pypvfs 1.0.3

Python bindings for Pinnacle Technology PVFS (virtual filesystem for physiological recordings)

pypvfs

pypvfs is the PyPI distribution name for Pinnacle Technology PVFS — a virtual filesystem used for physiological recordings. This is not related to the historical HPC “Parallel Virtual File System” or other unrelated “PVFS” acronyms.

After installation you import the pvfs_tools Python package (same module layout as in the former Morelia monolith):

pip install pypvfs

from pvfs_tools.Core.pvfs_data_file import PvfsDataFile

Optional extras

Extra	Purpose
video	WebM/VP8 export (WebMWriter uses PyAV). pip install "pypvfs[video]"
test	pytest + av for running the test suite locally
examples	pyedflib + av for the Tk GUI EDF and WebM examples

Native library

The package ships platform-specific native libraries under pvfs_tools.Core (.dll on Windows, .so on Linux). On PyPI you receive a prebuilt wheel for your platform, so pip install pypvfs does not require a C++ toolchain.

If you install from source (sdist or a checkout), pip will invoke scikit-build-core which compiles src/pvfs_tools/Core/pvfs.cpp and pvfs_wrapper.cpp via CMake. You will need:

• CMake ≥ 3.15 (pip installs cmake from PyPI into the isolated build env if your system lacks it)
• A C++17 compiler (g++/clang++ on Linux/macOS, MSVC 2019+ on Windows)

pip install -e ".[test]"   # builds the native libs and installs in editable mode
pytest

Examples (repository examples/)

Requires a working pypvfs install; GUI examples need a display.

Script	Extra deps	Description
examples/pvfs_create_cli.py	—	CLI to create a synthetic PVFS file
examples/pvfs_to_edf_converter.py	pyedflib	Tk GUI: PVFS → EDF+
examples/pvfs_to_video_converter.py	pypvfs[video]	Tk GUI: PVFS video → WebM

pip install pyedflib
python examples/pvfs_to_edf_converter.py

Documentation

• Quick reference: docs/pvfsQuickStart.txt

Tests

pip install -e ".[test]"
pytest

Repository

Source and issues: see Repository URL in pyproject.toml.

Publishing releases

Step-by-step (tags, GitKraken, PyPI checks): docs/RELEASE.md.

Do not rely on twine upload from a developer machine for routine releases (credentials drift and audit pain). Use the GitHub Action instead.

1. Configure trusted publishing on PyPI for this GitHub repository and workflow publish.yml (OIDC — no PYPI_API_TOKEN in repo secrets unless you choose that path).
2. Bump version in pyproject.toml, commit, then create and push a version tag matching your policy, e.g. v1.0.2:

   git tag v1.0.2
   git push origin v1.0.2
   

   Pushing tag v* runs .github/workflows/publish.yml: cibuildwheel builds a manylinux_2_28_x86_64 wheel inside the PyPA manylinux Docker image and a win_amd64 wheel on a Windows runner, then uploads both (plus the sdist) with pypa/gh-action-pypi-publish. The manylinux image guarantees a low glibc baseline so the published Linux wheel works on RHEL/CentOS/Rocky/Alma 8+, Ubuntu 20.04+, Debian 11+, and Amazon Linux 2023 — not just the build host.
3. Manual run: Actions → Publish to PyPI → Run workflow (same build/upload path).

Note: ptech-morelia already uses a similar pattern (tag push + matrix build + OIDC) in its packaging workflow — keep both projects on the same release habit so versions stay aligned when you cut coordinated releases.