udpate

ga4gh · Jul 17, 2024 · faedc8a · faedc8a
1 parent d4934cf
commit faedc8a
Show file tree

Hide file tree

Showing 7 changed files with 82 additions and 88 deletions.
diff --git a/notebooks/getting_started/3_Basic_Models.ipynb b/notebooks/getting_started/3_Basic_Models.ipynb
@@ -347,7 +347,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.1"
+   "version": "3.12.2"
   }
  },
  "nbformat": 4,

diff --git a/notebooks/getting_started/4_Exploring_the_AlleleTranslator.ipynb b/notebooks/getting_started/4_Exploring_the_AlleleTranslator.ipynb
@@ -513,7 +513,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.1"
+   "version": "3.12.2"
   }
  },
  "nbformat": 4,

diff --git a/notebooks/getting_started/5_Exploring_the_CnvTranslator.ipynb b/notebooks/getting_started/5_Exploring_the_CnvTranslator.ipynb
@@ -170,16 +170,16 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:CX.0M5VkV5v504_laQURFMEsqzZGcOF9YEw',\n",
-       " 'type': <VrsType.CN_CHANGE: 'CopyNumberChange'>,\n",
+       " 'type': 'CopyNumberChange',\n",
        " 'digest': '0M5VkV5v504_laQURFMEsqzZGcOF9YEw',\n",
        " 'location': {'id': 'ga4gh:SL.GSJAEJXFDz7Nq6VlJj5NTEku48MmteUU',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'GSJAEJXFDz7Nq6VlJj5NTEku48MmteUU',\n",
-       "  'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "  'sequenceReference': {'type': 'SequenceReference',\n",
        "   'refgetAccession': 'SQ.eK4D2MosgK_ivBkgi6FVPg5UXs1bYESm'},\n",
        "  'start': 45002866,\n",
        "  'end': 45015056},\n",
-       " 'copyChange': <CopyChange.EFO_0030067: 'efo:0030067'>}"
+       " 'copyChange': 'efo:0030067'}"
       ]
      },
      "execution_count": 4,
@@ -225,16 +225,16 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:CX.0BN4vrqPrLPAZYsQEAPnG4IS8AYeBGe1',\n",
-       " 'type': <VrsType.CN_CHANGE: 'CopyNumberChange'>,\n",
+       " 'type': 'CopyNumberChange',\n",
        " 'digest': '0BN4vrqPrLPAZYsQEAPnG4IS8AYeBGe1',\n",
        " 'location': {'id': 'ga4gh:SL.tydo6UFL8Y60L5Me3k8AJfljURO9vYn9',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'tydo6UFL8Y60L5Me3k8AJfljURO9vYn9',\n",
-       "  'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "  'sequenceReference': {'type': 'SequenceReference',\n",
        "   'refgetAccession': 'SQ.KEO-4XBcm1cxeo_DIQ8_ofqGUkp4iZhI'},\n",
        "  'start': 75502957,\n",
        "  'end': 76045032},\n",
-       " 'copyChange': <CopyChange.EFO_0030070: 'efo:0030070'>}"
+       " 'copyChange': 'efo:0030070'}"
       ]
      },
      "execution_count": 5,
@@ -292,12 +292,12 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:CN.O_QHImmfErh9jDFkJaypPPvUmnj7EM70',\n",
-       " 'type': <VrsType.CN_COUNT: 'CopyNumberCount'>,\n",
+       " 'type': 'CopyNumberCount',\n",
        " 'digest': 'O_QHImmfErh9jDFkJaypPPvUmnj7EM70',\n",
        " 'location': {'id': 'ga4gh:SL.hBVWalem_rNclxjmUuT9CHbEGCdlqW9L',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'hBVWalem_rNclxjmUuT9CHbEGCdlqW9L',\n",
-       "  'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "  'sequenceReference': {'type': 'SequenceReference',\n",
        "   'refgetAccession': 'SQ.HxuclGHh0XCDuF8x6yQrpHUBL7ZntAHc'},\n",
        "  'start': 85623,\n",
        "  'end': 57073230},\n",
@@ -346,12 +346,12 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:CN.WDzlT9oUq4IcQrVRWGH0dZnARnFBotCS',\n",
-       " 'type': <VrsType.CN_COUNT: 'CopyNumberCount'>,\n",
+       " 'type': 'CopyNumberCount',\n",
        " 'digest': 'WDzlT9oUq4IcQrVRWGH0dZnARnFBotCS',\n",
        " 'location': {'id': 'ga4gh:SL.H1Zh5xdBqamBjwVE9orWdY_uBkpEMH1V',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'H1Zh5xdBqamBjwVE9orWdY_uBkpEMH1V',\n",
-       "  'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "  'sequenceReference': {'type': 'SequenceReference',\n",
        "   'refgetAccession': 'SQ.5ZUqxCmDDgN4xTRbaSjN8LwgZironmB8'},\n",
        "  'start': 46111352,\n",
        "  'end': 46119948},\n",
@@ -385,7 +385,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.1"
+   "version": "3.12.2"
   }
  },
  "nbformat": 4,

diff --git a/notebooks/getting_started/6_Upcoming_features.ipynb b/notebooks/getting_started/6_Upcoming_features.ipynb
@@ -146,17 +146,16 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:VA.LK_4rOVxyEwrEpaOVd-BDFV0ocbO5vgV',\n",
-       " 'type': <VrsType.ALLELE: 'Allele'>,\n",
+       " 'type': 'Allele',\n",
        " 'digest': 'LK_4rOVxyEwrEpaOVd-BDFV0ocbO5vgV',\n",
        " 'location': {'id': 'ga4gh:SL.nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       "  'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "  'sequenceReference': {'type': 'SequenceReference',\n",
        "   'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'},\n",
        "  'start': 80656509,\n",
        "  'end': 80656510},\n",
-       " 'state': {'type': <VrsType.LIT_SEQ_EXPR: 'LiteralSequenceExpression'>,\n",
-       "  'sequence': 'TT'}}"
+       " 'state': {'type': 'LiteralSequenceExpression', 'sequence': 'TT'}}"
       ]
      },
      "execution_count": 4,
@@ -226,9 +225,9 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:SL.nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       " 'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       " 'type': 'SequenceLocation',\n",
        " 'digest': 'nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       " 'sequenceReference': {'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       " 'sequenceReference': {'type': 'SequenceReference',\n",
        "  'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'},\n",
        " 'start': 80656509,\n",
        " 'end': 80656510}"
@@ -261,7 +260,7 @@
     {
      "data": {
       "text/plain": [
-       "{'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       "{'type': 'SequenceReference',\n",
        " 'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'}"
       ]
      },
@@ -293,7 +292,7 @@
      "data": {
       "text/plain": [
        "{'id': None,\n",
-       " 'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
+       " 'type': 'SequenceReference',\n",
        " 'label': None,\n",
        " 'description': None,\n",
        " 'alternativeLabels': None,\n",
@@ -399,9 +398,9 @@
      "data": {
       "text/plain": [
        "{'id': 'refseq:NC_000005.10',\n",
-       " 'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
-       " 'label': 'GRCh38:chr5',\n",
-       " 'alternativeLabels': ['GRCh38:5'],\n",
+       " 'type': 'SequenceReference',\n",
+       " 'label': 'GRCh38:5',\n",
+       " 'alternativeLabels': ['GRCh38:chr5'],\n",
        " 'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'}"
       ]
      },
@@ -435,12 +434,12 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:SL.nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       " 'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       " 'type': 'SequenceLocation',\n",
        " 'digest': 'nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
        " 'sequenceReference': {'id': 'refseq:NC_000005.10',\n",
-       "  'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
-       "  'label': 'GRCh38:chr5',\n",
-       "  'alternativeLabels': ['GRCh38:5'],\n",
+       "  'type': 'SequenceReference',\n",
+       "  'label': 'GRCh38:5',\n",
+       "  'alternativeLabels': ['GRCh38:chr5'],\n",
        "  'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'},\n",
        " 'start': 80656509,\n",
        " 'end': 80656510}"
@@ -465,20 +464,19 @@
      "data": {
       "text/plain": [
        "{'id': 'ga4gh:VA.LK_4rOVxyEwrEpaOVd-BDFV0ocbO5vgV',\n",
-       " 'type': <VrsType.ALLELE: 'Allele'>,\n",
+       " 'type': 'Allele',\n",
        " 'digest': 'LK_4rOVxyEwrEpaOVd-BDFV0ocbO5vgV',\n",
        " 'location': {'id': 'ga4gh:SL.nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
-       "  'type': <VrsType.SEQ_LOC: 'SequenceLocation'>,\n",
+       "  'type': 'SequenceLocation',\n",
        "  'digest': 'nA5-KovovkH-5p3LF1657nkkeWFwrInI',\n",
        "  'sequenceReference': {'id': 'refseq:NC_000005.10',\n",
-       "   'type': <VrsType.SEQ_REF: 'SequenceReference'>,\n",
-       "   'label': 'GRCh38:chr5',\n",
-       "   'alternativeLabels': ['GRCh38:5'],\n",
+       "   'type': 'SequenceReference',\n",
+       "   'label': 'GRCh38:5',\n",
+       "   'alternativeLabels': ['GRCh38:chr5'],\n",
        "   'refgetAccession': 'SQ.aUiQCzCPZ2d0csHbMSbh2NzInhonSXwI'},\n",
        "  'start': 80656509,\n",
        "  'end': 80656510},\n",
-       " 'state': {'type': <VrsType.LIT_SEQ_EXPR: 'LiteralSequenceExpression'>,\n",
-       "  'sequence': 'TT'}}"
+       " 'state': {'type': 'LiteralSequenceExpression', 'sequence': 'TT'}}"
       ]
      },
      "execution_count": 11,
@@ -507,7 +505,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.1"
+   "version": "3.12.2"
   }
  },
  "nbformat": 4,

diff --git a/src/ga4gh/core/domain_models.py b/src/ga4gh/core/domain_models.py
@@ -11,14 +11,13 @@
     module name, e.g., `ga4gh.core.domain_models.Gene`
 """
 from typing import Literal, Union, List
-from enum import Enum
 
 from pydantic import Field, RootModel
 
 from ga4gh.core.entity_models import DomainEntity
 
 
-class CommonDomainType(str, Enum):
+class CommonDomainType:
     """Define GKS Common Domain Entity types"""
 
     PHENOTYPE = "Phenotype"
@@ -33,9 +32,9 @@ class CommonDomainType(str, Enum):
 class Phenotype(DomainEntity):
     """An observable characteristic or trait of an organism."""
 
-    type: Literal[CommonDomainType.PHENOTYPE] = Field(
+    type: Literal["Phenotype"] = Field(
         CommonDomainType.PHENOTYPE,
-        description=f'MUST be "{CommonDomainType.PHENOTYPE.value}".'
+        description=f'MUST be "{CommonDomainType.PHENOTYPE}".'
     )
 
 
@@ -44,18 +43,18 @@ class Disease(DomainEntity):
     of all or part of an organism and is not immediately due to any external injury.
     """
 
-    type: Literal[CommonDomainType.DISEASE] = Field(
+    type: Literal["Disease"] = Field(
         CommonDomainType.DISEASE,
-        description=f'MUST be "{CommonDomainType.DISEASE.value}".'
+        description=f'MUST be "{CommonDomainType.DISEASE}".'
     )
 
 
 class TraitSet(DomainEntity):
     """A set of phenotype and/or disease concepts that together constitute a condition."""
 
-    type: Literal[CommonDomainType.TRAIT_SET] = Field(
+    type: Literal["TraitSet"] = Field(
         CommonDomainType.TRAIT_SET,
-        description=f'MUST be "{CommonDomainType.TRAIT_SET.value}".'
+        description=f'MUST be "{CommonDomainType.TRAIT_SET}".'
     )
     traits: List[Union[Disease, Phenotype]] = Field(
         ...,
@@ -76,27 +75,27 @@ class Condition(RootModel):
 class TherapeuticAction(DomainEntity):
     """A therapeutic action taken that is intended to alter or stop a pathologic process."""
 
-    type: Literal[CommonDomainType.TR_ACTION] = Field(
+    type: Literal["TherapeuticAction"] = Field(
         CommonDomainType.TR_ACTION,
-        description=f'MUST be "{CommonDomainType.TR_ACTION.value}".'
+        description=f'MUST be "{CommonDomainType.TR_ACTION}".'
     )
 
 
 class TherapeuticAgent(DomainEntity):
     """An administered therapeutic agent that is intended to alter or stop a pathologic process."""
 
-    type: Literal[CommonDomainType.TR_AGENT] = Field(
+    type: Literal["TherapeuticAgent"] = Field(
         CommonDomainType.TR_AGENT,
-        description=f'MUST be "{CommonDomainType.TR_AGENT.value}".'
+        description=f'MUST be "{CommonDomainType.TR_AGENT}".'
     )
 
 
 class TherapeuticSubstituteGroup(DomainEntity):
     """A group of therapeutic procedures that may be treated as substitutes for one another."""
 
-    type: Literal[CommonDomainType.TR_SUB] = Field(
+    type: Literal["TherapeuticSubstituteGroup"] = Field(
         CommonDomainType.TR_SUB,
-        description=f'MUST be "{CommonDomainType.TR_SUB.value}".'
+        description=f'MUST be "{CommonDomainType.TR_SUB}".'
     )
     substitutes: List[Union[TherapeuticAction, TherapeuticAgent]] = Field(
         ...,
@@ -110,9 +109,9 @@ class CombinationTherapy(DomainEntity):
     performed in combination.
     """
 
-    type: Literal[CommonDomainType.TR_COMB] = Field(
+    type: Literal["CombinationTherapy"] = Field(
         CommonDomainType.TR_COMB,
-        description=f'MUST be "{CommonDomainType.TR_COMB.value}".'
+        description=f'MUST be "{CommonDomainType.TR_COMB}".'
     )
     components: List[Union[TherapeuticSubstituteGroup, TherapeuticAction, TherapeuticAgent]] = Field(
         ...,
@@ -136,7 +135,7 @@ class TherapeuticProcedure(RootModel):
 class Gene(DomainEntity):
     """A basic physical and functional unit of heredity."""
 
-    type: Literal[CommonDomainType.GENE] = Field(
+    type: Literal["Gene"] = Field(
         CommonDomainType.GENE,
-        description=f'MUST be "{CommonDomainType.GENE.value}".'
+        description=f'MUST be "{CommonDomainType.GENE}".'
     )
diff --git a/src/ga4gh/core/entity_models.py b/src/ga4gh/core/entity_models.py
@@ -14,7 +14,7 @@
 from typing import Any, Dict, Annotated, Optional, Union, List
 from enum import Enum
 
-from pydantic import BaseModel, Field, RootModel, StringConstraints, model_serializer, field_validator
+from pydantic import BaseModel, Field, RootModel, StringConstraints, model_serializer, ConfigDict
 
 from ga4gh.core import GA4GH_IR_REGEXP
 
@@ -119,6 +119,8 @@ class Coding(BaseModel):
 class ConceptMapping(BaseModel):
     """A mapping to a concept in a terminology or code system."""
 
+    model_config = ConfigDict(use_enum_values=True)
+
     coding: Coding = Field(..., description="A structured representation of a code for a defined concept in a terminology or code system.")
     relation: Relation = Field(..., description="A mapping relation between concepts as defined by the Simple Knowledge Organization System (SKOS).")
 
@@ -144,6 +146,8 @@ class Expression(BaseModel):
     variation include the HGVS and ISCN nomenclatures.
     """
 
+    model_config = ConfigDict(use_enum_values=True)
+
     syntax: Syntax = Field(..., description="The syntax used to describe the variation. The value should be one of the supported syntaxes.")
     value: str = Field(..., description="The expression of the variation in the specified syntax. The value should be a valid expression in the specified syntax.")
     syntax_version: Optional[str] = Field(None, description="The version of the syntax used to describe the variation. This is particularly important for HGVS expressions, as the syntax has evolved over time.")
@@ -176,17 +180,6 @@ class Entity(ABC, BaseModel):
     alternativeLabels: Optional[List[str]] = Field(None, description="Alternative name(s) for the Entity.")
     extensions: Optional[List[Extension]] = Field(None, description="A list of extensions to the entity. Extensions are not expected to be natively understood, but may be used for pre-negotiated exchange of message attributes between systems.")
 
-    @field_validator("type")
-    def validate_type(cls, v: str | Enum) -> str:
-        """Ensure that ``type`` field is represented as a string
-
-        :param v: Input value
-        :return: String representation of ``type`` field
-        """
-        if isinstance(v, Enum):
-            v = v.value
-        return v
-
 
 class DomainEntity(Entity, ABC):
     """An Entity that is specific to a particular biomedical domain such as disease,