as_discrete(levels) generates 'data_meta/series_annotation'.

python-package/lets_plot/mapping.py

@@ -3,7 +3,7 @@
 
 
 class MappingMeta:
- def __init__(self, variable, annotation, **parameters):
+ def __init__(self, variable, annotation, levels, **parameters):
  if variable is None:
  raise ValueError("variable can't be none")
 
@@ -12,10 +12,11 @@ def __init__(self, variable, annotation, **parameters):
 
  self.variable = variable
  self.annotation = annotation
+ self.levels = levels
  self.parameters = parameters
 
 
-def as_discrete(variable, label=None, order_by=None, order=None):
+def as_discrete(variable, label=None, order_by=None, order=None, levels=None):
  """
  The function is used to annotate a numeric data series as categorical data with the possibility of its ordering for the purposes of given visualization.
 
@@ -29,6 +30,8 @@ def as_discrete(variable, label=None, order_by=None, order=None):
  The variable name to order by.
  order : int
  The ordering direction. 1 for ascending, -1 for descending.
+ levels : list
+ The list of values that the variable can take.
 
  Returns
  -------
@@ -108,6 +111,6 @@ def as_discrete(variable, label=None, order_by=None, order=None):
  """
  if isinstance(variable, str):
  label = variable if label is None else label
- return MappingMeta(variable, 'as_discrete', label=label, order_by=order_by, order=order)
+ return MappingMeta(variable, 'as_discrete', levels=levels, label=label, order_by=order_by, order=order)
  # aes(x=as_discrete([1, 2, 3])) - pass as is
  return variable

python-package/lets_plot/plot/util.py

@@ -35,13 +35,21 @@ def as_annotated_data(raw_data: Any, raw_mapping: Any) -> Tuple:
  # mapping
  mapping = {}
  mapping_meta = []
+ # series annotations
+ series_meta = []
 
  if raw_mapping is not None:
  for aesthetic, variable in raw_mapping.as_dict().items():
  if aesthetic == 'name': # ignore FeatureSpec.name property
  continue
 
  if isinstance(variable, MappingMeta):
+ if variable.levels is not None:
+ series_meta.append({
+ 'column': variable.variable,
+ 'factor_levels': variable.levels
+ })
+
  mapping[aesthetic] = variable.variable
  mapping_meta.append({
  'aes': aesthetic,
@@ -54,9 +62,6 @@ def as_annotated_data(raw_data: Any, raw_mapping: Any) -> Tuple:
  if len(mapping_meta) > 0:
  data_meta.update({'mapping_annotations': mapping_meta})
 
- # series annotations
- series_meta = []
-
  if is_dict_or_dataframe(data):
  for column_name, values in data.items():
  if isinstance(values, Iterable):