Hello human, I'm @whedon, a robot that can help you with some common editorial tasks. @celliern, @@mstimberg it looks like you're currently assigned to review this paper 🎉.

⭐️ Important ⭐️

If you haven't already, you should seriously consider unsubscribing from GitHub notifications for this (https://github.com/openjournals/joss-reviews) repository. As a reviewer, you're probably currently watching this repository which means for GitHub's default behaviour you will receive notifications (emails) for all reviews 😿

To fix this do the following two things:

1. Set yourself as 'Not watching' https://github.com/openjournals/joss-reviews:

2. You may also like to change your default settings for this watching repositories in your GitHub profile here: https://github.com/settings/notifications

For a list of things I can do to help you, just type:

@whedon commands

For example, to regenerate the paper pdf after making changes in the paper's md or bib files, type:

@whedon generate pdf

Reference check summary:

OK DOIs

- 10.1145/2833157.2833162 is OK
- 10.21105/joss.01356 is OK
- 10.11588/ans.2015.100.20553 is OK
- 10.1109/MCSE.2009.52 is OK
- 10.1007/s10444-019-09666-0 is OK
- 10.5194/gmd-12-1165-2019 is OK
- 10.1109/MCSE.2007.53 is OK
- 10.1109/MCSE.2007.55 is OK
- 10.7717/peerj-cs.103 is OK
- 10.21105/joss.01277 is OK

MISSING DOIs

- https://doi.org/10.1137/110856976 may be missing for title: PyClaw: Accessible, Extensible, Scalable Tools for Wave Propagation Problems

INVALID DOIs

- https://doi.org/10.1038/s41592-019-0686-2 is INVALID because of 'https://doi.org/' prefix

PDF failed to compile for issue #2158 with the following error:

pandoc-citeproc: reference mstimberg not found
Error producing PDF.
! TeX capacity exceeded, sorry [input stack size=5000].
\reserved@a ->\def \reserved@a
*{\let @xs@assign @xs@expand@and@detokenize...
l.369 }

Looks like we failed to compile the PDF

I think there was an issue with @@mstimberg instead of @mstimberg being added as a reviewer. Maybe removing and readding will fix it?

This looks like a problem with a reference in the .md and .bib files - the .md file seems to be trying to reference something that is not in the bib file (or the bib file has a problem that leads to it not being parsed correctly)

👋 @openjournals/dev - can you look at this?

@whedon generate pdf

👉 Check article proof 📄 👈

@xuanxu I cannot edit the checklist and I did not receive an invite – is this because this is my first review for JOSS and I have to do something else first? Or is it because of the earlier @@mstimberg issue?

@mstimberg I guess it's the previous @@ issue, let's try to reinvite you.

@whedon re-invite @mstimberg as reviewer

OK, the reviewer has been re-invited.

@mstimberg please accept the invite by clicking this link: https://github.com/openjournals/joss-reviews/invitations

Thanks @xuanxu, it's working now 👍

First, I wanted to say that I'm impressed ! The library is clean and well organized and I will definitively add it in my research toolbox. It also fit well the Python scientific ecosystem, especially compared with my work () that has not be designed to easily write tensor-based PDE equation. This is a welcome approach as this is a very common task among a wide range of scientific topics.

The features offered by py-pde are multiple and the dependencies despite very narrowed mandatory dependencies (all of them being in the standard scientific stack and shipped with distributions like Anaconda). This should guarantee an easy deployment on clusters.

All tests have run with a fresh conda environment with py=3.6 on my computer.

I still have to play a bit with it to ensure the functional claims of the software. During this time, I have some optional (but welcome) improvement that I could suggest :

Packaging improvement

The optional dependencies could be listed in the setup.py (see in the setuptools doc) to make the process more automatic for the user.

Documentation improvement

Some information are available in the paper may be added also in the documentation: the method used for the operator discretization (finite difference), the temporal solvers available for the explicit as well as for the implicit ones. In general, a "advanced user documentation" could highly improve the adoption of the software : more detail on how to implement custom PDE (with or without numba), some info on the internals (discretization, time-steps choice), and if / how the library can be extended to fit more specific needs.

Similarly, there is no in-between the very short overview and trivial example and the full API documentation (which is welcome but hard to decipher).

Some doc-strings lack of details : the allowed boundary conditions are well described in the raised ValueError, but not in the doc-string itself.

There is more examples available on the repository than shown in the documentation : using Sphinx-Gallery, these examples could easily be integrated in the documentation.

The software seems to embed a benchmark tool : it could be nice to have a summary of the performance of implemented models with and without numba and with and without parallelization.

Community guidelines

It could be nice to add a Code of Conduct (see the scipy one for an extended example, and here for a standard one that you can easily adapt for your repository).

Dear authors and reviewers

We wanted to notify you that in light of the current COVID-19 pandemic, JOSS has decided to suspend submission of new manuscripts and to handle existing manuscripts (such as this one) on a "best efforts basis". We understand that you may need to attend to more pressing issues than completing a review or updating a repository in response to a review. If this is the case, a quick note indicating that you need to put a "pause" on your involvement with a review would be appreciated but is not required.

Thanks in advance for your understanding.

Arfon Smith, Editor in Chief, on behalf of the JOSS editorial team.

Dear @celliern, thank you very much for your suggestions! I managed to include most of your points in the revised code. Thank you especially for suggesting the example gallery – I didn't know about this, but I think it helps showcasing the features of the package tremendously.

I'm looking forward to the remaining comments, but I obviously understand if this will now take much longer than usual. Stay safe and healthy!

@david-zwicker Nice your gallery is wonderful !

Two minor points :

• "Do the authors clearly state what problems the software is designed to solve and who the target audience is?"
  This could be highlighted in the "Getting started" : I guess that such software is designed with researcher and student in mind (more than engineer / industrial target) due to the well known finite difference limitation (especially on complex geometry).

• The CoC is in the repo, but should be linked in the doc and / or in the repository README.
  I'm looking at the paper today.

As a side suggestion (not blocking for the publication) : you will have a strong limitation on hyperbolic equations. This can be (partially) mitigated by introducing "advection" operator that could either implement really simple Gudunov finite volume scheme or upwind finite difference scheme. This will extend the scope of the software.

Stay safe and healthy as well !

For the paper:

• there is still some typo : they could be easily spotted by a spell checker.
• pyclaw is the Clawpack python interface, clawpack is specialized in finite volume (especially Gudunov-like methods), not finite difference.
• you could cite the Dedalus Project for the pseudo-spectral methods.

Otherwise, the paper look fine. I will recommand the paper / software for publication once these different point are fixed.

You have written a nice scientic tool. Nice work !

I join @celliern in congratulating you to your software, this is really an excellent tool and I have rarely seen that much attention to code quality in scientific software (type annotations, etc.)!

I have reported two minor issues (zwicker-group/py-pde#2, zwicker-group/py-pde#3) and three suggestions (zwicker-group/py-pde#4, zwicker-group/py-pde#5, zwicker-group/py-pde#5) directly in the py-pde repository, but @celliern's thorough review certainly made my job much easier.

Fixing the typos in the article is the only real blocker for me – if I wanted to be really nitpicky, I could also mention that the capitalization in the references list is somewhat inconsistent (e.g. "Python" vs. "python").

Pending these minor issues I can easily recommend this paper for publication. 👍

Fixing the typos in the article is the only real blocker for me – if I wanted to be really nitpicky, I could also mention that the capitalization in the references list is somewhat inconsistent (e.g. "Python" vs. "python"

Please do be nitpicky - we need to fix all of these before publication :)

Thank you very much for your constructive and helpful comments. I think I implemented almost all of them, except for the following three:

• Adding the benchmarks to the documentations: I'm not quite sure how to do this, so that it is useful. In particular, the performance will depend quite heavily on the hardware on which the code runs. Moreover, the results of the benchmark need to be cached somehow, so they are not re-run every time the documentation is updated. However, I also do not want to include results obtained on my local machine in the repository. Taken together, I'm not quite sure how to implement this so it's useful.
• Adding notebook-gallery: I had some issues with the normal gallery and thus would not like to increase the complexity of this solution. Moreover, there is a binder link in the Readme that people can use to directly play with the code. This is arguably more useful than static notebooks.
• Advection operators: This would certainly be useful, but I'm not very experienced with this and I'm afraid that a general enough solution is not trivial (even restricted to Cartesian grids) – All examples I found were for 1d systems with a pre-defined advection direction.

However, all your other comments were extremely useful and made the project and in particular the documentation much better. I also revised the paper (I'm sorry for all the spelling mistakes). Let me know if you have further suggestions for improvement!

@mstimberg @celliern:
Both of you have checked all items.
Do you consider your review finished and recommend publication?

@xuanxu: In principle yes, but I'd like to have a final look at the revised paper. I guess the PDF has to be regenerated? I might be able to trigger this myself, let's try:

@whedon generate pdf

commands to whedon need to be at the start of a comment

@whedon generate pdf

👉 Check article proof 📄 👈

@xuanxu I had a final look and can confirm that from my side the review is finished and I recommend the paper publication. 👍

@david-zwicker No problem, they were non blocking suggestion ;)

For the advection specialized schemes I will try to do a pull request when I will have the time (I hope soon?).

@xuanxu I also had a final look on the documentation and paper, and I ecommend the paper publication !