maayanlab / appyter Goto Github PK

Would address #8

Currently 'textlistfield' is more like 'multicheckboxfield'

Now that we're using eventlet, we're getting this issue: eventlet/eventlet#526

Maybe we need to use gevent instead.

Carriage returns do not clear the output of a cell. The output is only displayed at the end of the evaluation of the cell, even if prints are called during evaluation.

StringField does not currently support clickable example inputs like FileField does. Thank you!

When selecting a file to upload, the file is uploaded fine and the green bar shows that the upload is done by the text box above still says "Choose file". The expected behavior is that the uploaded file name will be displayed in the text box.

https://components.nteract.io/

We might need to know about the profile at jupyter notebook creation time, we might as well collect it at init. We could then parse it and we wouldn't need to specify profile when we run jupyter-template, we could grab it from the notebook.

This should continue to support existing components but make them more modular and easier to extend in the future

Right now there are no margins and no title

People should be able to go back to an already evaluated notebook and see the results.

Also if you get disconnected during evaluation, your client should continue where it left off instead of restarting everything.

For this effort, the session ID should be available in the URL

To reproduce run a tcgaEnrichrViewer instance, then go back with the browser back button, change the cancer type and press submit... The results will run the old/previous settings...

For example, if I declare and initiate my list list, then try to use list as my default for a StringField, jupyter notebook will return the issue of: Error: undeclared variable(s) {list}.

Re-rendering it on every request is slow

from IPython.display import display, Markdown
display(Markdown("str"))
is not working. When I run the jupyter-template, it just shows that <IPython.core.display.Markdown object>

The working directory should be the main notebook so that external scripts can be called if necessary, but files should automatically get put in a session based subdirectory referenced in the jupyter notebook.

There are various places we could cache things to speed up execution over time and/or not do redundant work.

• caching network requests
• automatic function call memoization with joblib's Memory
• cell with dependencies caching (docker-style)
• duplicate input parameters

We need a landing page that provides all high quality Jupyter Templates implementations...

To provide an example analysis to users, it would be great to give an option of loading example datasets or showing the example analysis results.

The default is 0~10 which is pretty useless usually; it should default to no min/max..

I'm thinking this should just be an external component -- it's possible that such a thing already exists for jupyter notebook.

This functionality, likely from the debugging mode, should be disabled in production

After uploading files, the progress bar changes colors from blue to green but it still looks like the upload did not finish. Make it clearer that the file was successfully uploaded.

support copy/pasting a list with newline character as separator?

For instance, a dependent module that is meant to be called.

This would speed things up and lower the overhead of tons of HTTP requests

Using TextListField, when 1 or more options are selected , an empty string is returned.

Including:

• dependent scripts
• requirements.txt
• ...

• Google Colab
• MyBinder
• Kubernetes Job

I tried to display a markdown saying "PRE~~" on the top of the plot but it is placed on the bottom

It should already be possible given that one can inspect an appyter for field settings

Could help with #8

https://github.com/jupyter/nbclient

• nbclient now replaces nbconvert's executeprocessor

• nbclient can facilitate what yieldingexecuteprocessor achieves

• nbclient.client.register_output_hook

  • can be used for #35

May or may not enable some ipywidgets to function

Necessary for jupyter-template-catalog deployment