def test_partial_mappings_french_diacritics(self):
"""
Make sure words ending in "icité" are analyzed the same way whether or not there
is an accent.
"""
index_name = "stub_index"
mapping = {"dynamic_templates": MULTILINGUAL_TEXT}
# Create the index and set a mapping that includes the pattern we want to test
ES_INDICES_CLIENT.create(index=index_name)
ES_INDICES_CLIENT.put_mapping(index=index_name, body=mapping)
# The index needs to be closed before we set an analyzer
ES_INDICES_CLIENT.close(index=index_name)
ES_INDICES_CLIENT.put_settings(body=ANALYSIS_SETTINGS,
index=index_name)
ES_INDICES_CLIENT.open(index=index_name)
self.assertEqual(
ES_INDICES_CLIENT.analyze(
body='{"analyzer": "french", "text": "électricité"}',
index=index_name)["tokens"][0]["token"],
"electricit",
)
self.assertEqual(
ES_INDICES_CLIENT.analyze(
body='{"analyzer": "french", "text": "electricite"}',
index=index_name)["tokens"][0]["token"],
"electricit",
)
def test_partial_mappings_code(self):
"""Make sure our code analyzer works as expected."""
index_name = "stub_index"
# Create the index and set a mapping that includes the pattern we want to test
ES_INDICES_CLIENT.create(index=index_name)
# The index needs to be closed before we set an analyzer
ES_INDICES_CLIENT.close(index=index_name)
ES_INDICES_CLIENT.put_settings(body=ANALYSIS_SETTINGS,
index=index_name)
ES_INDICES_CLIENT.open(index=index_name)
self.assertEqual(
[
t["token"] for t in ES_INDICES_CLIENT.analyze(
body='{"analyzer": "code_trigram", "text": "003rst"}',
index=index_name,
)["tokens"]
],
[
"003",
"003r",
"003rs",
"003rst",
"03r",
"03rs",
"03rst",
"3rs",
"3rst",
"rst",
],
)
self.assertEqual(
[
t["token"] for t in ES_INDICES_CLIENT.analyze(
body='{"analyzer": "code", "text": "003rst"}',
index=index_name)["tokens"]
],
["003rst"],
)