Merge convert-numeric into encodings

.coveragerc

@@ -2,6 +2,7 @@
 branch = true
 source =
     mplaltair
+omit = *tests*
 
 [report]
 exclude_lines =

mplaltair/__init__.py

@@ -1,35 +1,23 @@
 import matplotlib
 import altair
 
+from ._convert import _convert
 
-# TODO rename this?
-def convert(encoding, *, figure=None):
+def convert(chart):
     """Convert an altair encoding to a Matplotlib figure
 
 
     Parameters
     ----------
-    encoding
-        The Altair encoding of the plot.
-
-    figure : matplotib.figure.Figure, optional
-        # TODO: generalize this to 'thing that supports gridspec slicing?
+    chart
+        The Altair chart object generated by Altair
 
     Returns
     -------
-    figure : matplotlib.figure.Figure
-        The Figure with all artists in it (ready to be saved or shown)
-
     mapping : dict
         Mapping from parts of the encoding to the Matplotlib artists.  This is
         for later customization.
 
 
     """
-    if figure is None:
-        from matplotlib import pyplot as plt
-        figure = plt.figure()
-
-    mapping = {}
-
-    return figure, mapping
+    return _convert(chart)

mplaltair/_convert.py

@@ -1 +1,129 @@
+import matplotlib.dates as mdates
+from ._data import _locate_channel_data, _locate_channel_dtype
 
+def _allowed_ranged_marks(enc_channel, mark):
+    """TODO: DOCS
+    """
+    return mark in ['area', 'bar', 'rect', 'rule'] if enc_channel in ['x2', 'y2'] else True
+
+def _process_x(dtype, data):
+    """Returns the MPL encoding equivalent for Altair x channel
+    """
+    return ('x', data)
+
+
+def _process_y(dtype, data):
+    """Returns the MPL encoding equivalent for Altair y channel
+    """
+    return ('y', data)
+
+
+def _process_x2(dtype, data):
+    """Returns the MPL encoding equivalent for Altair x2 channel
+    """
+    raise NotImplementedError
+
+
+def _process_y2(dtype, data):
+    """Returns the MPL encoding equivalent for Altair y2 channel
+    """
+    raise NotImplementedError
+
+
+def _process_color(dtype, data):
+    """Returns the MPL encoding equivalent for Altair color channel
+    """
+    if dtype == 'quantitative':
+        return ('c', data)
+    elif dtype == 'nominal':
+        raise NotImplementedError
+    elif dtype == 'ordinal':
+        return ('c', data)
+    else:  # temporal
+        return ('c', data)
+
+
+def _process_fill(dtype, data):
+    """Returns the MPL encoding equivalent for Altair fill channel
+    """
+    return _process_color(dtype, data)
+
+
+def _process_shape(dtype, data):
+    """Returns the MPL encoding equivalent for Altair shape channel
+    """
+    raise NotImplementedError
+
+
+def _process_opacity(dtype, data):
+    """Returns the MPL encoding equivalent for Altair opacity channel
+    """
+    raise NotImplementedError
+
+
+def _process_size(dtype, data):
+    """Returns the MPL encoding equivalent for Altair size channel
+    """
+    if dtype == 'quantitative':
+        return ('s', data)
+    elif dtype == 'nominal':
+        raise NotImplementedError
+    elif dtype == 'ordinal':
+        return ('s', data)
+    elif dtype == 'temporal':
+        raise NotImplementedError
+
+
+def _process_stroke(dtype, data):
+    """Returns the MPL encoding equivalent for Altair stroke channel
+    """
+    raise NotImplementedError
+
+_mappings = {
+    'x': _process_x,
+    'y': _process_y,
+    'x2': _process_x2,
+    'y2': _process_y2,
+    'color': _process_color,
+    'fill': _process_fill,
+    'shape': _process_shape,
+    'opacity': _process_opacity,
+    'size': _process_size,
+    'stroke': _process_stroke,
+}
+
+def _convert(chart):
+    """Convert an altair encoding to a Matplotlib figure
+
+
+    Parameters
+    ----------
+    chart
+        The Altair chart.
+
+    Returns
+    -------
+    mapping : dict
+        Mapping from parts of the encoding to the Matplotlib artists.  This is
+        for later customization.
+    """
+    mapping = {}
+
+    if not chart.to_dict().get('encoding'):
+        raise ValueError("Encoding not provided with the chart specification")
+
+    for enc_channel, enc_spec in chart.to_dict()['encoding'].items():
+        if not _allowed_ranged_marks(enc_channel, chart.to_dict()['mark']):
+            raise ValueError("Ranged encoding channels like x2, y2 not allowed for Mark: {}".format(chart['mark']))
+
+    for channel in chart.to_dict()['encoding']:
+        data = _locate_channel_data(chart, channel)
+        dtype = _locate_channel_dtype(chart, channel)
+        if dtype == 'temporal':
+            try:
+                data = mdates.date2num(data)  # Convert dates to Matplotlib dates
+            except AttributeError:
+                raise
+        mapping[_mappings[channel](dtype, data)[0]] = _mappings[channel](dtype, data)[1]
+
+    return mapping

mplaltair/_data.py

@@ -0,0 +1,64 @@
+from ._exceptions import ValidationError
+
+def _locate_channel_dtype(chart, channel):
+    """Locates dtype used for each channel
+        Parameters
+        ----------
+        chart
+            The Altair chart
+        channel
+            The Altair channel being examined
+
+        Returns
+        -------
+        A string representing the data type from the Altair chart ('quantitative', 'ordinal', 'numeric', 'temporal')
+        """
+
+    channel_val = chart.to_dict()['encoding'][channel]
+    if channel_val.get('type'):
+        return channel_val.get('type')
+    else:
+        # TODO: find some way to deal with 'value' so that, opacity, for instance, can be plotted with a value defined
+        if channel_val.get('value'):
+            raise NotImplementedError
+        raise NotImplementedError
+
+
+def _locate_channel_data(chart, channel):
+    """Locates data used for each channel
+
+    Parameters
+    ----------
+    chart
+        The Altair chart
+    channel
+        The Altair channel being examined
+
+    Returns
+    -------
+    A numpy ndarray containing the data used for the channel
+
+    Raises
+    ------
+    ValidationError
+        Raised when the specification does not contain any data attribute
+
+    """
+
+    channel_val = chart.to_dict()['encoding'][channel]
+    if channel_val.get('value'):
+        return channel_val.get('value')
+    elif channel_val.get('aggregate'):
+        return _aggregate_channel()
+    elif channel_val.get('timeUnit'):
+        return _handle_timeUnit()
+    else:  # field is required if the above are not present.
+        return chart.data[channel_val.get('field')].values
+
+
+def _aggregate_channel():
+    raise NotImplementedError
+
+
+def _handle_timeUnit():
+    raise NotImplementedError

mplaltair/_exceptions.py

@@ -0,0 +1,2 @@
+class ValidationError(Exception):
+    pass