Skip to content

Commit 132c9de

Browse files
fixup! test: create tests for chat_template_base64 field
Signed-off-by: Harikrishnan Balagopal <harikrishmenon@gmail.com>
1 parent e66f8c8 commit 132c9de

2 files changed

Lines changed: 18 additions & 7 deletions

File tree

tests/artifacts/predefined_data_configs/__init__.py

Lines changed: 9 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -44,13 +44,18 @@
4444
PREDEFINED_DATA_CONFIGS, "mt_data_granite_3_1B_tokenize_and_mask_handler.yaml"
4545
)
4646
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_BASE64_HANDLER = os.path.join(
47-
PREDEFINED_DATA_CONFIGS, "mt_data_granite_3_1B_tokenize_and_mask_base64_handler.yaml"
47+
PREDEFINED_DATA_CONFIGS,
48+
"mt_data_granite_3_1B_tokenize_and_mask_base64_handler.yaml",
4849
)
49-
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_INVALID_BASE64_HANDLER = os.path.join(
50-
PREDEFINED_DATA_CONFIGS, "mt_data_granite_3_1B_tokenize_and_mask_invalid_base64_handler.yaml"
50+
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_INVALID_BASE64_HANDLER = (
51+
os.path.join(
52+
PREDEFINED_DATA_CONFIGS,
53+
"mt_data_granite_3_1B_tokenize_and_mask_invalid_base64_handler.yaml",
54+
)
5155
)
5256
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_BASE64_HANDLER_EXPECTED_CHAT_TEMPLATE = os.path.join(
53-
PREDEFINED_DATA_CONFIGS, "mt_data_granite_3_1B_tokenize_and_mask_base64_handler_expected_chat_template.txt"
57+
PREDEFINED_DATA_CONFIGS,
58+
"mt_data_granite_3_1B_tokenize_and_mask_base64_handler_expected_chat_template.txt",
5459
)
5560
DATA_CONFIG_YAML_STREAMING_INPUT_OUTPUT = os.path.join(
5661
PREDEFINED_DATA_CONFIGS, "tokenize_and_apply_input_masking_streaming.yaml"

tests/test_sft_trainer.py

Lines changed: 9 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1400,15 +1400,21 @@ def test_data_config_chat_template_as_base64():
14001400
expected_chat_template_path = DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_BASE64_HANDLER_EXPECTED_CHAT_TEMPLATE
14011401
with open(expected_chat_template_path, "r", encoding="utf-8") as f:
14021402
expected_chat_template = f.read()
1403-
data_config_path = DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_BASE64_HANDLER
1403+
data_config_path = (
1404+
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_BASE64_HANDLER
1405+
)
14041406
assert os.path.isfile(data_config_path)
14051407
data_config = load_and_validate_data_config(data_config_path)
14061408
parsed_chat_template = data_config.dataprocessor.chat_template
14071409
assert parsed_chat_template is not None, "the chat_template wasn't parsed correctly"
1408-
assert data_config.dataprocessor.chat_template == expected_chat_template, "the chat_template wasn't parsed correctly"
1410+
assert (
1411+
data_config.dataprocessor.chat_template == expected_chat_template
1412+
), "the chat_template wasn't parsed correctly"
14091413
# --------------------------------------------
14101414
with pytest.raises(ValueError):
1411-
data_config_path = DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_INVALID_BASE64_HANDLER
1415+
data_config_path = (
1416+
DATA_CONFIG_MULTITURN_CHAT_TOKENIZE_AND_MASKING_DATA_INVALID_BASE64_HANDLER
1417+
)
14121418
data_config = load_and_validate_data_config(data_config_path)
14131419

14141420

0 commit comments

Comments
 (0)