tinygnn 0.1.2__tar.gz

tinygnn-0.1.2/CITATION.cff

@@ -0,0 +1,39 @@
+# This CITATION.cff file was generated with cffinit.
+# Visit https://bit.ly/cffinit to generate yours today!
+
+cff-version: 1.2.0
+title: "TinyGNN: A Zero-Dependency C++ Inference Engine for Sparse Graph Neural Networks"
+message: >-
+  If you use this software, please cite it using the
+  metadata from this file.
+type: software
+authors:
+  - given-names: Jai Ansh Singh
+    family-names: Bindra
+    affiliation: JBAC EdTech
+  - given-names: Anubhav
+    family-names: Choudhery
+    affiliation: JBAC EdTech
+repository-code: "https://github.com/JaiAnshSB/TinyGNN"
+url: "https://github.com/JaiAnshSB/TinyGNN"
+abstract: >-
+  TinyGNN is a dependency-free C++17 inference engine optimized for
+  Sparse Graph Neural Networks (GCN, GraphSAGE, GAT). It features
+  hybrid Dense/SparseCSR tensor storage, AVX2 SIMD intrinsics,
+  OpenMP thread parallelism, operator fusion for GAT and GraphSAGE
+  kernels, and Python bindings via pybind11 with zero-copy NumPy,
+  SciPy, and PyTorch interoperability.
+keywords:
+  - graph neural networks
+  - sparse inference
+  - GCN
+  - GraphSAGE
+  - GAT
+  - C++17
+  - pybind11
+  - SIMD
+  - operator fusion
+  - zero-dependency
+license: MIT
+version: 0.1.0
+date-released: "2026-02-28"

tinygnn-0.1.2/CMakeLists.txt

@@ -0,0 +1,92 @@
+cmake_minimum_required(VERSION 3.16)
+project(TinyGNN VERSION 0.1.0 LANGUAGES CXX)
+
+# ── C++17 standard ──────────────────────────────────────────────────────────
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+set(CMAKE_CXX_EXTENSIONS OFF)
+
+# ── Compiler warnings ──────────────────────────────────────────────────────
+if(MSVC)
+    add_compile_options(/W4)
+else()
+    add_compile_options(-Wall -Wextra -Wpedantic)
+    # Static link on MinGW to avoid ld dynamic linking issues
+    if(MINGW)
+        set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -static")
+    endif()
+endif()
+
+# ── Core library ────────────────────────────────────────────────────────────
+add_library(tinygnn_core
+    src/tensor.cpp
+    src/graph_loader.cpp
+    src/ops.cpp
+    src/layers.cpp
+    src/model.cpp
+)
+
+target_include_directories(tinygnn_core PUBLIC
+    ${CMAKE_CURRENT_SOURCE_DIR}/include
+)
+
+# ── OpenMP (Phase 8) ───────────────────────────────────────────────────────
+find_package(OpenMP)
+if(OpenMP_CXX_FOUND)
+    target_link_libraries(tinygnn_core PUBLIC OpenMP::OpenMP_CXX)
+endif()
+
+# ── AVX2 + FMA intrinsics (Phase 8) ────────────────────────────────────────
+include(CheckCXXCompilerFlag)
+check_cxx_compiler_flag("-mavx2" COMPILER_SUPPORTS_AVX2)
+check_cxx_compiler_flag("-mfma"  COMPILER_SUPPORTS_FMA)
+if(COMPILER_SUPPORTS_AVX2 AND COMPILER_SUPPORTS_FMA)
+    target_compile_options(tinygnn_core PRIVATE -mavx2 -mfma)
+endif()
+
+# ── Tests ───────────────────────────────────────────────────────────────────
+enable_testing()
+
+add_executable(test_tensor tests/test_tensor.cpp)
+target_link_libraries(test_tensor PRIVATE tinygnn_core)
+add_test(NAME TensorTests COMMAND test_tensor)
+
+add_executable(test_graph_loader tests/test_graph_loader.cpp)
+target_link_libraries(test_graph_loader PRIVATE tinygnn_core)
+add_test(NAME GraphLoaderTests COMMAND test_graph_loader)
+
+add_executable(test_matmul tests/test_matmul.cpp)
+target_link_libraries(test_matmul PRIVATE tinygnn_core)
+add_test(NAME MatmulTests COMMAND test_matmul)
+
+add_executable(test_spmm tests/test_spmm.cpp)
+target_link_libraries(test_spmm PRIVATE tinygnn_core)
+add_test(NAME SpmmTests COMMAND test_spmm)
+
+add_executable(test_activations tests/test_activations.cpp)
+target_link_libraries(test_activations PRIVATE tinygnn_core)
+add_test(NAME ActivationsTests COMMAND test_activations)
+
+add_executable(test_gcn tests/test_gcn.cpp)
+target_link_libraries(test_gcn PRIVATE tinygnn_core)
+add_test(NAME GCNTests COMMAND test_gcn)
+
+add_executable(test_graphsage tests/test_graphsage.cpp)
+target_link_libraries(test_graphsage PRIVATE tinygnn_core)
+add_test(NAME GraphSAGETests COMMAND test_graphsage)
+
+add_executable(test_gat tests/test_gat.cpp)
+target_link_libraries(test_gat PRIVATE tinygnn_core)
+add_test(NAME GATTests COMMAND test_gat)
+
+add_executable(test_e2e tests/test_e2e.cpp)
+target_link_libraries(test_e2e PRIVATE tinygnn_core)
+add_test(NAME E2ETests COMMAND test_e2e)
+
+# ── Benchmark (Phase 8) ────────────────────────────────────────────────────
+add_executable(bench_parallel benchmarks/bench_parallel.cpp)
+target_link_libraries(bench_parallel PRIVATE tinygnn_core)
+
+# ── Benchmark (Phase 9: Operator Fusion) ──────────────────────────────────
+add_executable(bench_fusion benchmarks/bench_fusion.cpp)
+target_link_libraries(bench_fusion PRIVATE tinygnn_core)

tinygnn-0.1.2/CONTRIBUTING.md

@@ -0,0 +1,142 @@
+# Contributing to TinyGNN
+
+Thank you for your interest in contributing to TinyGNN! This document provides guidelines and information for contributors.
+
+## Authors
+
+**Jai Ansh Singh Bindra and Anubhav Choudhery (under JBAC EdTech)**
+
+## Getting Started
+
+### Prerequisites
+
+- **C++17 compiler**: GCC 8+, Clang 7+, or MSVC 2019+
+- **CMake 3.16+**
+- **Python 3.8+** (for Python bindings)
+- **pybind11 ≥ 2.11** (for Python bindings)
+
+### Setting Up the Development Environment
+
+```bash
+# Clone the repository
+git clone https://github.com/JaiAnshSB/TinyGNN.git
+cd TinyGNN
+
+# Create a virtual environment
+python -m venv .venv
+source .venv/bin/activate  # Linux/macOS
+# .venv\Scripts\activate   # Windows
+
+# Install development dependencies
+pip install -e ".[dev]"
+
+# Build and test C++
+cmake -S . -B build -DCMAKE_BUILD_TYPE=Debug
+cmake --build build --parallel
+ctest --test-dir build --output-on-failure
+
+# Run Python tests
+python -m pytest tests/test_python_bindings.py -v
+```
+
+## Development Workflow
+
+### Code Style
+
+**C++ (backend)**:
+- C++17 standard
+- 4-space indentation
+- `snake_case` for functions and variables
+- `PascalCase` for types and classes
+- Comprehensive documentation comments (Doxygen-style)
+- All public functions must have `@throws` documentation
+
+**Python (frontend)**:
+- Follow PEP 8
+- Type hints where practical
+- Docstrings for all public functions
+
+### Adding a New Operation
+
+1. **Declare** in the appropriate header (`include/tinygnn/ops.hpp` or `layers.hpp`)
+2. **Implement** in the corresponding source (`src/ops.cpp` or `src/layers.cpp`)
+3. **Bind** in `python/tinygnn_ext.cpp`
+4. **Export** in `python/tinygnn/__init__.py`
+5. **Test** with both C++ assertions and Python pytest
+
+### Running Tests
+
+```bash
+# C++ tests (all 9 test suites, 18,000+ assertions)
+cmake --build build --parallel
+ctest --test-dir build --output-on-failure
+
+# Python tests (49 tests)
+python -m pytest tests/test_python_bindings.py -v
+
+# Memory safety (Linux/WSL only)
+bash scripts/sanitizers.sh       # ASan + UBSan
+bash scripts/valgrind_all.sh     # Memcheck + Helgrind
+```
+
+### Building Documentation
+
+```bash
+# C++ docs (Doxygen)
+doxygen Doxyfile
+# Output: docs/doxygen/html/index.html
+
+# Python docs (Sphinx)
+pip install sphinx sphinx-rtd-theme breathe
+cd docs && make html
+# Output: docs/_build/html/index.html
+```
+
+## Pull Request Guidelines
+
+1. **Branch from `main`** and target `main` for your PR
+2. **Keep commits focused** — one logical change per commit
+3. **Write descriptive commit messages** following the format:
+   ```
+   Phase N: Brief description of change
+
+   Detailed explanation of what changed and why.
+   Include test results where applicable.
+   ```
+4. **All existing tests must pass** — CI will verify this automatically
+5. **Add tests for new functionality** — aim for comprehensive coverage
+6. **Update documentation** when adding public API surface
+
+## Architecture Overview
+
+```
+include/tinygnn/     → C++ public headers (tensor, ops, layers, model)
+src/                 → C++ implementations
+python/              → pybind11 bindings + Python package
+tests/               → C++ test suites + Python pytest
+benchmarks/          → Performance benchmark programs
+scripts/             → Build helpers, training, validation
+docs/                → Sphinx + Doxygen documentation
+.github/workflows/   → GitHub Actions CI/CD
+```
+
+### Design Principles
+
+- **Zero dependencies** — no Eigen, BLAS, Boost, or external libraries
+- **Fail-fast validation** — descriptive exceptions for all invalid inputs
+- **In-place where possible** — minimize memory allocations
+- **Cache-aware** — (i,k,j) loop order, AVX2 SIMD, OpenMP parallelism
+- **Operator fusion** — fused GAT/SAGE kernels reduce memory bandwidth
+
+## Reporting Issues
+
+When reporting bugs, please include:
+- Operating system and compiler version
+- Python version (if applicable)
+- Steps to reproduce
+- Expected vs. actual behavior
+- Relevant error messages or stack traces
+
+## License
+
+By contributing to TinyGNN, you agree that your contributions will be licensed under the MIT License.

tinygnn-0.1.2/Doxyfile

@@ -0,0 +1,89 @@
+# =============================================================================
+#  TinyGNN — Doxygen Configuration
+#  Doxyfile
+#
+#  Generates C++ API documentation from the annotated header files.
+#
+#  Usage:
+#    doxygen Doxyfile
+#
+#  Output: docs/doxygen/html/index.html
+# =============================================================================
+
+# ── Project metadata ─────────────────────────────────────────────────────────
+PROJECT_NAME           = "TinyGNN"
+PROJECT_NUMBER         = "0.1.0"
+PROJECT_BRIEF          = "A zero-dependency C++17 inference engine for Sparse Graph Neural Networks"
+PROJECT_LOGO           =
+
+# ── Input ────────────────────────────────────────────────────────────────────
+INPUT                  = include/tinygnn \
+                         src
+INPUT_ENCODING         = UTF-8
+FILE_PATTERNS          = *.hpp *.cpp
+RECURSIVE              = YES
+
+# ── Output ───────────────────────────────────────────────────────────────────
+OUTPUT_DIRECTORY       = docs/doxygen
+GENERATE_HTML          = YES
+HTML_OUTPUT            = html
+HTML_THEME             = 
+GENERATE_LATEX         = NO
+GENERATE_XML           = YES
+XML_OUTPUT             = xml
+
+# ── Extraction ───────────────────────────────────────────────────────────────
+EXTRACT_ALL            = YES
+EXTRACT_PRIVATE        = NO
+EXTRACT_STATIC         = YES
+EXTRACT_LOCAL_CLASSES  = YES
+EXTRACT_ANON_NSPACES   = NO
+
+# ── Source browsing ──────────────────────────────────────────────────────────
+SOURCE_BROWSER         = YES
+INLINE_SOURCES         = NO
+STRIP_CODE_COMMENTS    = NO
+REFERENCED_BY_RELATION = YES
+REFERENCES_RELATION    = YES
+
+# ── Namespaces ───────────────────────────────────────────────────────────────
+HIDE_UNDOC_MEMBERS     = NO
+HIDE_UNDOC_CLASSES     = NO
+SHOW_NAMESPACES        = YES
+
+# ── Graphs ───────────────────────────────────────────────────────────────────
+HAVE_DOT               = YES
+DOT_IMAGE_FORMAT       = svg
+INTERACTIVE_SVG        = YES
+CLASS_DIAGRAMS         = YES
+COLLABORATION_GRAPH    = YES
+INCLUDE_GRAPH          = YES
+INCLUDED_BY_GRAPH      = YES
+CALL_GRAPH             = NO
+CALLER_GRAPH           = NO
+GRAPHICAL_HIERARCHY    = YES
+DIRECTORY_GRAPH        = YES
+DOT_GRAPH_MAX_NODES    = 100
+MAX_DOT_GRAPH_DEPTH    = 0
+
+# ── Preprocessing ────────────────────────────────────────────────────────────
+ENABLE_PREPROCESSING   = YES
+MACRO_EXPANSION        = YES
+EXPAND_ONLY_PREDEF     = NO
+PREDEFINED             = __AVX2__ \
+                         _OPENMP
+
+# ── Warnings ─────────────────────────────────────────────────────────────────
+WARNINGS               = YES
+WARN_IF_UNDOCUMENTED   = YES
+WARN_IF_DOC_ERROR      = YES
+WARN_NO_PARAMDOC       = YES
+
+# ── Styling ──────────────────────────────────────────────────────────────────
+GENERATE_TREEVIEW      = YES
+DISABLE_INDEX          = NO
+FULL_SIDEBAR           = NO
+HTML_COLORSTYLE        = LIGHT
+HTML_DYNAMIC_MENUS     = YES
+HTML_DYNAMIC_SECTIONS  = YES
+SEARCHENGINE           = YES

tinygnn-0.1.2/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025-2026 Jai Ansh Singh Bindra and Anubhav Choudhery (under JBAC EdTech)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

tinygnn-0.1.2/MANIFEST.in

@@ -0,0 +1,40 @@
+# Include C++ source and headers for source distributions
+include CMakeLists.txt
+include README.md
+include LICENSE
+include pyproject.toml
+include setup.py
+
+# C++ sources
+recursive-include include *.hpp
+recursive-include src *.cpp
+
+# pybind11 bindings
+include python/tinygnn_ext.cpp
+recursive-include python/tinygnn *.py
+
+# Benchmarks (for reference, not required for build)
+recursive-include benchmarks *.cpp
+
+# Scripts
+recursive-include scripts *.py *.sh
+
+# Tests
+recursive-include tests *.cpp *.py
+
+# Documentation
+recursive-include docs *
+include Doxyfile
+include CITATION.cff
+include CONTRIBUTING.md
+
+# Exclude build artifacts
+global-exclude *.o *.obj *.exe *.dll *.so *.dylib *.pyd
+global-exclude __pycache__ *.pyc
+prune build
+prune build-cpp
+prune .venv
+prune venv
+prune datasets
+prune weights
+prune .git