dataset

Functions to read machine translation evaluation datasets

source

read_source_ref

 read_source_ref (source_path, ref_path)

Read the testset into two arrays and return them


source

download_read_set

 download_read_set (base_path, source_language_code, target_language_code,
                    test_set_name)

Downloads data set if it is not cached. Return source and reference arrays.


source

read_own_set

 read_own_set (base_path, source_language_code, target_language_code,
               test_set_name, date='')

Reads already present non-sacrebleu test set. Return source and reference arrays.


source

get_translated_test_set

 get_translated_test_set (base_path, sourcelang, targetlang, mtengine,
                          test_set_name, test_date, domain='')

Read MT hypothesis translations for specified MT engine


source

read_tsv_set

 read_tsv_set (tsv_file)

Reads complete evaluation set from TSV file containing source, hypothesis and reference