Python: Test popular data science libraries

[seeM]

We should ensure that Positron's basic functionality works well for all of the popular Python data science libraries. That could involve manually trying out these libraries, and perhaps adding some tests for each of them.

Here's a list of popular Python data science libraries that we might want to consider, sorted by category:

• data
  • pandas
  • numpy
  • networkx
  • scrapy
  • beautiful soup
  • pillow
• machine learning
  • scikit-learn
  • natural language toolkit (nltk)
  • statsmodels
  • gensim
  • xgboost
  • lightgbm
  • spacy
  • pycaret
  • eli5
• computer vision
  • opencv
  • scikit-image
• deep learning
  • tensorflow
  • torch
  • keras
  • theano
  • caffe
• visualization
  • matplotlib
  • scipy
  • seaborn
  • plotly
  • ggplot
  • altair
  • autoviz
  • pydot
• web
  • sqlalchemy
  • flask
  • requests
  • django

[isabelizimm]

similar to https://github.com/posit-dev/positron/issues/1073 but testing rather than custom env pane--should the list of libraries be the same?

other potential libraries:

• data validation
  • pandera
  • pydantic
• dataframe
  • polars
  • ibis
  • siuba
• viz ++
  • plotnine

we could really make this list infinitely long, i would even suggest things like pickle, etc.

[isabelizimm]

In case others are interested: my notes on these libraries while testing

[isabelizimm]

this research spike is considered closed for help pane work, please refer to the spreadsheet for findings