obographs 0.0.1__py3-none-any.whl → 0.0.2__py3-none-any.whl

obographs/model.py

@@ -13,10 +13,15 @@ import json
 import logging
 from collections import defaultdict
 from pathlib import Path
-from typing import Any, Literal, TypeAlias, overload
+from typing import TYPE_CHECKING, Any, Literal, TypeAlias, overload
 
 from pydantic import BaseModel, Field
 
+if TYPE_CHECKING:
+    import curies
+
+    from .standardized import StandardizedGraph
+
 __all__ = [
     "Definition",
     "Edge",
@@ -58,10 +63,10 @@ OBO_SYNONYM_TO_OIO: dict[str, SynonymPredicate] = {
 class Property(BaseModel):
     """Represent a property inside a metadata element."""
 
-    pred: str = Field(...)
-    val: str = Field(
-        ...,
-    )
+    pred: str
+    val: str
+    xrefs: list[str] | None = None
+    meta: Meta | None = None
 
 
 class Definition(BaseModel):
@@ -74,7 +79,7 @@ class Definition(BaseModel):
 class Xref(BaseModel):
     """Represents a cross-reference."""
 
-    val: str = Field(...)
+    val: str
 
 
 class Synonym(BaseModel):
@@ -87,6 +92,7 @@ class Synonym(BaseModel):
         default_factory=list,
         description="A list of CURIEs/IRIs for provenance for the synonym",
     )
+    meta: Meta | None = None
 
 
 class Meta(BaseModel):
@@ -132,11 +138,18 @@ class Graph(BaseModel):
     domainRangeAxioms: list[Any] = Field(default_factory=list)  # noqa:N815
     propertyChainAxioms: list[Any] = Field(default_factory=list)  # noqa:N815
 
+    def standardize(self, converter: curies.Converter) -> StandardizedGraph:
+        """Standardize the graph."""
+        from .standardized import StandardizedGraph
+
+        return StandardizedGraph.from_obograph_raw(self, converter)
+
 
 class GraphDocument(BaseModel):
     """Represents a list of OBO graphs."""
 
     graphs: list[Graph]
+    meta: Meta | None = None
 
 
 def get_id_to_node(graph: Graph) -> dict[str, Node]:

obographs/standardized.py

@@ -0,0 +1,210 @@
+"""Standardize an OBO graph."""
+
+from __future__ import annotations
+
+from curies import Converter, Reference
+from pydantic import BaseModel, Field
+from typing_extensions import Self
+
+from obographs.model import Definition, Edge, Graph, Meta, Node, NodeType, Property, Synonym, Xref
+
+__all__ = [
+    "StandardizedDefinition",
+    "StandardizedEdge",
+    "StandardizedGraph",
+    "StandardizedMeta",
+    "StandardizedNode",
+    "StandardizedXref",
+]
+
+
+class StandardizedProperty(BaseModel):
+    """A standardized property."""
+
+    predicate: Reference
+    value: Reference
+    xrefs: list[Reference] | None = None
+    meta: StandardizedMeta | None = None
+
+    @classmethod
+    def from_obograph_raw(cls, prop: Property, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(
+            predicate=_curie_or_uri_to_ref(prop.pred, converter),
+            value=_curie_or_uri_to_ref(prop.val, converter),
+        )
+
+
+class StandardizedDefinition(BaseModel):
+    """A standardized definition."""
+
+    value: str | None = Field(default=None)
+    xrefs: list[Reference] | None = Field(default=None)
+
+    @classmethod
+    def from_obograph_raw(cls, definition: Definition | None, converter: Converter) -> Self | None:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        if definition is None:
+            return None
+        return cls(
+            value=definition.val,
+            xrefs=_parse_list(definition.xrefs, converter),
+        )
+
+
+class StandardizedXref(BaseModel):
+    """A standardized database cross-reference."""
+
+    reference: Reference
+
+    @classmethod
+    def from_obograph_raw(cls, xref: Xref, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(reference=_curie_or_uri_to_ref(xref.val, converter))
+
+
+class StandardizedSynonym(BaseModel):
+    """A standardized synonym."""
+
+    text: str
+    predicate: Reference
+    type: Reference | None = None
+    xrefs: list[Reference] | None = None
+
+    @classmethod
+    def from_obograph_raw(cls, synonym: Synonym, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(
+            text=synonym.val,
+            predicate=Reference(prefix="oboInOwl", identifier=synonym.pred),
+            type=synonym.synonymType and _curie_or_uri_to_ref(synonym.synonymType, converter),
+            xrefs=_parse_list(synonym.xrefs, converter),
+        )
+
+
+class StandardizedMeta(BaseModel):
+    """A standardized meta object."""
+
+    definition: StandardizedDefinition | None
+    subsets: list[Reference] | None = None
+    xrefs: list[StandardizedXref] | None = None
+    synonyms: list[StandardizedSynonym] | None = None
+    comments: list[str] | None = None
+    deprecated: bool = False
+    version: str | None = None
+    properties: list[StandardizedProperty] | None = None
+
+    @classmethod
+    def from_obograph_raw(cls, meta: Meta | None, converter: Converter) -> Self | None:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        if meta is None:
+            return None
+        return cls(
+            definition=StandardizedDefinition.from_obograph_raw(meta.definition, converter),
+            subsets=[_curie_or_uri_to_ref(subset, converter) for subset in meta.subsets]
+            if meta.subsets
+            else None,
+            xrefs=[StandardizedXref.from_obograph_raw(xref, converter) for xref in meta.xrefs]
+            if meta.xrefs
+            else None,
+            synonyms=[
+                StandardizedSynonym.from_obograph_raw(synonym, converter)
+                for synonym in meta.synonyms
+            ]
+            if meta.synonyms
+            else None,
+            comments=meta.comments,
+            version=meta.version,
+            deprecated=meta.deprecated,
+            properties=[
+                StandardizedProperty.from_obograph_raw(p, converter)
+                for p in meta.basicPropertyValues
+            ]
+            if meta.basicPropertyValues
+            else None,
+        )
+
+
+class StandardizedNode(BaseModel):
+    """A standardized node."""
+
+    reference: Reference
+    label: str | None = Field(None)
+    meta: StandardizedMeta | None = None
+    type: NodeType = Field(..., description="Type of node")
+
+    @classmethod
+    def from_obograph_raw(cls, node: Node, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(
+            reference=_curie_or_uri_to_ref(node.id, converter),
+            label=node.lbl,
+            meta=StandardizedMeta.from_obograph_raw(node.meta, converter),
+            type=node.type,
+        )
+
+
+class StandardizedEdge(BaseModel):
+    """A standardized edge."""
+
+    subject: Reference
+    predicate: Reference
+    object: Reference
+    meta: StandardizedMeta | None = None
+
+    @classmethod
+    def from_obograph_raw(cls, node: Edge, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(
+            subject=_curie_or_uri_to_ref(node.sub, converter),
+            predicate=_curie_or_uri_to_ref(node.pred, converter),
+            object=_curie_or_uri_to_ref(node.obj, converter),
+            meta=StandardizedMeta.from_obograph_raw(node.meta, converter),
+        )
+
+
+class StandardizedGraph(BaseModel):
+    """A standardized graph."""
+
+    id: str | None = None
+    meta: StandardizedMeta | None = None
+    nodes: list[StandardizedNode] = Field(default_factory=list)
+    edges: list[StandardizedEdge] = Field(default_factory=list)
+
+    # TODO other bits
+
+    @classmethod
+    def from_obograph_raw(cls, graph: Graph, converter: Converter) -> Self:
+        """Instantiate by standardizing a raw OBO Graph object."""
+        return cls(
+            id=graph.id,
+            meta=StandardizedMeta.from_obograph_raw(graph.meta, converter),
+            nodes=[StandardizedNode.from_obograph_raw(node, converter) for node in graph.nodes],
+            edges=[StandardizedEdge.from_obograph_raw(edge, converter) for edge in graph.edges],
+        )
+
+
+def _parse_list(ss: list[str] | None, converter: Converter) -> list[Reference] | None:
+    if not ss:
+        return None
+    return [_curie_or_uri_to_ref(x, converter) for x in ss]
+
+
+#: defined in https://github.com/geneontology/obographs/blob/6676b10a5cce04707d75b9dd46fa08de70322b0b/obographs-owlapi/src/main/java/org/geneontology/obographs/owlapi/FromOwl.java#L36-L39
+BUILTINS = {
+    "is_a": Reference(prefix="rdfs", identifier="subClassOf"),
+    "subPropertyOf": Reference(prefix="rdfs", identifier="subPropertyOf"),
+    "type": Reference(prefix="rdf", identifier="type"),
+    "inverseOf": Reference(prefix="owl", identifier="inverseOf"),
+}
+
+
+def _curie_or_uri_to_ref(s: str, converter: Converter) -> Reference:
+    if s in BUILTINS:
+        return BUILTINS[s]
+    if converter.is_uri(s):
+        p, o = converter.parse_uri(s)
+        return Reference(prefix=p, identifier=o)
+    elif converter.is_curie(s):
+        pass
+    raise ValueError(f"can't parse string: {s}")

obographs/version.py

@@ -12,7 +12,7 @@ __all__ = [
     "get_version",
 ]
 
-VERSION = "0.0.1"
+VERSION = "0.0.2"
 
 
 def get_git_hash() -> str:

{obographs-0.0.1.dist-info → obographs-0.0.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: obographs
-Version: 0.0.1
+Version: 0.0.2
 Summary: A python data model for OBO Graphs
 Keywords: snekpack,cookiecutter
 Author: Charles Tapley Hoyt
@@ -23,6 +23,8 @@ Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Typing :: Typed
 Requires-Dist: pydantic
+Requires-Dist: curies
+Requires-Dist: typing-extensions
 Requires-Dist: sphinx>=8 ; extra == 'docs'
 Requires-Dist: sphinx-rtd-theme>=3.0 ; extra == 'docs'
 Requires-Dist: sphinx-automodapi ; extra == 'docs'
@@ -89,11 +91,25 @@ data model.
 import obographs
 
 url = "https://raw.githubusercontent.com/geneontology/obographs/refs/heads/master/examples/abox.json"
-graph_document = obographs.read(url)
+graph_raw = obographs.read(url)
 ```
 
-Note that the OBO Graph JSON schema uses non-pythonic names. The underlying data
-model does not attempt to give better names to the fields.
+The OBO Graph JSON schema uses non-Pythonic names, and it's inherently not aware
+of semantics - it uses a combination of URIs and ad-hoc symbols as identifiers.
+`obographs` implements a standardization workflow that creates new data
+structures with parsed/normalized URIs and symbols that has Pythonic nams. Use
+it like:
+
+```python
+import curies
+
+converter = curies.get_bioregistry_converter()
+
+graph = graph_raw.standardize(converter)
+```
+
+Now, this graph object will have nice Pythonic names and references parsed as
+`curies.Reference` objects.
 
 ## 🚀 Installation
 

obographs-0.0.2.dist-info/RECORD

@@ -0,0 +1,10 @@
+obographs/version.py,sha256=04839f3ad0ad481ebb750f5228278ee8a55b26c17bd61c0a3f2ccc6da50204ae,961
+obographs/__init__.py,sha256=93807a9cc6f4002ac923a84a8c5e1a9b6301c5368cd61a957c9115adfaef1a56,254
+obographs/model.py,sha256=3387390f1a30fe47b22c7c346dfdf2b6a595ccbc7cd8d7d09debf4a13f8f6e6c,6846
+obographs/py.typed,sha256=01ba4719c80b6fe911b091a7c05124b64eeece964e09c058ef8f9805daca546b,1
+obographs/standardized.py,sha256=15a58dcb9af39a01565fdb096389caf2702f94052e933f9368d93caf620022f5,7099
+obographs-0.0.2.dist-info/licenses/LICENSE,sha256=4be0ec343e3bf11fd54321a6b576d5616ebb7d18898f741f63c517209e33bcb2,1076
+obographs-0.0.2.dist-info/WHEEL,sha256=e3765529bb0cc791d07188d72ec6a759d7625ff6d3a5e4b710d25409bae03770,79
+obographs-0.0.2.dist-info/entry_points.txt,sha256=9a9819cedd2186e28d5d42ddce5e3de1417b0db2b07392ff35f9adc7c86a8619,50
+obographs-0.0.2.dist-info/METADATA,sha256=3e577f9dcce8afd6b2b0694935c320b898e5253ba6f0dcf8520b4d1a996ef3a3,13429
+obographs-0.0.2.dist-info/RECORD,,

obographs-0.0.1.dist-info/RECORD

@@ -1,9 +0,0 @@
-obographs/version.py,sha256=0ce3c0b54472caa26dc5eacd65d21c9f7a242a94f5f9ea4d972cf08c2775018b,961
-obographs/__init__.py,sha256=93807a9cc6f4002ac923a84a8c5e1a9b6301c5368cd61a957c9115adfaef1a56,254
-obographs/model.py,sha256=2032a2fb41761be360fbae3280ae5db570c12eb528bd3215c9d2e39494934ada,6441
-obographs/py.typed,sha256=01ba4719c80b6fe911b091a7c05124b64eeece964e09c058ef8f9805daca546b,1
-obographs-0.0.1.dist-info/licenses/LICENSE,sha256=4be0ec343e3bf11fd54321a6b576d5616ebb7d18898f741f63c517209e33bcb2,1076
-obographs-0.0.1.dist-info/WHEEL,sha256=e3765529bb0cc791d07188d72ec6a759d7625ff6d3a5e4b710d25409bae03770,79
-obographs-0.0.1.dist-info/entry_points.txt,sha256=9a9819cedd2186e28d5d42ddce5e3de1417b0db2b07392ff35f9adc7c86a8619,50
-obographs-0.0.1.dist-info/METADATA,sha256=7a7a77f96ddea3e845896d0a37e44ef7a559187f7c01738c77d2fb1a881ffcb9,12972
-obographs-0.0.1.dist-info/RECORD,,