Coverage for bioimageio/spec/dataset/v0_3.py: 98%

1from typing import TYPE_CHECKING, Any, ClassVar, Dict, Literal, Optional, cast 

2 

3from pydantic import model_validator 

4 

5from .._internal.common_nodes import InvalidDescr 

6from .._internal.io import FileDescr as FileDescr 

7from .._internal.io_basics import AbsoluteFilePath as AbsoluteFilePath 

8from .._internal.io_basics import Sha256 as Sha256 

9from .._internal.url import HttpUrl as HttpUrl 

10from ..generic.v0_3 import VALID_COVER_IMAGE_EXTENSIONS as VALID_COVER_IMAGE_EXTENSIONS 

11from ..generic.v0_3 import Author as Author 

12from ..generic.v0_3 import BadgeDescr as BadgeDescr 

13from ..generic.v0_3 import BioimageioConfig as BioimageioConfig 

14from ..generic.v0_3 import CiteEntry as CiteEntry 

15from ..generic.v0_3 import Config as Config 

16from ..generic.v0_3 import DeprecatedLicenseId as DeprecatedLicenseId 

17from ..generic.v0_3 import ( 

18 DocumentationSource, 

19 GenericDescrBase, 

20 LinkedResourceBase, 

21 _author_conv, # pyright: ignore[reportPrivateUsage] 

22 _maintainer_conv, # pyright: ignore[reportPrivateUsage] 

23) 

24from ..generic.v0_3 import Doi as Doi 

25from ..generic.v0_3 import LicenseId as LicenseId 

26from ..generic.v0_3 import LinkedResource as LinkedResource 

27from ..generic.v0_3 import Maintainer as Maintainer 

28from ..generic.v0_3 import OrcidId as OrcidId 

29from ..generic.v0_3 import RelativeFilePath as RelativeFilePath 

30from ..generic.v0_3 import ResourceId as ResourceId 

31from ..generic.v0_3 import Uploader as Uploader 

32from ..generic.v0_3 import Version as Version 

33from .v0_2 import DatasetDescr as DatasetDescr02 

34 

35 

36class DatasetId(ResourceId): 

37 pass 

38 

39 

40class DatasetDescr(GenericDescrBase): 

41 """A bioimage.io dataset resource description file (dataset RDF) describes a dataset relevant to bioimage 

42 processing. 

43 """ 

44 

45 implemented_type: ClassVar[Literal["dataset"]] = "dataset" 

46 if TYPE_CHECKING: 

47 type: Literal["dataset"] = "dataset" 

48 else: 

49 type: Literal["dataset"] 

50 

51 id: Optional[DatasetId] = None 

52 """bioimage.io-wide unique resource identifier 

53 assigned by bioimage.io; version **un**specific.""" 

54 

55 parent: Optional[DatasetId] = None 

56 """The description from which this one is derived""" 

57 

58 source: Optional[HttpUrl] = None 

59 """"URL to the source of the dataset.""" 

60 

61 @model_validator(mode="before") 

62 @classmethod 

63 def _convert(cls, data: Dict[str, Any], /) -> Dict[str, Any]: 

64 if ( 

65 data.get("type") == "dataset" 

66 and isinstance(fv := data.get("format_version"), str) 

67 and fv.startswith("0.2.") 

68 ): 

69 old = DatasetDescr02.load(data) 

70 if isinstance(old, InvalidDescr): 

71 return data 

72 

73 return cast( 

74 Dict[str, Any], 

75 (cls if TYPE_CHECKING else dict)( 

76 attachments=( 

77 [] 

78 if old.attachments is None 

79 else [FileDescr(source=f) for f in old.attachments.files] 

80 ), 

81 authors=[ 

82 _author_conv.convert_as_dict(a) for a in old.authors 

83 ], # pyright: ignore[reportArgumentType] 

84 badges=old.badges, 

85 cite=[ 

86 {"text": c.text, "doi": c.doi, "url": c.url} for c in old.cite 

87 ], # pyright: ignore[reportArgumentType] 

88 config=old.config, # pyright: ignore[reportArgumentType] 

89 covers=old.covers, 

90 description=old.description, 

91 documentation=cast(DocumentationSource, old.documentation), 

92 format_version="0.3.0", 

93 git_repo=old.git_repo, # pyright: ignore[reportArgumentType] 

94 icon=old.icon, 

95 id=None if old.id is None else DatasetId(old.id), 

96 license=old.license, # type: ignore 

97 links=old.links, 

98 maintainers=[ 

99 _maintainer_conv.convert_as_dict(m) for m in old.maintainers 

100 ], # pyright: ignore[reportArgumentType] 

101 name=old.name, 

102 source=old.source, 

103 tags=old.tags, 

104 type=old.type, 

105 uploader=old.uploader, 

106 version=old.version, 

107 **(old.model_extra or {}), 

108 ), 

109 ) 

110 

111 return data 

112 

113 

114class LinkedDataset(LinkedResourceBase): 

115 """Reference to a bioimage.io dataset.""" 

116 

117 id: DatasetId 

118 """A valid dataset `id` from the bioimage.io collection."""