luisscoccola/persistable

density-based clustering for exploratory data analysis based on multi-parameter persistence

persistable's Issues

Clustering Parameters Question

This question is non-blocking for my JOSS review.

Is there a plan to make the parameters stateful or transferrable? For example, after I select some desirable parameters using the persistable method, where do the parameters "live," so to speak, and how could I use these parameters to cluster on another dataset after reloading? I do understand that there is a quick_cluster method in your API, which I appreciate, but as I understand it this method automatically selects parameters based upon the ranges provided. Sorry again if this is due to my misunderstanding of the tool, which is why I add this issue as non-blocking.

Installation requirements

The installation went very smoothly, relying on pip to pull down all the requirements. It would be helpful, however, to list the top level installation requirements in the documentation up front so that users can have a clear idea of what they will be pulling in before they do the install (I know this is covered by requirements.txt but putting it in the README and readthedocs would be beneficial).

#openjournals/joss-reviews#5022

PyPi Registration Question

This issue is non-blocking for my JOSS review.

Sorry if I missed it in the documentation somewhere, but is there a plan to register the package on PyPi and perhaps conda-forge? I noticed that the name persistable is already taken on PyPi, which is rather unfortunate. Is this the reason, or is something else keeping PyPi registration? Mind you, I don't consider it a big deal to not register on these servers, but there is perhaps added visibility for your package that the convenience for installation could bring.

Error listing upon suggested tests

These errors occurred when I ran the following tests shown on the GitHub page after installing persistable on the command line, which went without problems:

pip install pytest playwright pytest-playwright
python -m playwright install --with-deps
pip install -r requirements.txt
python -m setup build_ext --inplace
pytest .

As suggested, I'm sending them on. See attached screen grab.

Dash warnings in ``start_UI(inline=True)``

When running in a notebook on a clean install the start_UI(inline=True) works, but results in a large number of warnings being printed to the notebook from Dash. You may want to find a way to catch or suppress these warnings as they can be distracting.

openjournals/joss-reviews#5022

API Index

This is a relatively low-level request that I have for the documentation, which I otherwise enjoy, but I would like to see an index of the core types and methods of the package in the API reference section. For example, I wonder if it is possible in the sphinx docs setup that you use for the documentation to, in the API section, have a list of types and their methods that link to their full definitions further down the page (in the way that you have currently).

If this is limited by the tools that sphinx provides, it is no worries. I just know that other packages such as those for documenting Julia code allow for an index functionality.

Community guidelines

The Contributing section lists what to do for people wishing to contribute, but it would be beneficial if you could also cover those seeking help with issues (e.g. how to file an issue, and what the expectations are for issue formatting if any), and simply with questions relating to the library (e.g. a discussions section). As a further step it could be helpful to include a code of conduct for the repository, just to make expectations clear.

openjournals/joss-reviews#5022