hgraph

HGraph is a functional-reactive time-series engine with a Python-first user experience and a native C++ runtime. Most users author and test graphs with the hgraph Python package; library authors can use the C++ API directly when they need native integration or maximum performance. Both paths wire and execute through the same runtime.

The C++-backed runtime is published under the hgraph distribution name:

python -m pip install hgraph

The distribution exposes the supported hgraph authoring package and its private native _hgraph extension. The 0.8 line replaces the Python runtime maintained on the release/0.5 branch while retaining Python as the primary public API. One wheel per supported platform covers CPython 3.12 and later through the CPython stable ABI. The supported Python and platform policy is recorded in docs/source/developer_guide/release_readiness.rst.

Start with docs/source/getting_started.rst and use docs/source/reference/ for the supported Python types, decorators, operators and modules.

cmake -S . -B build                 # configure (fmt + Catch2 fetched if absent)
cmake --build build -j              # build hgraph_core + tests
ctest --test-dir build --output-on-failure

Requires a C++23 compiler and CMake >= 3.25. Python/nanobind are not needed for the default build (bindings are opt-in via -DHGRAPH_BUILD_PYTHON_BINDINGS=ON).

First-party extensions are co-developed in extensions/ but remain separate native and Python distributions. Kafka is built in-tree for development with -DHGRAPH_BUILD_KAFKA_EXTENSION=ON, or independently from extensions/kafka/ against an installed hgraph SDK. Its wheel is selected from the uv workspace after making that matching SDK discoverable:

CMAKE_PREFIX_PATH=/path/to/hgraph/sdk \
  uv build --wheel --package hgraph-kafka --python 3.12

The core build does not enable the extension by default and does not acquire a Kafka or librdkafka dependency.

Sphinx docs live under docs/source (uv sync --extra docs, then uv run sphinx-build -W -b html docs/source docs/_build/html):

• Getting started — docs/source/getting_started.rst: install the wheel and run a first graph in Python.
• User guide — docs/source/user_guide/: the concepts the runtime implements and the primary Python authoring track (python/: quick start, common tasks, tutorial, programming model). Native C++ authoring is an advanced section for library authors.
• Python API reference — docs/source/reference/: curated reference pages plus a generated inventory of wildcard exports, lazy operators and public submodules.
• Specification — docs/source/specification/: a language-neutral definition of HGraph semantics.
• Developer guide — the authoritative design records (docs/source/developer_guide/): architecture, data structures, wiring, nested graphs, mesh, services, error handling, operators, roadmap.

The narrative documentation's Python examples are executable. They are checked against a real runtime by sphinx-build -b doctest, which needs an importable hgraph; CI runs both that and the warning-free HTML build.

• AGENTS.md — canonical project direction: goals, build philosophy, source layout, dependency policy, git hygiene.
• CLAUDE.md — the operational working guide: the enforced design-first workflow (docs change in the same commit as code), guardrails, architecture map, and current state.