gsoc

GSOC NOTES:

Use the unit framework to convert the data into categorical numbers Enhance the unit framework to be more dynamic

Categorical ToDo

Close out PR:

0. remove old str tests: tests_units_string_tests (don't need anymore)

1. change name to StrCategoryConvertor, category.py

2. explicitly register all string types on 2 and 3 and throw from future import up top

3. remove .use('agg') from tests

4. make empty list in axis init

5. uses axis_init dict to convert single value

6. update axis_unit & rerender - can't change original order

7. make image test

ax.plot(['a','b'])
ax.plot(['a', 'b', 'c'])
ax.plot(['b', 'c', 'd'])

0. categorical colors

1. sorted categories

2. hierarchical categories

3. mouseover report label by snapping to nearest one

4. Clean up docs and write examples following code guide, can use blog posts as basis,

Other

• scatter plot legend needs background

• Py3.5 dict - matplotlib.cbook - index_of

  • write asSequence wrapper over np.asarray in cbook, factor out np.asarray and replace with asSequence

• Enum (convertor)

• Support for pd.series category (look at support for regular series type) - this is a PR to Pandas

misc notes:

build: export MPLLOCALFREETYPE=1 (& update docs)

Python Framework Bug: pythonw -m "nose" test_categorical.py

labels only exist after rendering (fig.canvas.draw(), plt.draw(), plt.show())