def __init__(self, limit=10000, **options):
Annotation.__init__(self, **options)
self.limit = limit
def __init__(self, description, attributes, inner_attributes=None,
mode='split_workbooks', single_workbook_name=None,
min_tokens=2, max_tokens=3000, **options):
'''
Classify sentences based on some of their attributes, and write
them to XLS files. Lists of sentences will be taken from different
users, distributed as evenly as possible (cf. MixUsers).
sentences:
list of sentences (same as Filter.sentences). [Inherited
from Annotation base class]
description:
a textual description of the annotation. Will appear in a
'meta' sheet of each workbook. Mandatory.
attributes:
a single attribute, or a tuple of attributes, by which
the sentences will be classified into annotation files (the
exact behavior depends on the mode)
inner_attributes:
only applicable for the split workbooks mode. Combinations of
the inner attributes will form the names of sheets within
each workbook
mode:
'single_workbook': one workbook with different sheet for each
attribute combination
'split_workbooks': separate workbook for each attribute combination
'single_sheet': single workbook with a single sheet for all
attribute combinations (but combination will appear in a
dedicated spreadsheet column)
single_workbook_name:
name of single workbook, if applicable
min_tokens:
if an attribute combination has less than this number of occurences
discard this combination (do not write it to annotation files)
max_tokens:
stop writing sentences with any given attribute combination after
this amount of tokens of this combination have been written
Example:
ByAttributeAnnotation("30 to 40 dative verbs, topicalized",
sentences=sentences, attributes=['lemma', 'argument'])
'''
Annotation.__init__(self, description, **options)
self.attributes = attributes
if mode not in ('split_workbooks', 'single_workbook', 'single_sheet'):
raise ValueError('unknown mode "%s"' % mode)
if mode != 'split_workbooks' and inner_attributes:
raise ValueError('inner_attributes can only be set when' \
'mode is "split_workbooks"')
self.mode = mode
self.inner_attributes = inner_attributes
self.single_workbook_name = single_workbook_name
self.min_tokens = min_tokens
self.max_tokens = max_tokens
self.workbooks = {}
def __init__(self, description, subcat_filter, max_sentences=None):
Annotation.__init__(self, description)
self.subcat_filter = subcat_filter
self.max_sentences = max_sentences
