Overview

We need to refine the way users specify colors, to simplify the public API and eliminate inconsistencies. The challenge is to handle all of the ways in which a user might want to specify per-series / per-datum colors, while avoiding potential ambiguities.

Use-cases

Specify one explicit color for everything.
```
  toyplot.plot(x, y, color="red")
```

Specify an explicit color per series.

  toyplot.plot(x, y, color=["red", "cinnamon", "cardinal"])

Specify a palette for the series.

  toyplot.plot(x, y, color=toyplot.color.brewer("reds"))

Specify a colormap for the series.

  toyplot.plot(x, y, color=toyplot.color.LinearMap(toyplot.color.Palette(["red", "green", "blue"])))

Specify per-datum scalar values and have them mapped.
```
  toyplot.plot(x, y, color=temperature)
```

Specify per-datum scalar values and have them mapped, but override a few "special" values.

  colors = toyplot.color.broadcast(temperature)
  colors[numpy.argsort(temperature)[0]] = "blue"
  colors[numpy.argsort(temperature)[-1]] = "red"
  toyplot.plot(x, y, color=colors)

Specify explicit per-datum color values for everything (i.e. map them yourself).
```
  toyplot.plot(x, y, color=mycolors)
```

API

~~(values) - map the values with the default colormap.~~
~~(colormap) - Use the given colormap to map series $[0, N)$, or datum $[0, N)$ if there's only one series.~~
~~(palette) - Use the palette with a linear colormap to map series $[0, N)$, or datum $[0, N)$ if there's only one series.~~
~~(values, colormap) - map values with the given colormap.~~
~~(values, palette) - map values with a linear colormap and the given palette.~~

toyplot.color.broadcast(...) always returns a numpy color array (array with dtype == toyplot.color.dtype), with the given shape.

Values

~~"value" - a single CSS color value.~~
~~(r, g, b) - a single RGB color value.~~
~~(r, g, b, a) - a single RGBA color value.~~
~~toyplot.color.rgb() - a single RGB color value.~~
~~toyplot.color.rgba() - a single RGBA color value.~~
~~[value1, value2, ...] - 1D heterogeneous collection of CSS and tuple color values.~~
- We don't allow scalar values in this case, because their domain would be ambiguous.
- Have to be careful with the implementation here - we don't want a list of nothing-but-tuples to be misinterpreted as a 2D numpy array of scalars.
~~numpy string array with shape $M$ - 1D collection of CSS color values.~~
numpy string array with shape $M \times N$ - 2D collection of CSS color values.
~~numpy numeric array with shape $M$ - 1D collection of scalar values for mapping.~~
~~numpy numeric array with shape $M \times N$ - 2D collection of scalar values for mapping.~~
~~numpy color array with shape $M$ - 1D collection of color values.~~
numpy color array with shape $M \times N$ - 2D collection of color values.



In [1]:

    
import numpy
x = numpy.linspace(0, 1)
y = numpy.column_stack((x ** 2, x ** 3))
numpy.random.seed(1234)
temperature = numpy.random.uniform(size=y.shape)



In [2]:

    
import toyplot



In [3]:

    
toyplot.scatterplot(x, y, width=300);









    




Save as .csv



In [4]:

    
toyplot.scatterplot(x, y, color="red", width=300);









    




Save as .csv



In [5]:

    
toyplot.scatterplot(x, y, color=["red", "blue"], width=300);









    




Save as .csv



In [6]:

    
toyplot.scatterplot(x, y, fill=temperature, width=300);









    




Save as .csv



In [7]:

    
toyplot.scatterplot(x, y, fill=(temperature, toyplot.color.brewer("BlueRed")), width=300);









    




Save as .csv



In [8]:

    
toyplot.scatterplot(x, y, fill=(temperature, toyplot.color.LinearMap()), width=300);









    




Save as .csv



In [9]:

    
colors = toyplot.color.broadcast((temperature, toyplot.color.brewer("Greys")), y.shape)
colors[20:30, 1] = toyplot.color.css("yellow")
colors.flat[numpy.argmin(temperature)] = toyplot.color.css("blue")
colors.flat[numpy.argmax(temperature)] = toyplot.color.css("red")
toyplot.scatterplot(x, y, fill=colors, width=300);









    




Save as .csv



In [10]:

    
x = numpy.linspace(0, 1)
y = x ** 2



In [11]:

    
toyplot.bars(y, fill=(x, toyplot.color.brewer("Reds")), width=300);









    




Save as .csv



In [12]:

    
toyplot.bars(y, fill=toyplot.color.brewer("Reds"), width=300);









    




Save as .csv



In [13]:

    
toyplot.fill(x, numpy.column_stack((y, y/2, y/3, y/4, y/5)), baseline="symmetric", fill=(numpy.arange(5), toyplot.color.brewer("Reds")), width=300);









    




Save as .csv



In [14]:

    
toyplot.fill(x, numpy.column_stack((y, y/2, y/3, y/4, y/5)), baseline="symmetric", fill=toyplot.color.brewer("Reds"), width=300);









    




Save as .csv



In [15]:

    
toyplot.scatterplot(x, numpy.column_stack((y, y/2, y/3, y/4, y/5)), color=(numpy.arange(5), toyplot.color.brewer("Reds")), width=300);









    




Save as .csv



In [16]:

    
toyplot.scatterplot(x, numpy.column_stack((y, y/2, y/3, y/4, y/5)), color=toyplot.color.brewer("Reds"), width=300);









    




Save as .csv



In [17]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.text(x, y, y, fill=(x, toyplot.color.brewer("Reds")));









    




Save as .csv



In [18]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.text(x, y, y, fill=toyplot.color.brewer("Reds"));









    




Save as .csv



In [19]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.hlines(x, stroke=(x, toyplot.color.brewer("Reds")));









    




Save as .csv



In [20]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.hlines(x, stroke=toyplot.color.brewer("Reds"));









    




Save as .csv



In [21]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.vlines(x, stroke=(x, toyplot.color.brewer("Reds")));









    




Save as .csv



In [22]:

    
canvas = toyplot.Canvas(width=300)
axes = canvas.axes()
axes.vlines(x, stroke=toyplot.color.brewer("Reds"));









    




Save as .csv



In [ ]: