mirror of
https://github.com/explosion/spaCy.git
synced 2024-11-12 04:38:28 +03:00
18 lines
573 B
Python
18 lines
573 B
Python
|
# -*- coding: utf-8 -*-
|
||
|
|
|
||
|
|
from __future__ import unicode_literals
|
||
|
|
|
||
|
|
import pytest
|
||
|
|
|
||
|
|
|
||
|
|
@pytest.mark.parametrize('text', ["αριθ.", "τρισ.", "δισ.", "σελ."])
|
||
|
|
def test_tokenizer_handles_abbr(el_tokenizer, text):
|
||
|
|
tokens = el_tokenizer(text)
|
||
|
|
assert len(tokens) == 1
|
||
|
|
|
||
|
|
|
||
|
|
def test_tokenizer_handles_exc_in_text(el_tokenizer):
|
||
|
|
text = "Στα 14 τρισ. δολάρια το κόστος από την άνοδο της στάθμης της θάλασσας."
|
||
|
|
tokens = el_tokenizer(text)
|
||
|
|
assert len(tokens) == 14
|
||
|
|
assert tokens[2].text == "τρισ."
|