python_ext

⚠️ Warning: This is just a prototype/experimental ⚠️

Overview

• setup.py is used by python to create setuptools pakcage.
• As per setup.py and project settings: | Module | crono_exts.timetagger4vector | | Python Installation folder in vcxproj | C:\Dev\Python311 | | Python folder on the Windows Path | C:\Dev\Python311 |
• Extension project can be opened from Visual Studio using file timetagger4ext.vcxproj.

Directory Structure

Current structure

└───timetagger4ext
    ├───include
    ├───lib  # External libraries (driver)
    ├───src
    │   └───crono_exts  # The package source code
    └───tools  # Has project files, package setup files

Build the package from Command Line

Prerequisites**

1. Just make sure that the following packages are installed

<path-to-python-installation-folder>/pip install numpy setuptools wheel

2. You would need Visual Studio 2022 (not VS Code), you would need to enable C++ and python development in the Installer.
3. To configure it for your environment you would not to set the paths for your install , mentioned in the following document: Write C++ extensions for Python - Visual Studio (Windows) | Microsoft Learn.

Steps

cd to python_ext\timetagger4ext\tools and run:

python setup.py bdist_wheel sdist

It will build both the package binaries and source distribution. Output:

• In directory: python_ext\timetagger4ext\tools\dist
• Wheel package: crono_exts-0.1-cp311-cp311-win_amd64.whl.
• Source: crono_exts-0.1.tar.gz

Install the package

cd to python_ext\timetagger4ext\tools\dist, run:

pip install crono_exts-0.1-cp311-cp311-win_amd64.whl

You should find:

Successfully installed crono-exts-0.1

Next steps:

• Install the package and make sure that the driver dll is installed on the system folder.
• Ensure that the extension APIs are exported, and use them in a sample python app.
• Ensure all source is added to the source package.
• Fill clasifiers in setup.py.
• Try the package on test.pypi, could be using twine

  twine check dist/*
  twine upload dist/* -r testpypi
  

• Support Linux.

ReadOut App

Overview

• Sample application that uses the extension.
• Uses crono_exts.timetagger4vector extension.

Run from the command line.

cd TimeTagger4ReadOut\src python TimeTagger4ReadOut\src

Create MSVS Project (other IDEs are probably better)

Prerequisites

• Check prerequisites of the extension.

Steps

• Create "Python Application" project from Visual Studio.
• If not created, then create an environment from python installation: Right click TimeTagger4ReadOut -> Python Environments then Add Environment then select the underlying environment from Base interpreter, assuming called env1.
• Install numpy: Right click TimeTagger4ReadOut -> Python Environments -> env1 then Manage Python Packages..., then enter numpy and install it.
• Install our extension: Right click TimeTagger4ReadOut -> Python Environments -> env1 then Open Interactive Window, then:

  pip install <path-to-whl>
  

Extras

Which Python Installation to Use on Windows for MSVS

Here are trials with the results, just to avoid in future any one has problems, last one works successfully:

1 - Using bundled python

Installing numpy on Visual Studio bundled python:

C:\Program Files (x86)\Microsoft Visual Studio\Shared\Python39_64>.\scripts\pip install numpy

This installed it on:

C:\Users\basse\AppData\Roaming\Python\Python39\site-packages

Conclusion

• A bit annoying setting the paths, and is an old version, trying with the latest python versions.

2 - Using latest python version 3.12

• Installing latest python: 3.12.
• When setting it as an environment in Visual Studio, a warning that this "This python version is not supported by Visual Studio".
• No problem with python immediately previous version, it's supported.

Conclusion

• Not supported by MSVS.

3 - Using an external python 3.11.9

• Installing python version 3.11.9, asking installer to set it in the environment variables.
• Select it as an environment in Visual Studio Project: Solution Explorer -> TimeTagger4ReadOut -> Python Environments -> Add Environment. Select it from "Base interpreter".
• Install numpy from the project environment.
  • Right click 'env (Ptyhon 3.11(64-bit))' in the Solution Explorer
  • Click "Manage Python Packages"
  • Enter "numpy", as it to install it.
  • Installing numpy on the python environment (C:\Dev\Python311) externally somehow doesn't avail it for the project, the above way works well.
• Update link and include paths in the project to C:\Dev\Python311.
• Build all succeeds.
• When running, import numpy as np is OK, but: line:

  import timetagger4vector as tt4v
  

  Caused a run-time error

  DLL load failed while importing timetagger4vector: The specified module could not be found.
  

  • I copied xtdc4_driver_64.dll & timetagger4vector.pyd to the .py folder (F:\Work\crono-all\timetagger4python\TimeTagger4ReadOut) and it worked.

Conclusion

• Used successfully.