itables can be a great alternative here, as it eliminates the need for additional dependencies. Here's how you can use it:

from itables import show
show(df)

If you prefer to avoid the IPython dependency, you can consider the following approach:

Additionally, do you have any suggestions for writing tests for this integration?

itables can be a great alternative here, as it eliminates the need for additional dependencies. Here's how you can use it:

I assume astropy picked the other library for a reason?

What is the difference between them?

Also is this an AI answer?

Additionally, do you have any suggestions for writing tests for this integration?

Honestly, I do not. How did astropy do it?

Since astropy are dealing with these types of dependencies, I think investigating how they do that and copying their solution is the simplest path forward.

I assume astropy picked the other library for a reason?

I believe maybe they didn't pick itables because "itables appears to change how dataframes are rendered Notebook-wide which I felt isn't something we're going for." said by exitflynn, although this is not a problem now because itable have show(df) method which only converts particular dataframes rather than all.

What is the difference between them?

If we use itables we can directly use their show method which can display them in jupyter notbook, but in astropy they have used ipydatagrid to single tables and returned them directly, which we cannot do in sunpy as there can be multiple tables in a fido result, also this is the reason why we cannot use astropy's show_in_notebook method.

Edit: itables looks much better and feels faster to use.

Also is this an AI answer?

Neither the code nor the comments are ai (although I use grammarly), I guess it is a problem? I will stop that now.

Additionally, do you have any suggestions for writing tests for this integration?

Honestly, I do not. How did astropy do it?

Since astropy are dealing with these types of dependencies, I think investigating how they do that and copying their solution is the simplest path forward.

Yes, they just checked isInstance(t.show_in_table(), Datagrid) but again this wouldn't work here, because we are not returning Datagrid as they are.

I assume astropy picked the other library for a reason?

I believe maybe they didn't pick itables because "itables appears to change how dataframes are rendered Notebook-wide which I felt isn't something we're going for." said by exitflynn, although this is not a problem now because itable have show(df) method which only converts particular dataframes rather than all.

How does it change them compared to the other library?

What is the difference between them?

If we use itables we can directly use their show method which can display them in jupyter notbook, but in astropy they have used ipydatagrid to single tables and returned them directly, which we cannot do in sunpy as there can be multiple tables in a fido result, also this is the reason why we cannot use astropy's show_in_notebook method.

So how does both libraries handle having multiple tables and rendering them in a notebook?

Also is this an AI answer?

Neither the code nor the comments are ai (although I use grammarly), I guess it is a problem? I will stop that now.

No that's fine. It's a common tool and its different from Gen AI.

Additionally, do you have any suggestions for writing tests for this integration?

Honestly, I do not. How did astropy do it?
Since astropy are dealing with these types of dependencies, I think investigating how they do that and copying their solution is the simplest path forward.

Yes, they just checked isInstance(t.show_in_table(), Datagrid) but again this wouldn't work here, because we are not returning Datagrid as they are.

Hmm, I expect that properly testing this kind of output is painful and why it was not done. Can you check how the quicklook method is tested inside sunpy?

I assume astropy picked the other library for a reason?

I believe maybe they didn't pick itables because "itables appears to change how dataframes are rendered Notebook-wide which I felt isn't something we're going for." said by exitflynn, although this is not a problem now because itable have show(df) method which only converts particular dataframes rather than all.

How does it change them compared to the other library?

itable has this feature as well

from itables import init_notebook_mode

init_notebook_mode(all_interactive=True)

This converts all dataframes into those interactive tables, this is what exitflynn meant although i guess he missed or they added show(df) new method later, which makes changes to particular dataframes only

What is the difference between them?

If we use itables we can directly use their show method which can display them in jupyter notbook, but in astropy they have used ipydatagrid to single tables and returned them directly, which we cannot do in sunpy as there can be multiple tables in a fido result, also this is the reason why we cannot use astropy's show_in_notebook method.

So how does both libraries handle having multiple tables and rendering them in a notebook?

Naturally they cannot render multiple tables, but itable renders the tables in distinct boxes with search button (image in the comment above) which looks good.

Additionally, do you have any suggestions for writing tests for this integration?

Honestly, I do not. How did astropy do it?
Since astropy are dealing with these types of dependencies, I think investigating how they do that and copying their solution is the simplest path forward.

Yes, they just checked isInstance(t.show_in_table(), Datagrid) but again this wouldn't work here, because we are not returning Datagrid as they are.

Hmm, I expect that properly testing this kind of output is painful and why it was not done. Can you check how the quicklook method is tested inside sunpy?

Yes I had a look, I will try using mocking.

Naturally they cannot render multiple tables, but itable renders the tables in distinct boxes with search button (image in the comment above) which looks good.

So in your opinion, which library should we use?

Yes I had a look, I will try using mocking.

Thanks

Naturally they cannot render multiple tables, but itable renders the tables in distinct boxes with search button (image in the comment above) which looks good.

So in your opinion, which library should we use?

itables, for some reasons I am not able to run the test locally I am pushing to test it.
Let me know if we are sticking with ipydatagrid, I will make changes accordingly and fix pre-commit and changelog

Also can you please clarify what you mean by attrs table, and how they are generated, a code example maybe?

Also can you please clarify what you mean by attrs table, and how they are generated, a code example maybe?

https://docs.sunpy.org/en/stable/tutorial/acquiring_data/index.html#searching-for-data

Also can you please clarify what you mean by attrs table, and how they are generated, a code example maybe?

https://docs.sunpy.org/en/stable/tutorial/acquiring_data/index.html#searching-for-data

Is the position and the name of search function acceptable? I have created a _create_table function to reduce redundancy of code

Looks fine with me.

1. Why are we using itables over ipydatagrid? Both seem relatively popular but the former is maintained by a single person and the latter is maintained by the Jupyter project.

Hi, ipydatagrid felt much slower and more laggy when I used it, which is why we decided to go with tables. Ref comment

2. How are we handling the itables/ipydatagrid dependency? Is an additional optional dependency category for notebook-related things being added?

I'm not quite sure could you please let me know where the dependency should be placed

2. How are we handling the itables/ipydatagrid dependency? Is an additional optional dependency category for notebook-related things being added?

I've now added an optional jupyter extra which also has ipywidgets in it for the parfive progress bars.

Changed test to be offline.

Thank you again @Prtm2110