14903020140

Committed 08 May 2025 09:14AM UTC coverage: 80.091% (-0.05%) from 80.14%

Build # 14903020140

Build Type

Pull #1778

github

Committed by

web-flow

Commit Message

Merge b61299f43 into 2d15f20af

Pull Request Pull Request #1778: Add support to mix args and textual query in load_dataset

Run Details

1642 of 2037 branches covered (80.61%)

Branch coverage included in aggregate %.

10270 of 12836 relevant lines covered (80.01%)

0.8 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

94.59

src/unitxt/dataset_utils.py

from json.decoder import JSONDecodeError
from typing import Any, Dict, Optional

from .artifact import Artifact, UnitxtArtifactNotFoundError, fetch_artifact
from .logging_utils import get_logger
from .parsing_utils import parse_key_equals_value_string_to_dict
from .register import _reset_env_local_catalogs, register_all_artifacts
from .settings_utils import get_settings
from .standard import DatasetRecipe

logger = get_logger()
settings = get_settings()


def fetch(artifact_name: str, overwrite_kwargs: Optional[Dict[str, Any]]=None):
    try:
        artifact, _ = fetch_artifact(artifact_name, overwrite_kwargs=overwrite_kwargs)
        return artifact
    except (UnitxtArtifactNotFoundError, JSONDecodeError):
        return None


def parse(query: str) -> dict:
    return parse_key_equals_value_string_to_dict(query)


def get_dataset_artifact(dataset, overwrite_kwargs: Optional[Dict[str, Any]]=None):
    if isinstance(dataset, DatasetRecipe):
        return dataset
    assert isinstance(
        dataset, str
    ), "dataset should be string description of recipe, or recipe object."
    _reset_env_local_catalogs()
    register_all_artifacts()
    recipe = fetch(dataset, overwrite_kwargs=overwrite_kwargs)
    if recipe is None:
        args = parse(dataset)
        if "__type__" not in args:
            args["__type__"] = settings.default_recipe
        if overwrite_kwargs is not None:
            args.update(overwrite_kwargs)
        recipe = Artifact.from_dict(args)
    return recipe

1	from json.decoder import JSONDecodeError	1✔
2	from typing import Any, Dict, Optional	1✔
3
4	from .artifact import Artifact, UnitxtArtifactNotFoundError, fetch_artifact	1✔
5	from .logging_utils import get_logger	1✔
6	from .parsing_utils import parse_key_equals_value_string_to_dict	1✔
7	from .register import _reset_env_local_catalogs, register_all_artifacts	1✔
8	from .settings_utils import get_settings	1✔
9	from .standard import DatasetRecipe	1✔
10
11	logger = get_logger()	1✔
12	settings = get_settings()	1✔
13
14
15	def fetch(artifact_name: str, overwrite_kwargs: Optional[Dict[str, Any]]=None):	1✔
16	try:	1✔
17	artifact, _ = fetch_artifact(artifact_name, overwrite_kwargs=overwrite_kwargs)	1✔
18	return artifact	×
19	except (UnitxtArtifactNotFoundError, JSONDecodeError):	1✔
20	return None	1✔
21
22
23	def parse(query: str) -> dict:	1✔
24	return parse_key_equals_value_string_to_dict(query)	1✔
25
26
27	def get_dataset_artifact(dataset, overwrite_kwargs: Optional[Dict[str, Any]]=None):	1✔
28	if isinstance(dataset, DatasetRecipe):	1✔
29	return dataset	×
30	assert isinstance(	1✔
31	dataset, str
32	), "dataset should be string description of recipe, or recipe object."
33	_reset_env_local_catalogs()	1✔
34	register_all_artifacts()	1✔
35	recipe = fetch(dataset, overwrite_kwargs=overwrite_kwargs)	1✔
36	if recipe is None:	1✔
37	args = parse(dataset)	1✔
38	if "__type__" not in args:	1✔
39	args["__type__"] = settings.default_recipe	1✔
40	if overwrite_kwargs is not None:	1✔
41	args.update(overwrite_kwargs)	1✔
42	recipe = Artifact.from_dict(args)	1✔
43	return recipe	1✔

IBM / unitxt / 14903020140

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous