self, library: str, *, lang: Optional[str] = None) -> ApiReport:
+        return await run_api(await self._ensure_manager(), library, lang=lang)
+
+    async def rebuild(self) -> SyncResult:
+        await self.close()
+        manager = await self._ensure_manager(wipe=True)
+        return await run_rebuild(manager, root=self._root)

agent_coderag-1.3.4/code_rag/api/models.py

@@ -0,0 +1,26 @@
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class SyncFileError(BaseModel):
+    file: str
+    message: str
+
+
+class SyncResult(BaseModel):
+    status: str
+    indexed_files: int
+    errors: list[SyncFileError] = Field(default_factory=list)
+
+
+class ApiReport(BaseModel):
+    library: str
+    language: str
+    report: str
+
+
+class SetupResult(BaseModel):
+    model_config = ConfigDict(protected_namespaces=())
+
+    model_dir: str
+    downloaded: list[str]
+    skipped: list[str]

{agent_coderag-1.3.2 → agent_coderag-1.3.4}/code_rag/core/constants.py

@@ -4,6 +4,10 @@ Central location for project-wide constants.
 
 # Embedding Model Constants
 EMBEDDING_DIM = 384
+EMBEDDING_BATCH_SIZE = 32
+EMBEDDING_REQUEST_TIMEOUT = 30
+EMBEDDING_PROBE_TEXT = "probe"
+LOCAL_EMBEDDING_MODEL_ID = "local:mini-lm"
 MAX_TOKEN_LENGTH = 512
 PAD_ID = 0
 PAD_TOKEN = "[PAD]"  # nosec B105

agent_coderag-1.3.4/code_rag/core/interfaces.py

@@ -0,0 +1,113 @@
+from abc import ABC, abstractmethod
+from typing import List, Optional
+from .models import KnowledgeUnit, Relation
+
+
+class IParser(ABC):
+    """Interface for extracting structure from code."""
+
+    @abstractmethod
+    async def distill_file(self, file_path: str) -> List[KnowledgeUnit]:
+        pass  # pragma: no cover
+
+
+class IEmbedder(ABC):
+    """Port for local ONNX and remote OpenAI-compatible embedders."""
+
+    @property
+    @abstractmethod
+    def dimension(self) -> int:
+        """Bound embedding width. No IO. Raise if unbound."""  # pragma: no cover
+
+    @property
+    @abstractmethod
+    def model_id(self) -> str:
+        """Stable id written to index_meta (local:mini-lm or remote model)."""  # pragma: no cover
+
+    @abstractmethod
+    def bind_dimension(self, dim: int) -> None:
+        """Record the index dimension before search/upsert."""  # pragma: no cover
+
+    @abstractmethod
+    async def aembed(self, texts: list[str]) -> list[list[float]]:
+        """Return len(texts) L2-normalized rows, 1:1 with input order."""  # pragma: no cover
+
+    @abstractmethod
+    async def close(self) -> None:
+        """Release ONNX session and/or HTTP resources."""  # pragma: no cover
+
+
+class IStorage(ABC):
+    """Interface for storing the index."""
+
+    @abstractmethod
+    async def upsert_unit(
+        self, unit: KnowledgeUnit, vector: Optional[List[float]] = None
+    ):
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def has_embedding(self, unit_id: str) -> bool:
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def list_units(self) -> List[KnowledgeUnit]:
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def mark_embedding_model_synced(self) -> None:
+        pass  # pragma: no cover
+
+    @property
+    @abstractmethod
+    def embedding_model_dirty(self) -> bool:
+        pass  # pragma: no cover
+
+    @property
+    @abstractmethod
+    def embedder(self) -> IEmbedder:
+        """Bound embedder used for vector search and upsert."""  # pragma: no cover
+
+    @abstractmethod
+    async def get_unit(self, unit_id: str) -> Optional[KnowledgeUnit]:
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def search_units(self, query: str, limit: int = 5) -> List[KnowledgeUnit]:
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def upsert_relation(self, relation: Relation):
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def get_relations(
+        self, unit_id: str, direction: str = "out"
+    ) -> List[Relation]:
+        pass  # pragma: no cover
+
+    @abstractmethod
+    async def set_dependency_path(self, lib_name: str, path: str) -> None:
+        """Caches the absolute path to a library's JAR/binary."""
+
+    @abstractmethod
+    async def get_dependency_path(self, lib_name: str) -> Optional[str]:
+        """Retrieves the cached path for a library."""
+
+    @abstractmethod
+    async def delete_stale_units(
+        self, file_path: str, current_unit_ids: List[str]
+    ) -> None:
+        """Removes units that are no longer present in the given file."""
+
+    @abstractmethod
+    async def close(self) -> None:
+        """Closes the storage connection and releases resources."""
+
+
+class IIntelligence(ABC):
+    """Interface for LLM-based analysis (distillation, embeddings)."""
+
+    @abstractmethod
+    async def summarize(self, code: str, unit_name: str) -> str:
+        pass  # pragma: no cover