build(python): faster copy of cpp files for the python build

Author: Neiko2002

examples/README.md

@@ -28,10 +28,12 @@ We use [`uv`](https://docs.astral.sh/uv/) for fast Python environment and depend
 Navigating into any example directory and running `uv sync` or `uv run` will automatically set up a isolated Python virtual environment, build the local `deglib` C++ bindings, and install all required dependencies:
 
 ```bash
-cd examples/paper_reproduction
+cd examples/knng
 uv sync
 ```
 
 ## Projects
 
-- [`paper_reproduction`](./paper_reproduction/): Reproduce DEG paper search benchmarks (Recall vs QPS) on datasets mentioned in the root `readme.md` (`Audio`, `Enron`, `SIFT1M`, `DEEP1M`, `GloVe-100`).
+- [`knng`](./knng/): k-Nearest Neighbor Graph (k-NNG) construction benchmark using EVP quantization and FP16 reranking (SISAP 2026 Challenge Task 1).
+- [`static_data`](./static_data/): DEG paper search benchmark reproduction (Recall vs. QPS) on static datasets (`sift1m`, `deep1m`, `glove-100`, `audio`, `enron`).
+- [`dynamic_data`](./dynamic_data/): DEG dynamic data streaming benchmark (AddHalf, AddHalfRemoveAndAddOneAtATime, AddAllRemoveHalf).

examples/dynamic_data/README.md

@@ -53,12 +53,19 @@ uv run python main.py sift1m --graph-dir /data/deg_graphs/sift1m/dynamic
 uv run python main.py all --graph-dir /data/deg_graphs --threads 4
 ```
 
-## Setup
+## Prerequisites: Building the Python Library (`deglib`)
+
+The `deglib` Python library contains C++ pybind11 bindings (`deglib_cpp`). 
+
+Since `pyproject.toml` references `deglib` locally (`path = "../../python"`), running `uv sync` automatically copies the C++ sources from `cpp/`, invokes CMake, and compiles/installs the latest `deglib` C++ bindings into the local environment:
 
 ```bash
+# Install dependencies & compile latest deglib C++ bindings
 uv sync
 ```
 
+> **Note:** A C++ compiler (MSVC on Windows, GCC/Clang on Linux/macOS) and CMake must be available in your system `PATH`. `setup.py` handles copying C++ files, running CMake, and compiling the extension during the `uv sync` step.
+
 ## Key Differences from `static_data`
 
 - Iterates over 3 `DataStreamType`s instead of a single `AddAll`

examples/knng/README.md

@@ -13,6 +13,19 @@ The approach demonstrates **Mode 4 (`evp-rerank`)**, which achieves state-of-the
 3. **Graph Exploration**: Exploration for vertex $i$ walks the DEG graph neighborhood using fast EVP bit-level inner product distances to collect candidates (`evpK = 50`).
 4. **FP16 Candidate Reranking**: Exact inner-product distances are computed using `deglib_cpp.floats_to_fp16` and `deglib_cpp.fp16_to_floats` for candidate sets to produce final $k$-nearest neighbor edges.
 
+## Prerequisites: Building the Python Library (`deglib`)
+
+The `deglib` Python library contains C++ pybind11 bindings (`deglib_cpp`). 
+
+Since `pyproject.toml` references `deglib` locally (`path = "../../python"`), running `uv sync` automatically copies the C++ sources from `cpp/`, invokes CMake, and compiles/installs the latest `deglib` C++ bindings into the local environment:
+
+```bash
+# Install dependencies & compile latest deglib C++ bindings
+uv sync
+```
+
+> **Note:** A C++ compiler (MSVC on Windows, GCC/Clang on Linux/macOS) and CMake must be available in your system `PATH`. `setup.py` handles copying C++ files, running CMake, and compiling the extension during the `uv sync` step.
+
 ## Running the Benchmark with `uv`
 
 ### 1. Run Full Benchmark (200K Wikipedia BGE-M3 vectors)

examples/static_data/README.md

@@ -10,18 +10,18 @@ This example project demonstrates how to download paper datasets from `readme.md
 - `deep1m` (96D, 1M base vectors, L2 distance)
 - `glove-100` (100D, 1.18M base vectors, Angular / InnerProduct distance)
 
-## Prerequisites & UV Setup
-
-1. **Install `uv`** (if not already installed):
-   - **Linux/macOS**: `curl -LsSf https://astral.sh/uv/install.sh | sh`
-   - **Windows (PowerShell)**: `powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"`
-   - **Pip**: `pip install uv`
-
-2. **Sync Dependencies**:
-   Inside this directory, initialize the environment and install dependencies (including building local `deglib` bindings):
-   ```bash
-   uv sync
-   ```
+## Prerequisites: Building the Python Library (`deglib`)
+
+The `deglib` Python library contains C++ pybind11 bindings (`deglib_cpp`). 
+
+Since `pyproject.toml` references `deglib` locally (`path = "../../python"`), running `uv sync` automatically copies the C++ sources from `cpp/`, invokes CMake, and compiles/installs the latest `deglib` C++ bindings into the local environment:
+
+```bash
+# Install dependencies & compile latest deglib C++ bindings
+uv sync
+```
+
+> **Note:** A C++ compiler (MSVC on Windows, GCC/Clang on Linux/macOS) and CMake must be available in your system `PATH`. `setup.py` handles copying C++ files, running CMake, and compiling the extension during the `uv sync` step.
 
 ## Running the Benchmark
 

python/setup.py

@@ -50,26 +50,16 @@ def finalize_options(self):
         pass
 
     def run(self):
+        ignore_dirs = shutil.ignore_patterns('external', 'cmake-build*', 'build', 'benchmark', '.venv', '.git*')
         copy_dirs = [
             (os.path.join('..', 'cpp'), 'lib')
         ]
 
         for src, dst in copy_dirs:
             print(f"Copying {src} to {dst}")
-            shutil.copytree(src, dst, dirs_exist_ok=True)
+            shutil.copytree(src, dst, dirs_exist_ok=True, ignore=ignore_dirs)
         print("Files copied successfully.")
 
-        rm_dirs = [
-            os.path.join('lib', 'external'),
-            os.path.join('lib', 'cmake-build-default'),
-            os.path.join('lib', 'build'),
-            os.path.join('lib', 'benchmark'),
-        ]
-        for rm_dir in rm_dirs:
-            if os.path.isdir(rm_dir):
-                print(f'Removing "{rm_dir}"')
-                shutil.rmtree(rm_dir)
-
 
 class CopySDist(sdist_class):
     def run(self):