Package sentence :: Module parallelsentence :: Class ParallelSentence

Class ParallelSentence

object --+
         |
        ParallelSentence

Known Subclasses:

A parallel sentence, that contains a source sentence, a number of target sentences, a reference and some attributes

Instance Methods

[hide private]

__init__(self, source, translations, reference=None, attributes={}, rank_name='rank', **kwargs)
Constructor @type source SimpleSentence @param source The source text of the parallel sentence @type translations list ( SimpleSentence ) @param translations A list of given translations @type reference SimpleSentence @param reference The desired translation provided by the system @type attributes dict { String name , String value } @param the attributes that describe the parallel sentence source code

__str__(self)
str(x)

source code

__lt__(self, other)

source code

__gt__(self, other)

source code

__eq__(self, other)

source code

get_rank(self)
provide the rank value of the parallel sentence return: the rank value rtype: string

source code

Ranking

get_ranking(self)
returns a ranking list, containing the ranks of the included target translations

source code

dict([(string,string), ...])

get_attributes(self)
provide all attributes

source code

set([string, ...])

get_attribute_names(self)
provide all attribute names

source code

string

get_attribute(self, name)
provide the value of a particular attribute

source code

get_target_attribute_values(self, attribute_name, sub=None)

source code

get_filtered_target_attribute_values(self, attribute_name, filter_attribute_name, filter_attribute_value)

source code

add_attributes(self, attributes)

source code

set_langsrc(self, langsrc)

source code

set_langtgt(self, langtgt)

source code

set_id(self, id)

source code

get_compact_id(self)

source code

get_tuple_id(self)

source code

get_compact_judgment_id(self)

source code

get_judgment_id(self)

source code

has_judgment_id(self)

source code

add_judgment_id(self, value)

source code

get_source(self)

source code

set_source(self, src)

source code

get_translations(self)

source code

set_translations(self, tgt)

source code

get_reference(self)

source code

set_reference(self, ref)

source code

get_nested_attribute_names(self)

source code

get_nested_attributes(self)
function that gathers all the features of the nested sentences to the parallel sentence object, by prefixing their names accordingly

source code

recover_attributes(self)
Moves the attributes back to the nested sentences

source code

serialize(self)

source code

_prefix(self, listitems, prefix)

source code

merge_parallelsentence(self, ps, attribute_replacements={}, **kwargs)
Augment the parallelsentence with another parallesentence. source code

get_pairwise_parallelsentences(self, replacement=True, **kwargs)
Create a set of all available parallel sentence pairs (in tgt) from one ParallelSentence object.

source code

import_indexed_parallelsentence(self, parallelsentence, target_attribute_names, keep_attributes_general=[], keep_attributes_source=[], keep_attributes_target=[]) source code

import_missing_parallelsentence(self, target_attribute_names, keep_attributes_general=[], keep_attributes_source=[], keep_attributes_target=[]) source code

remove_ties(self)
Function that modifies the current parallel sentence by removing the target translations that create ties.

source code

Inherited from object: __delattr__, __format__, __getattribute__, __hash__, __new__, __reduce__, __reduce_ex__, __repr__, __setattr__, __sizeof__, __subclasshook__

Instance Variables

[hide private]

SimpleSentence

ref
a reference translation

SimpleSentence

src
the source sentence

[SimpleSentence, ...]

tgt
a list of target sentences / translations

Properties

[hide private]

Inherited from object: __class__

Method Details

[hide private]

init(self, source, translations, reference=None, attributes=`{}`, rank_name=`'rank'`, **kwargs)
(Constructor)

source code

Constructor @type source SimpleSentence @param source The source text of the parallel sentence @type translations list ( SimpleSentence ) @param translations A list of given translations @type reference SimpleSentence @param reference The desired translation provided by the system @type attributes dict { String name , String value } @param the attributes that describe the parallel sentence

Parameters:

sort_translations (boolean) - Whether translations should be sorted based on the system name

Overrides: object.__init__

str(self)
(Informal representation operator)

source code

str(x)

Overrides: object.__str__: (inherited documentation)

get_ranking(self)

source code

returns a ranking list, containing the ranks of the included target translations

Returns: Ranking: the ranking list

get_attributes(self)

source code

provide all attributes

Returns: dict([(string,string), ...]): the parallel sentence attributes dictionary

get_attribute_names(self)

source code

provide all attribute names

Returns: set([string, ...]): a set with the names of the attributes

get_attribute(self, name)

source code

provide the value of a particular attribute

Returns: string: the value of the attribute with the specified name

merge_parallelsentence(self, ps, attribute_replacements=`{}`, **kwargs)

source code

Augment the parallelsentence with another parallesentence. Merges attributes of source, target and reference sentences and adds target sentences whose system doesn't exist. attributes of target sentences that have a common system.

Parameters:

ps (sentence.parallelsentence.ParallelSentence) - Object of ParallelSentence() with one source sentence and more target sentences
add_missing (boolean) - If translation outputs are missing from the first file but exist in the second, add them (default: True)

get_pairwise_parallelsentences(self, replacement=True, **kwargs)

source code

Create a set of all available parallel sentence pairs (in tgt) from one ParallelSentence object.

Parameters:

ps (sentence.parallelsentence.ParallelSentence
kwargs:
) - Object of ParallelSetnece() with one source sentence and more target sentences
replacement (boolean) - If enabled, creates pairs with all possible combinations with replacement
include_references (boolean) - Include references as system translations from system "_ref" and lowest rank
filter_unassigned (boolean) - If enabled, it filters out pairs with rank = "-1", which means no value was assigned It should not be turned on for test-sets
restrict_ranks ([int, ...]) - Filter pairs to keep only for the ones that include the given ranks. Don't filter if list empty. Before using this, make sure that the ranks are normalized
p (a list of PairwiseParallelSentence() objects)

remove_ties(self)

source code

Function that modifies the current parallel sentence by removing the target translations that create ties. Only first translation for each rank is kept

Class ParallelSentence

__init__(self, source, translations, reference=None, attributes={}, rank_name='rank', **kwargs) (Constructor)

__str__(self) (Informal representation operator)

get_ranking(self)

get_attributes(self)

get_attribute_names(self)

get_attribute(self, name)

merge_parallelsentence(self, ps, attribute_replacements={}, **kwargs)

get_pairwise_parallelsentences(self, replacement=True, **kwargs)

remove_ties(self)

init(self, source, translations, reference=None, attributes=`{}`, rank_name=`'rank'`, **kwargs)
(Constructor)

str(self)
(Informal representation operator)

merge_parallelsentence(self, ps, attribute_replacements=`{}`, **kwargs)