nu-radio / radiotools Goto Github PK

Since we have implemented some basic unittests so far, we should implement CI to exploit the full power of these tests.
When I have some free time I will give it a shot and will try to start with one of the most common CI's (e.g. something like travis ci). At the beginning only some simple routines are needed I think.
Any objections @cg-laser ?

Executing from radiotools import helper on Python3.11 fails with

>>> from radiotools import helper
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/homebrew/lib/python3.11/site-packages/radiotools/helper.py", line 4, in <module>
    from radiotools.atmosphere import models as atm
  File "/opt/homebrew/lib/python3.11/site-packages/radiotools/atmosphere/models.py", line 3, in <module>
    from past.builtins import xrange
  File "/opt/homebrew/lib/python3.11/site-packages/past/builtins/__init__.py", line 43, in <module>
    from past.builtins.noniterators import (filter, map, range, reduce, zip)
  File "/opt/homebrew/lib/python3.11/site-packages/past/builtins/noniterators.py", line 24, in <module>
    from past.types import basestring
  File "/opt/homebrew/lib/python3.11/site-packages/past/types/__init__.py", line 25, in <module>
    from .oldstr import oldstr
  File "/opt/homebrew/lib/python3.11/site-packages/past/types/oldstr.py", line 5, in <module>
    from collections import Iterable
ImportError: cannot import name 'Iterable' from 'collections' (/opt/homebrew/Cellar/[email protected]/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/collections/__init__.py)

When running: the test code in coordinatesystems.py (python2 radiotools/coordinatesystems.py)

File "radiotools/coordinatesystems.py", line 138, in transform_to_vxB_vxvxB nX, nY = station_position.shape ValueError: need more than 1 value to unpack
This happens when transforming a single station position.
Line 143 looks like it should catch this case (array with shape (3,)) however it never reaches this point.
if(len(station_position.shape) == 1):
Adding somethink like
# if a single station position is transformed: (3,) -> (1, 3) if station_position.ndim == 1: station_position = np.expand_dims(station_position, axis=0)
would solve this issue. However than the test code does not return same values for the vector is both coordinate systems (are they supposed to do that?).

The homepage mentions that the default unit for angles is radians, however everywhere in the documentation and examples degrees are used. This is particularly confusing when looking at the documentation of the __init__ function for cstrafo, where the zenith and azimuth descriptions are in degrees. If this could be updated to use radians instead, it would help first time users like me tremendously 👍🏽

I was trying to install the latest version in my Python 3.9 virtual environment, but kept getting this error:

ImportError: urllib3 v2.0 only supports OpenSSL 1.1.1+

Even though I downgraded my urllib package to version 1.26.6. After digging around for a while I managed to track down the issue to the project's TOML file. With TOML files, an isolated build environment is created. In this build environment, urllib v2.0+ was installed, leading to the crash. A simple fix for this is to add the build requirement "urllib3 < 2.0".

Can this be added or does this potentially cause (security) issues?

Hi,
this might be a good place to share opinions about the conversion to a python 2 and 3 compatible library?!
I started a corresponding branch! Please share any opinion.

Issue
Creating a radiotools.atmosphere.refractivity.RefractivityTable with curved=True throws an error by the attempt at creating an inhomogenous numpy array. In get_cached_table_curved_atmosphere, the list self._distances is appended with arrays of different sizes (distances = np.arange(...) with variable ending and fixed stepsize). Creating an array from this (line 234) throws the error.

Minimal working example

from radiotools.atmosphere import refractivity
tab = refractivity.RefractivityTable(curved=True)

Hi @cg-laser,
this is related to the merged PR #38. The observation level have to be used consistently in functions like:
_get_vertical_height_numeric or _get_distance_xmax. Do you agree?
What I do not quite see is how that effects the calculation via taylor expansions. It is necessary to just include the obersvation_level there too? Or does one need to repeat the determination of the used parameter.
Cheers

Hi @cg-laser,
is there a reason why using while in the code below:

helpers.py:

def get_normalized_angle(angle, degree=False, interval=np.deg2rad([0, 360])): import collections if degree: interval = np.rad2deg(interval) delta = interval[1] - interval[0] if(isinstance(angle, (collections.Sequence, np.ndarray))): angle[angle >= interval[1]] -= delta angle[angle < interval[0]] += delta else: while (angle >= interval[1]): angle -= delta while (angle < interval[0]): angle += delta return angle

PS: is there a better way to reference code from the repo?

On a fresh install of a python environment with Python 3.10.12 and pip 22.0.2 and installing this project via pip install radiotools

importing radiotools fails with

ModuleNotFoundError: No module named 'past'
ModuleNotFoundError: No module named 'scipy'

add

It seems that on the last PR merged the py2 pipeline failed. However the failure seemed not to be related to the changes (based on the log). Or are they? Regardless of this. Is it worth keeping py2 compatibility and the checks or should we remove them?