elixir-dx 1.0.0

Python wrapper for Data Explorer

go from data to knowledge

ElixirDx

A collaborative notebook platform that enables teams to use and visualize data, together.

Requirements

Python 3.8+

Installation

pip install dx

Then import the package:

import dx

Usage

The dx library currently enables DEX media type visualization of pandas DataFrame and Series objects, as well as numpy ndarray objects. This can be handled in two ways:

• explicit dx.display() calls
• setting the display_mode to update the IPython display formatter for a session

With dx.display()

dx.display() will display a single dataset using the DEX media type. It currently supports:

• pandas DataFrame objects

  import pandas as pd
  import random
  
  df = pd.DataFrame({
      'random_ints': [random.randint(0, 100) for _ in range(500)],
      'random_floats': [random.random() for _ in range(500)],
  })
  dx.display(df)
  

• tabular data as dict or list types

  dx.display([
    [1, 5, 10, 20, 500],
    [1, 2, 3, 4, 5],
    [0, 0, 0, 0, 1]
  ])
  

• .csv or .json filepaths

  df = dx.random_dataframe()
  df.to_csv("dx_docs_sample.csv", index=False)
  
  dx.display("dx_docs_sample.csv")
  

With dx.set_display_mode()

Using either "simple" or "enhanced" display modes will allow dx will update the current IPython display formatters to allow DEX media type visualization of pandas DataFrame objects for an entire notebook / kernel session instead of the default DataFrame display output.

Details

This will adjust pandas options to:

• increasing the number of rows displayed to 50000 from pandas default of 60
• increasing the number of columns displayed to 50 from pandas default of 20
• enabling html.table_schema (False by default in pandas)

This will also handle some basic column cleaning and generate a schema for the DataFrame using pandas.io.json.build_table_schema. Depending on the display mode, the data will be transformed into either a list of dictionaries or list of lists of columnar values.

• "simple" - list of dictionaries
• "enhanced" - list of lists

NOTE: Unlike dx.display(), this only affects pandas DataFrames (or any types set in settings.RENDERABLE_OBJECTS); it does not affect the display of .csv/.json file data, or dict/list outputs

• dx.set_display_mode("simple")

  import dx
  import numpy as np
  import pandas as pd
  
  # enable DEX display outputs from now on
  dx.set_display_mode("simple")
  
  df = pd.read_csv("dx_docs_sample.csv")
  df
  

  df2 = pd.DataFrame(
      [
          [1, 5, 10, 20, 500],
          [1, 2, 3, np.nan, 5],
          [0, 0, 0, np.nan, 1]
      ],
      columns=['a', 'b', 'c', 'd', 'e']
  )
  df2
  

If, at any point, you want to go back to the default display formatting (vanilla pandas output), use the "plain" display mode. This will revert the IPython display format update to its original state and put the pandas options back to their default values.

• dx.set_display_mode("plain")

  # revert to original pandas display outputs from now on
  dx.set_display_mode("plain")
  
  df = pd.read_csv("dx_docs_sample.csv")
  df
  

  df2 = pd.DataFrame(
      [
          [1, 5, 10, 20, 500],
          [1, 2, 3, np.nan, 5],
          [0, 0, 0, np.nan, 1]
      ],
      columns=['a', 'b', 'c', 'd', 'e']
  )
  df2
  

Generating Sample Data

Documentation coming soon!

Contributing

See CONTRIBUTING.md.

Code of Conduct

We follow the ElixirNote code of conduct.

LICENSE

See LICENSE.md.