Code quality

For usability and reliability of the software the code needs to be high quality. For this project it means meeting the four criteria described below. Any addition should pass all of them.

Look, read and feel like Python

When using classes and functions from the package it should feel like you are using Python tools and be forces to write C++ style code with Python syntax.

with xray_box.shutter_open():
    for th in threshold:
        d.vthreshold = th
        d.acq()

should be preferred over

N = len(threshold)
xray_box.open_shutter()
for i in range(N):
    d.dacs.set_dac('vthreshold', threshold[i])
    d.acq()
xray_box.close_shutter()

even if the difference might seem small.

Have documentation

Classes and functions should be documented with doc-strings in the source code. Preferably with examples. The syntax to be used is numpy-sphinx.

def function(arg):
    """
    This is a function that does something

    Parameters
    ----------
    arg: int
        An argument

    Returns
    --------
    value: double
        Returns a value

    """
    return np.sin(arg+np.pi)

Pass static analysis with pylint

Yes, anything less than 9/10 just means that you are lazy. If there is a good reason why to diverge, then we can always add an exception.

Currently the following additions are made:

• good-names: x, y, ax, im etc.
• function arguments 10
• Whitelist: numpy, _sls

Tested code

Last but not least… actually last just because of the long list included. All code that goes in should have adequate tests. If a new function does not have a minimum of one test it does not get added.

Unit-tests with pytest and mocker

----------- coverage: platform linux, python 3.6.4-final-0 -----------
Name                         Stmts   Miss  Cover
------------------------------------------------
sls_detector/__init__.py         4      0   100%
sls_detector/decorators.py      14      3    79%
sls_detector/detector.py       461    115    75%
sls_detector/eiger.py          150     64    57%
sls_detector/errors.py           7      0   100%
sls_detector/jungfrau.py        59     26    56%
------------------------------------------------
TOTAL                          695    208    70%


========= 78 passed in 0.60 seconds =========

Simple integration tests

These tests require a detector connected. Performs simple tasks like setting exposure time and reading back to double check the value

----------- coverage: platform linux, python 3.6.4-final-0 -----------
Name                         Stmts   Miss  Cover
------------------------------------------------
sls_detector/__init__.py         4      0   100%
sls_detector/decorators.py      14      0   100%
sls_detector/detector.py       461    103    78%
sls_detector/eiger.py          150     20    87%
sls_detector/errors.py           7      0   100%
sls_detector/jungfrau.py        59     26    56%
------------------------------------------------
TOTAL                          695    149    79%


========= 67 passed, 1 skipped in 16.66 seconds =========

Complex integration test

Typical measurements. Might require X-rays. Tests are usually evaluated from plots