JOSS comment (from tamnva)

[tamnva]

Dear authors of AquaFetch,

First of all, I would like to thank the authors for creating this package for the community. This issue was created to store my comments for the manuscript submitted to JOSS (openjournals/joss-reviews#8051). Please find below the reasons for the unchecked items in JOSS reviewe.

General checks

1. Contribution and authorship: please explain the contribution to the package developemnt of all coauthors and the list of people that committed to the package github repo

Functionality

1. Installation/Installation instructions: The dev_requirements.txt and requirements.txt file should list of dependency packages, including the python version.

2. Functionality

• The code provided in the Usage section of the README file is quite confusing and does not work, mixing of both python script and outputs.
• The package name is "aqua-fetch" but when import in python using the name "aqua_fetch" (for example from aqua_fetch import RainfallRunoff). Please check the name following standard naming convention, for example using PEP 8 https://peps.python.org/pep-0008/

Sorry I cannot further review the environment as the I have issue with importing the software, could the authors please have a look at this so I could continue my review

Note: attached is the environment that I am working with and the error when I importing the function

from aqua_fetch import RainfallRunoff
Traceback (most recent call last):
File "", line 1, in
File "C:\Users\user_name\AppData\Local\anaconda3\envs\aqua1\Lib\site-packages\aqua_fetch_init_.py", line 7, in
from .rr import RainfallRunoff
File "C:\Users\user_name\AppData\Local\anaconda3\envs\aqua1\Lib\site-packages\aqua_fetch\rr_init.py", line 14, in
from .utils import _RainfallRunoff
File "C:\Users\user_name\AppData\Local\anaconda3\envs\aqua1\Lib\site-packages\aqua_fetch\rr\utils.py", line 31, in
class _RainfallRunoff(Datasets):
File "C:\Users\user_name\AppData\Local\anaconda3\envs\aqua1\Lib\site-packages\aqua_fetch\rr\utils.py", line 881, in _RainfallRunoff
) -> plt.Axes:
^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'Axes'

[tamnva]

environment.txt

[jeiloh]

The command pip install aqua-fetch from PyPI fails. Initially, it throws AttributeError: 'NoneType' object has no attribute 'Axes'. I think this is because matplotlib is not installed as a base dependency.

After manually installing matplotlib, a different error appears on import: AttributeError: 'NoneType' object has no attribute 'Dataset'. This seems xarray is also missing from the base dependencies required for import.
Confirmed that installing via pip install "aqua-fetch[all] @ git+https://github.com/hyex-research/AquaFetch.git" worked correctly.

@AtrCheema Could the PyPI package be updated, or the base dependency specification (install_requires) be reviewed to ensure pip install aqua-fetch provides a working installation out-of-the-box?

[tamnva]

Dear authors of AquaFetch,

Please find below my comments (in addition to the comments in my first post)

###General checks

• Substantial scholarly effort: Could the authors provide some examples where the software has already been cited in academic papers?
• Whether the software is sufficiently useful that it is likely to be cited by other researchers working in this domain: I think this is usefull in the case that someone wants to automatically download the data and convert to specific format for further analysis (e.g., large sample analysis or machine learning modelling exercise).

Functionality

• There are errors in importing functions, running the examples codes (either in the readme.md file or in the documentation, please find the error message in the attached files, the environment file was also attached)

Documentation

• Please update the documentation and rerun the python scripts included in this documentation to make sure that they works, there are also some error in the documentations (https://aquafetch.readthedocs.io/en/latest/auto_examples/index.html)
• Automated tests: I think the current testing using the github workflow (https://github.com/hyex-research/AquaFetch/blob/master/.github/workflows/draft-pdf.yml) is for testing the creation of the pdf file for JOSS, not testing the package functionality itself, is there a way to include testing of the package functionality within this workflow?

error_1.txt
error_2.txt
error_3.txt
error_4.txt
error_5.txt
error_6.txt
error_7.txt

environment.txt

[tamnva]

Just one small comment. The function dataset.fetch() in the readme example requires very high computational resources (e.g., with the CAMELS_AUS data) eventhough users just want to extract a single features. It would be nice if the developers can optimize this

[AtrCheema]

Hi @tamnva do you mean a extracting data for a single (dynamic) feature or extracting data for a single station?

and you mean the function is slow with xarray installation or without xarray (minimal installation)?

[tamnva]

Hi, in both cases (when I extracting single dynamic feature and extracting data for a single stations?) - the code I took from your usage section in the README (it takes around half a minutes with my PC). I have xarray installed already