Skip to content

Commit

Permalink
save language_code in dict meta
Browse files Browse the repository at this point in the history
  • Loading branch information
kmike committed Feb 8, 2015
1 parent 8f845a9 commit 2e76fc4
Show file tree
Hide file tree
Showing 3 changed files with 8 additions and 5 deletions.
5 changes: 3 additions & 2 deletions pymorphy2/opencorpora_dict/compile.py
Expand Up @@ -32,7 +32,7 @@


def convert_to_pymorphy2(opencorpora_dict_path, out_path, source_name,
overwrite=False, compile_options=None):
language_code, overwrite=False, compile_options=None):
"""
Convert a dictionary from OpenCorpora XML format to
Pymorphy2 compacted format.
Expand All @@ -50,7 +50,8 @@ def convert_to_pymorphy2(opencorpora_dict_path, out_path, source_name,
parsed_dict = parse_opencorpora_xml(opencorpora_dict_path)
simplify_tags(parsed_dict)
compiled_dict = compile_parsed_dict(parsed_dict, compile_options)
save_compiled_dict(compiled_dict, out_path, source_name=source_name)
save_compiled_dict(compiled_dict, out_path,
source_name=source_name, language_code=language_code)


def compile_parsed_dict(parsed_dict, compile_options=None):
Expand Down
3 changes: 2 additions & 1 deletion pymorphy2/opencorpora_dict/storage.py
Expand Up @@ -76,7 +76,7 @@ def load_dict(path, gramtab_format='opencorpora-int'):
)


def save_compiled_dict(compiled_dict, out_path, source_name):
def save_compiled_dict(compiled_dict, out_path, source_name, language_code):
"""
Save a compiled_dict to ``out_path``
``out_path`` should be a name of folder where to put dictionaries.
Expand Down Expand Up @@ -122,6 +122,7 @@ def _dawg_len(dawg):
prediction_suffixes_dawg_lenghts.append(_dawg_len(prediction_suffixes_dawg))

write_meta(_f('meta.json'), [
['language_code', language_code],
['format_version', CURRENT_FORMAT_VERSION],
['pymorphy2_version', pymorphy2.__version__],
['compiled_at', datetime.datetime.utcnow().isoformat()],
Expand Down
5 changes: 3 additions & 2 deletions tests/test_opencorpora_dict.py
Expand Up @@ -56,8 +56,9 @@ def test_convert_to_pymorphy2(self, tmpdir):
'min_ending_freq': 0,
'paradigm_prefixes': lang.ru.PARADIGM_PREFIXES,
}
convert_to_pymorphy2(self.XML_PATH, out_path, overwrite=True,
compile_options=options, source_name='toy')
convert_to_pymorphy2(self.XML_PATH, out_path,
source_name='toy', language_code='ru',
overwrite=True, compile_options=options)

# use it
morph = pymorphy2.MorphAnalyzer(out_path)
Expand Down

0 comments on commit 2e76fc4

Please sign in to comment.