knowledge-rag 3.5.0__tar.gz → 3.5.2__tar.gz

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: knowledge-rag
-Version: 3.5.0
+Version: 3.5.2
 Summary: Local RAG System for Claude Code — Hybrid search + Cross-encoder Reranking + 12 MCP Tools. Zero external servers.
 Project-URL: Homepage, https://github.com/lyonzin/knowledge-rag
 Project-URL: Repository, https://github.com/lyonzin/knowledge-rag
@@ -19,7 +19,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Topic :: Text Processing :: Indexing
-Requires-Python: <3.13,>=3.11
+Requires-Python: >=3.11
 Requires-Dist: beautifulsoup4>=4.12.0
 Requires-Dist: chromadb>=1.4.0
 Requires-Dist: fastembed[reranking]>=0.4.0
@@ -40,10 +40,11 @@ Description-Content-Type: text/markdown
 
 <div align="center">
 
-![Version](https://img.shields.io/badge/version-3.5.0-blue.svg)
-![Python](https://img.shields.io/badge/python-3.11%20%7C%203.12-green.svg)
+![Version](https://img.shields.io/badge/version-3.5.2-blue.svg)
+![Python](https://img.shields.io/badge/python-3.11%2B-green.svg)
 ![License](https://img.shields.io/badge/license-MIT-yellow.svg)
 ![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux%20%7C%20macOS-lightgrey.svg)
+![GPU](https://img.shields.io/badge/GPU-NVIDIA%20CUDA-76B900.svg?logo=nvidia)
 [![CI](https://github.com/lyonzin/knowledge-rag/actions/workflows/ci.yml/badge.svg)](https://github.com/lyonzin/knowledge-rag/actions/workflows/ci.yml)
 [![CodeQL](https://github.com/lyonzin/knowledge-rag/actions/workflows/security.yml/badge.svg)](https://github.com/lyonzin/knowledge-rag/actions/workflows/security.yml)
 [![Glama Score](https://glama.ai/mcp/servers/lyonzin/knowledge-rag/badges/score.svg)](https://glama.ai/mcp/servers/lyonzin/knowledge-rag)
@@ -63,35 +64,47 @@ Your documents become instantly searchable inside Claude Code — with reranking
 
 **12 MCP Tools** | **Hybrid Search + Cross-Encoder Reranking** | **Markdown-Aware Chunking** | **100% Local, Zero Cloud**
 
-[What's New](#whats-new-in-v350) | [Supported Formats](#supported-formats) | [Installation](#installation) | [Configuration](#configuration) | [API Reference](#api-reference) | [Architecture](#architecture)
+[What's New](#whats-new-in-v352) | [Supported Formats](#supported-formats) | [Installation](#installation) | [Configuration](#configuration) | [API Reference](#api-reference) | [Architecture](#architecture)
 
 </div>
 
 ---
 
-## What's New in v3.5.0
+## What's New in v3.5.2
 
 ### GPU-Accelerated Embeddings (Optional)
 
-ONNX embeddings can now run on NVIDIA GPUs for **5-10x faster indexing**. Opt-in — CPU remains the default.
+ONNX embeddings can run on NVIDIA GPUs for **5-10x faster indexing**. Opt-in — CPU remains the default.
 
 ```bash
+# NVIDIA GPU (requires CUDA 12.x drivers)
 pip install knowledge-rag[gpu]
+
+# Also install CUDA 12 runtime libraries (if not using CUDA Toolkit 12.x)
+pip install nvidia-cublas-cu12 nvidia-cudnn-cu12 nvidia-cuda-runtime-cu12
 ```
 
 ```yaml
 # config.yaml
 models:
   embedding:
-    gpu: true   # Falls back to CPU if CUDA unavailable
+    gpu: true   # Automatic CPU fallback if CUDA is unavailable
 ```
 
+**How it works:**
+- Sets `CUDAExecutionProvider` as primary, `CPUExecutionProvider` as fallback
+- Auto-discovers CUDA 12 DLLs from pip-installed NVIDIA packages (no manual PATH config)
+- If GPU init fails for any reason, falls back to CPU silently with a `[WARN]` log
+- `gpu: false` (default) forces CPU-only mode — zero CUDA overhead, clean logs
+
 Ideal for large knowledge bases (1000+ documents) where full rebuilds take minutes on CPU. After the initial index, incremental reindexing (`force: true`) takes seconds regardless.
 
 ### Recent Highlights
 
+- **v3.5.2** — CUDA DLL auto-discovery from pip packages, graceful GPU→CPU fallback, explicit CPU provider (no CUDA noise when `gpu: false`), BASE_DIR resolution fix for editable installs
+- **v3.5.1** — Remove Python `<3.13` upper bound — 3.13 and 3.14 now supported
+- **v3.5.0** — Optional GPU acceleration, supported formats table, full README rewrite
 - **v3.4.3** — MCP stdout save/restore fix (v3.4.2 broke JSON-RPC responses)
-- **v3.4.1** — `pip install` auto-detects project dir from venv location, Linux/macOS `install.sh`
 - **v3.4.0** — Persistent model cache, exclude patterns, Jupyter Notebook parser, inotify resilience, MetaTrader support
 
 See [Changelog](#changelog) for full history.
@@ -319,7 +332,7 @@ flowchart LR
 
 ### Prerequisites
 
-- Python 3.11 or 3.12 (**NOT** 3.13+ — onnxruntime incompatibility)
+- Python 3.11+
 - Claude Code CLI
 - ~200MB disk for model cache (auto-downloaded on first run)
 - *Optional:* NVIDIA GPU + CUDA for [accelerated embeddings](#gpu-accelerated-embeddings-optional) (`pip install knowledge-rag[gpu]`)
@@ -950,17 +963,10 @@ knowledge-rag/
 
 ### Python version mismatch
 
-ChromaDB depends on onnxruntime which requires Python 3.11 or 3.12. Python 3.13+ is **NOT** supported.
+Requires Python 3.11 or newer.
 
 ```bash
-# Check version
-python --version
-
-# Windows: use specific version
-py -3.12 -m venv venv
-
-# Linux/macOS: use specific version
-python3.12 -m venv venv
+python --version    # Must be 3.11+
 ```
 
 ### FastEmbed model download fails
@@ -1018,6 +1024,17 @@ With ~200 documents, expect ~300-500MB RAM. The embedding model (~50MB) and rera
 
 ## Changelog
 
+### v3.5.2 (2026-04-16)
+
+- **NEW**: Auto-discovery of CUDA 12 DLLs from pip-installed NVIDIA packages — no manual PATH configuration needed
+- **NEW**: Graceful GPU→CPU fallback with `[WARN]` log when CUDA init fails (missing drivers, wrong version, etc.)
+- **FIX**: Explicit `CPUExecutionProvider` when `gpu: false` — eliminates noisy CUDA probe errors in logs
+- **FIX**: BASE_DIR resolution now correctly prefers directories with `config.yaml` over those with only `config.example.yaml` (fixes editable installs)
+
+### v3.5.1 (2026-04-16)
+
+- **FIX**: Removed Python upper bound constraint (`<3.13` → `>=3.11`). Python 3.13 and 3.14 now supported — onnxruntime ships wheels for both.
+
 ### v3.5.0 (2026-04-16)
 
 - **NEW**: Optional GPU acceleration for ONNX embeddings — `pip install knowledge-rag[gpu]` + `models.embedding.gpu: true` in config. 5-10x faster indexing on NVIDIA GPUs with automatic CPU fallback.

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/README.md

@@ -2,10 +2,11 @@
 
 <div align="center">
 
-![Version](https://img.shields.io/badge/version-3.5.0-blue.svg)
-![Python](https://img.shields.io/badge/python-3.11%20%7C%203.12-green.svg)
+![Version](https://img.shields.io/badge/version-3.5.2-blue.svg)
+![Python](https://img.shields.io/badge/python-3.11%2B-green.svg)
 ![License](https://img.shields.io/badge/license-MIT-yellow.svg)
 ![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20Linux%20%7C%20macOS-lightgrey.svg)
+![GPU](https://img.shields.io/badge/GPU-NVIDIA%20CUDA-76B900.svg?logo=nvidia)
 [![CI](https://github.com/lyonzin/knowledge-rag/actions/workflows/ci.yml/badge.svg)](https://github.com/lyonzin/knowledge-rag/actions/workflows/ci.yml)
 [![CodeQL](https://github.com/lyonzin/knowledge-rag/actions/workflows/security.yml/badge.svg)](https://github.com/lyonzin/knowledge-rag/actions/workflows/security.yml)
 [![Glama Score](https://glama.ai/mcp/servers/lyonzin/knowledge-rag/badges/score.svg)](https://glama.ai/mcp/servers/lyonzin/knowledge-rag)
@@ -25,35 +26,47 @@ Your documents become instantly searchable inside Claude Code — with reranking
 
 **12 MCP Tools** | **Hybrid Search + Cross-Encoder Reranking** | **Markdown-Aware Chunking** | **100% Local, Zero Cloud**
 
-[What's New](#whats-new-in-v350) | [Supported Formats](#supported-formats) | [Installation](#installation) | [Configuration](#configuration) | [API Reference](#api-reference) | [Architecture](#architecture)
+[What's New](#whats-new-in-v352) | [Supported Formats](#supported-formats) | [Installation](#installation) | [Configuration](#configuration) | [API Reference](#api-reference) | [Architecture](#architecture)
 
 </div>
 
 ---
 
-## What's New in v3.5.0
+## What's New in v3.5.2
 
 ### GPU-Accelerated Embeddings (Optional)
 
-ONNX embeddings can now run on NVIDIA GPUs for **5-10x faster indexing**. Opt-in — CPU remains the default.
+ONNX embeddings can run on NVIDIA GPUs for **5-10x faster indexing**. Opt-in — CPU remains the default.
 
 ```bash
+# NVIDIA GPU (requires CUDA 12.x drivers)
 pip install knowledge-rag[gpu]
+
+# Also install CUDA 12 runtime libraries (if not using CUDA Toolkit 12.x)
+pip install nvidia-cublas-cu12 nvidia-cudnn-cu12 nvidia-cuda-runtime-cu12
 ```
 
 ```yaml
 # config.yaml
 models:
   embedding:
-    gpu: true   # Falls back to CPU if CUDA unavailable
+    gpu: true   # Automatic CPU fallback if CUDA is unavailable
 ```
 
+**How it works:**
+- Sets `CUDAExecutionProvider` as primary, `CPUExecutionProvider` as fallback
+- Auto-discovers CUDA 12 DLLs from pip-installed NVIDIA packages (no manual PATH config)
+- If GPU init fails for any reason, falls back to CPU silently with a `[WARN]` log
+- `gpu: false` (default) forces CPU-only mode — zero CUDA overhead, clean logs
+
 Ideal for large knowledge bases (1000+ documents) where full rebuilds take minutes on CPU. After the initial index, incremental reindexing (`force: true`) takes seconds regardless.
 
 ### Recent Highlights
 
+- **v3.5.2** — CUDA DLL auto-discovery from pip packages, graceful GPU→CPU fallback, explicit CPU provider (no CUDA noise when `gpu: false`), BASE_DIR resolution fix for editable installs
+- **v3.5.1** — Remove Python `<3.13` upper bound — 3.13 and 3.14 now supported
+- **v3.5.0** — Optional GPU acceleration, supported formats table, full README rewrite
 - **v3.4.3** — MCP stdout save/restore fix (v3.4.2 broke JSON-RPC responses)
-- **v3.4.1** — `pip install` auto-detects project dir from venv location, Linux/macOS `install.sh`
 - **v3.4.0** — Persistent model cache, exclude patterns, Jupyter Notebook parser, inotify resilience, MetaTrader support
 
 See [Changelog](#changelog) for full history.
@@ -281,7 +294,7 @@ flowchart LR
 
 ### Prerequisites
 
-- Python 3.11 or 3.12 (**NOT** 3.13+ — onnxruntime incompatibility)
+- Python 3.11+
 - Claude Code CLI
 - ~200MB disk for model cache (auto-downloaded on first run)
 - *Optional:* NVIDIA GPU + CUDA for [accelerated embeddings](#gpu-accelerated-embeddings-optional) (`pip install knowledge-rag[gpu]`)
@@ -912,17 +925,10 @@ knowledge-rag/
 
 ### Python version mismatch
 
-ChromaDB depends on onnxruntime which requires Python 3.11 or 3.12. Python 3.13+ is **NOT** supported.
+Requires Python 3.11 or newer.
 
 ```bash
-# Check version
-python --version
-
-# Windows: use specific version
-py -3.12 -m venv venv
-
-# Linux/macOS: use specific version
-python3.12 -m venv venv
+python --version    # Must be 3.11+
 ```
 
 ### FastEmbed model download fails
@@ -980,6 +986,17 @@ With ~200 documents, expect ~300-500MB RAM. The embedding model (~50MB) and rera
 
 ## Changelog
 
+### v3.5.2 (2026-04-16)
+
+- **NEW**: Auto-discovery of CUDA 12 DLLs from pip-installed NVIDIA packages — no manual PATH configuration needed
+- **NEW**: Graceful GPU→CPU fallback with `[WARN]` log when CUDA init fails (missing drivers, wrong version, etc.)
+- **FIX**: Explicit `CPUExecutionProvider` when `gpu: false` — eliminates noisy CUDA probe errors in logs
+- **FIX**: BASE_DIR resolution now correctly prefers directories with `config.yaml` over those with only `config.example.yaml` (fixes editable installs)
+
+### v3.5.1 (2026-04-16)
+
+- **FIX**: Removed Python upper bound constraint (`<3.13` → `>=3.11`). Python 3.13 and 3.14 now supported — onnxruntime ships wheels for both.
+
 ### v3.5.0 (2026-04-16)
 
 - **NEW**: Optional GPU acceleration for ONNX embeddings — `pip install knowledge-rag[gpu]` + `models.embedding.gpu: true` in config. 5-10x faster indexing on NVIDIA GPUs with automatic CPU fallback.

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/mcp_server/__init__.py

@@ -8,7 +8,7 @@ import sys  # noqa: I001
 _original_stdout = sys.stdout
 sys.stdout = sys.stderr
 
-__version__ = "3.5.0"
+__version__ = "3.5.2"
 __author__ = "Ailton Rocha (Lyon.)"
 
 from .config import Config  # noqa: E402

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/mcp_server/config.py

@@ -51,12 +51,17 @@ _venv_dir = _venv_project_dir()
 
 if os.environ.get("KNOWLEDGE_RAG_DIR"):
     BASE_DIR = Path(os.environ["KNOWLEDGE_RAG_DIR"])
+elif _venv_dir is not None and (_venv_dir / "config.yaml").exists():
+    # Prefer venv parent if it has an actual config.yaml (editable installs, PyPI installs)
+    BASE_DIR = _venv_dir
+elif _is_project_root(_source_dir) and (_source_dir / "config.yaml").exists():
+    BASE_DIR = _source_dir
+elif _is_project_root(Path.cwd()):
+    BASE_DIR = Path.cwd()
 elif _is_project_root(_source_dir):
     BASE_DIR = _source_dir
 elif _is_project_root(_venv_dir):
     BASE_DIR = _venv_dir
-elif _is_project_root(Path.cwd()):
-    BASE_DIR = Path.cwd()
 else:
     BASE_DIR = _venv_dir if _venv_dir is not None else Path.cwd()
 

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/mcp_server/server.py

@@ -19,7 +19,7 @@ Features:
     - CRUD operations via MCP tools (add, update, remove docs)
 
 Autor:   Lyon (Ailton Rocha)
-Versao:  3.5.0
+Versao:  3.5.2
 Data:    2026-04-16
 """
 
@@ -138,17 +138,60 @@ class FastEmbedEmbeddings:
     Model: BAAI/bge-small-en-v1.5 (384-dim, MTEB score 62.x)
     """
 
+    @staticmethod
+    def _setup_cuda_dll_paths():
+        """Add NVIDIA CUDA 12 pip package DLL paths to os.environ['PATH'].
+
+        When onnxruntime-gpu is installed alongside nvidia-cublas-cu12 etc.,
+        the DLLs live under site-packages/nvidia/*/bin/ and onnxruntime can't
+        find them unless they're on PATH. This is a no-op if the dirs don't exist.
+        """
+        import os
+        import site
+
+        site_dirs = site.getsitepackages() if hasattr(site, "getsitepackages") else []
+        nvidia_libs = [
+            "nvidia/cublas/bin",
+            "nvidia/cudnn/bin",
+            "nvidia/cuda_runtime/bin",
+            "nvidia/cufft/bin",
+            "nvidia/curand/bin",
+            "nvidia/cusolver/bin",
+            "nvidia/cusparse/bin",
+            "nvidia/nvjitlink/bin",
+            "nvidia/cuda_nvrtc/bin",
+        ]
+        added = []
+        for sp in site_dirs:
+            for lib in nvidia_libs:
+                p = os.path.join(sp, lib)
+                if os.path.isdir(p) and p not in os.environ.get("PATH", ""):
+                    os.environ["PATH"] = p + os.pathsep + os.environ.get("PATH", "")
+                    added.append(lib.split("/")[1])
+        if added:
+            print(f"[INFO] CUDA DLL paths added for: {', '.join(dict.fromkeys(added))}")
+
     def __init__(self, model: str = None):
         self.model_name = model or config.embedding_model
         self._dim = config.embedding_dim
         kwargs = {"model_name": self.model_name, "cache_dir": str(config.models_cache_dir)}
         if config.gpu_acceleration:
+            self._setup_cuda_dll_paths()
             kwargs["providers"] = ["CUDAExecutionProvider", "CPUExecutionProvider"]
             print(f"[INFO] Loading embedding model: {self.model_name} ({self._dim}D) [GPU accelerated]...")
+            try:
+                self._model = TextEmbedding(**kwargs)
+                print("[INFO] Embedding model loaded successfully [GPU]")
+            except (ValueError, RuntimeError) as e:
+                print(f"[WARN] GPU init failed ({e}), falling back to CPU...")
+                kwargs["providers"] = ["CPUExecutionProvider"]
+                self._model = TextEmbedding(**kwargs)
+                print("[INFO] Embedding model loaded successfully [CPU fallback]")
         else:
+            kwargs["providers"] = ["CPUExecutionProvider"]
             print(f"[INFO] Loading embedding model: {self.model_name} ({self._dim}D)...")
-        self._model = TextEmbedding(**kwargs)
-        print("[INFO] Embedding model loaded successfully")
+            self._model = TextEmbedding(**kwargs)
+            print("[INFO] Embedding model loaded successfully")
 
     def __call__(self, input: List[str]) -> List[List[float]]:
         """

{knowledge_rag-3.5.0 → knowledge_rag-3.5.2}/pyproject.toml

@@ -4,11 +4,11 @@ build-backend = "hatchling.build"
 
 [project]
 name = "knowledge-rag"
-version = "3.5.0"
+version = "3.5.2"
 description = "Local RAG System for Claude Code — Hybrid search + Cross-encoder Reranking + 12 MCP Tools. Zero external servers."
 readme = "README.md"
 license = {text = "MIT"}
-requires-python = ">=3.11,<3.13"
+requires-python = ">=3.11"
 authors = [
     {name = "Lyon.", email = "lyonzin@users.noreply.github.com"},
 ]