BLUME, Better Looking, Universal, Matplotlib Engine

blume, like bloom, a place where seedlings, orphans looking for a new home, to turn into flowers.

Better Looking?

Better Looking? Really?

Maybe BL stands for something else?

BL big and little?

Universal

Well right now there’s only me here, so that’s hardly Universal.

But the code I am building here blume.magic and blume.farm are aimed at helping me build and run software that is exploring, modelling, simulating, forecasting and displaying information.

Matplotlib

Turns data into pictures.

There is not much that matplotlib cannot do in the world of drawing.

Engine

Run small snippets of code that create an interesting matplotlib figure?

The Farm and its Magic.

The idea with the farm is a collection of Balls you can start and run as you like.

Connections between balls turn them into a graph which networkx handles very nicely.

matplotlib

curio

networkx

Matplotlib and python and other magic wonders

curio

Beginning with Tables

The initial motivation for this

See tabledotpy for an account.

An initial seedling is blume.table.

The initial goal here is to provide an upgrade to the matplotlib.table module.

There is a version here that tries to work around problems in the current module.

There are improvements in the automatic font sizing and cell padding and some new options available.

In amongst all this are some changes to the API the current module (matplotlib>=3.1) module exposes.

There is a good chance if you switch your code to use the table here that everything will work just fine, hopefully with some slightly better looking tables.

So the question remains on how best to transition to the new table?

In particular,

  1. what does the end goal look like and

  2. do we need an interim fix to existing matplotlib table?

It may be too early to decide, but if there is some sort of working precedent to follow that would be great.

As far as the tables are concerned, my plan is to work on a few examples, using tables in a variety of ways and see what ideas come out from the API point of view.

Karma Pi too

A personal tool kit for working on collecting analysing and displaying data. How to do this in a collabortive way?

A collection of 50 or so short python modules that provide some tools for displaying and collecting data. Some aimed at a specific problem or data set, others more general purpose. Random explorations.

It is time to take some of the core ideas in karma pi and re-implement them here.

Data

A time, a place, a value.

And meta data explaining what the value represents.

But data often in rectangular grids with a particular focus on grids which are just projections of grids laid on the surface of a sphere.

And finally

Interactive matplotlib, with async support.

Blume Zen

Everything is perfect, blume no longer exists.

Big and Little.

Evolution of a module

An idea. 2-300 lines of code maybe?

Oh and I can start with foo.py which has half of what I need.

Hack around until it works well enough for current needs.

  1. it works

  2. it turns out to be rather more work than you had hoped

  3. it has created enough new ideas that it has been abandonned for now

  4. realised we have had this problem before and are going round in circles