pyobo 0.11.2__py3-none-any.whl → 0.12.1__py3-none-any.whl

pyobo/struct/struct.py

@@ -1,111 +1,209 @@
 """Data structures for OBO."""
 
-import gzip
+from __future__ import annotations
+
+import datetime
+import itertools as itt
 import json
 import logging
 import os
 import sys
-from collections import defaultdict
-from collections.abc import Collection, Iterable, Iterator, Mapping, Sequence
+import warnings
+from collections import ChainMap, defaultdict
+from collections.abc import Callable, Collection, Iterable, Iterator, Mapping, Sequence
 from dataclasses import dataclass, field
-from datetime import datetime
-from operator import attrgetter
 from pathlib import Path
 from textwrap import dedent
-from typing import (
-    Any,
-    Callable,
-    ClassVar,
-    Optional,
-    TextIO,
-    Union,
-)
+from typing import Annotated, Any, ClassVar, TextIO
 
 import bioregistry
 import click
+import curies
 import networkx as nx
 import pandas as pd
+import ssslm
+from curies import ReferenceTuple
+from curies import vocabulary as _cv
 from more_click import force_option, verbose_option
 from tqdm.auto import tqdm
-from typing_extensions import Literal
-
-from .reference import Reference, Referenced
-from .typedef import (
-    RelationHint,
-    TypeDef,
-    comment,
-    default_typedefs,
-    exact_match,
-    from_species,
-    get_reference_tuple,
-    has_ontology_root_term,
-    has_part,
-    is_a,
-    orthologous,
-    part_of,
-    see_also,
-    term_replaced_by,
+from typing_extensions import Self
+
+from . import vocabulary as v
+from .reference import (
+    OBOLiteral,
+    Reference,
+    Referenced,
+    _reference_list_tag,
+    comma_separate_references,
+    default_reference,
+    get_preferred_curie,
+    reference_escape,
+    reference_or_literal_to_str,
+)
+from .struct_utils import (
+    Annotation,
+    AnnotationsDict,
+    HasReferencesMixin,
+    IntersectionOfHint,
+    PropertiesHint,
+    ReferenceHint,
+    RelationsHint,
+    Stanza,
+    StanzaType,
+    UnionOfHint,
+    _chain_tag,
+    _ensure_ref,
+    _get_prefixes_from_annotations,
+    _get_references_from_annotations,
+    _tag_property_targets,
 )
-from .utils import comma_separate, obo_escape_slim
+from .utils import _boolean_tag, obo_escape_slim
 from ..api.utils import get_version
 from ..constants import (
+    BUILD_SUBDIRECTORY_NAME,
     DATE_FORMAT,
+    DEFAULT_PREFIX_MAP,
     NCBITAXON_PREFIX,
     RELATION_ID,
     RELATION_PREFIX,
     TARGET_ID,
     TARGET_PREFIX,
 )
-from ..identifier_utils import normalize_curie
-from ..utils.io import multidict, write_iterable_tsv
-from ..utils.misc import obo_to_owl
-from ..utils.path import get_prefix_obo_path, prefix_directory_join
+from ..utils.cache import write_gzipped_graph
+from ..utils.io import multidict, safe_open, write_iterable_tsv
+from ..utils.path import (
+    CacheArtifact,
+    get_cache_path,
+    get_relation_cache_path,
+    prefix_directory_join,
+)
+from ..version import get_version as get_pyobo_version
 
 __all__ = [
+    "Obo",
     "Synonym",
     "SynonymTypeDef",
-    "SynonymSpecificity",
-    "SynonymSpecificities",
     "Term",
-    "Obo",
-    "make_ad_hoc_ontology",
     "abbreviation",
     "acronym",
+    "make_ad_hoc_ontology",
 ]
 
 logger = logging.getLogger(__name__)
 
-SynonymSpecificity = Literal["EXACT", "NARROW", "BROAD", "RELATED"]
-SynonymSpecificities: Sequence[SynonymSpecificity] = ("EXACT", "NARROW", "BROAD", "RELATED")
+#: This is what happens if no specificity is given
+DEFAULT_SPECIFICITY: _cv.SynonymScope = "RELATED"
+
+#: Columns in the SSSOM dataframe
+SSSOM_DF_COLUMNS = [
+    "subject_id",
+    "subject_label",
+    "object_id",
+    "predicate_id",
+    "mapping_justification",
+    "confidence",
+    "contributor",
+]
+UNSPECIFIED_MATCHING_CURIE = "sempav:UnspecifiedMatching"
+FORMAT_VERSION = "1.4"
 
 
 @dataclass
-class Synonym:
+class Synonym(HasReferencesMixin):
     """A synonym with optional specificity and references."""
 
     #: The string representing the synonym
     name: str
 
     #: The specificity of the synonym
-    specificity: SynonymSpecificity = "EXACT"
+    specificity: _cv.SynonymScope | None = None
 
     #: The type of synonym. Must be defined in OBO document!
-    type: "SynonymTypeDef" = field(
-        default_factory=lambda: DEFAULT_SYNONYM_TYPE  # type:ignore
-    )
+    type: Reference | None = None
 
     #: References to articles where the synonym appears
-    provenance: list[Reference] = field(default_factory=list)
+    provenance: Sequence[Reference | OBOLiteral] = field(default_factory=list)
+
+    #: Extra annotations
+    annotations: list[Annotation] = field(default_factory=list)
+
+    #: Language tag for the synonym
+    language: str | None = None
+
+    def __lt__(self, other: Synonym) -> bool:
+        """Sort lexically by name."""
+        return self._sort_key() < other._sort_key()
+
+    def _get_references(self) -> defaultdict[str, set[Reference]]:
+        """Get all prefixes used by the typedef."""
+        rv: defaultdict[str, set[Reference]] = defaultdict(set)
+        rv[v.has_dbxref.prefix].add(v.has_dbxref)
+        if self.type is not None:
+            rv[self.type.prefix].add(self.type)
+        for provenance in self.provenance:
+            match provenance:
+                case Reference():
+                    rv[provenance.prefix].add(provenance)
+                case OBOLiteral(_, datatype, _language):
+                    rv[datatype.prefix].add(v._c(datatype))
+        for prefix, references in _get_references_from_annotations(self.annotations).items():
+            rv[prefix].update(references)
+        return rv
+
+    def _sort_key(self) -> tuple[str, _cv.SynonymScope, str]:
+        return (
+            self.name,
+            self.specificity or DEFAULT_SPECIFICITY,
+            self.type.curie if self.type else "",
+        )
+
+    @property
+    def predicate(self) -> curies.NamedReference:
+        """Get the specificity reference."""
+        return _cv.synonym_scopes[self.specificity or DEFAULT_SPECIFICITY]
 
-    def to_obo(self) -> str:
+    def to_obo(
+        self,
+        ontology_prefix: str,
+        synonym_typedefs: Mapping[ReferenceTuple, SynonymTypeDef] | None = None,
+    ) -> str:
         """Write this synonym as an OBO line to appear in a [Term] stanza."""
-        return f"synonym: {self._fp()}"
+        return f"synonym: {self._fp(ontology_prefix, synonym_typedefs)}"
 
-    def _fp(self) -> str:
-        x = f'"{self._escape(self.name)}" {self.specificity}'
-        if self.type and self.type.pair != DEFAULT_SYNONYM_TYPE.pair:
-            x = f"{x} {self.type.preferred_curie}"
-        return f"{x} [{comma_separate(self.provenance)}]"
+    def _fp(
+        self,
+        ontology_prefix: str,
+        synonym_typedefs: Mapping[ReferenceTuple, SynonymTypeDef] | None = None,
+    ) -> str:
+        if synonym_typedefs is None:
+            synonym_typedefs = {}
+
+        x = f'"{self._escape(self.name)}"'
+
+        # Add on the specificity, e.g., EXACT
+        synonym_typedef = _synonym_typedef_warn(ontology_prefix, self.type, synonym_typedefs)
+        if synonym_typedef is not None and synonym_typedef.specificity is not None:
+            x = f"{x} {synonym_typedef.specificity}"
+        elif self.specificity is not None:
+            x = f"{x} {self.specificity}"
+        elif self.type is not None:
+            # it's not valid to have a synonym type without a specificity,
+            # so automatically assign one if we'll need it
+            x = f"{x} {DEFAULT_SPECIFICITY}"
+
+        # Add on the synonym type, if exists
+        if self.type is not None:
+            x = f"{x} {reference_escape(self.type, ontology_prefix=ontology_prefix)}"
+
+        # the provenance list is required, even if it's empty :/
+        x = f"{x} [{comma_separate_references(self.provenance)}]"
+
+        # OBO flat file format does not support language,
+        # but at least we can mention it here as a comment
+        if self.language:
+            x += f" ! language: {self.language}"
+
+        return x
 
     @staticmethod
     def _escape(s: str) -> str:
@@ -113,113 +211,100 @@ class Synonym:
 
 
 @dataclass
-class SynonymTypeDef(Referenced):
+class SynonymTypeDef(Referenced, HasReferencesMixin):
     """A type definition for synonyms in OBO."""
 
     reference: Reference
-    specificity: Optional[SynonymSpecificity] = None
+    specificity: _cv.SynonymScope | None = None
 
-    def to_obo(self) -> str:
+    def __hash__(self) -> int:
+        # have to re-define hash because of the @dataclass
+        return hash((self.__class__, self.prefix, self.identifier))
+
+    def to_obo(self, ontology_prefix: str) -> str:
         """Serialize to OBO."""
-        rv = f'synonymtypedef: {self.preferred_curie} "{self.name}"'
+        rv = f"synonymtypedef: {reference_escape(self.reference, ontology_prefix=ontology_prefix)}"
+        name = self.name or ""
+        rv = f'{rv} "{name}"'
         if self.specificity:
             rv = f"{rv} {self.specificity}"
         return rv
 
-    @classmethod
-    def from_text(
-        cls,
-        text: str,
-        specificity: Optional[SynonymSpecificity] = None,
-        *,
-        lower: bool = True,
-    ) -> "SynonymTypeDef":
-        """Get a type definition from text that's normalized."""
-        identifier = (
-            text.replace("-", "_")
-            .replace(" ", "_")
-            .replace('"', "")
-            .replace(")", "")
-            .replace("(", "")
-        )
-        if lower:
-            identifier = identifier.lower()
-        return cls(
-            reference=Reference(prefix="obo", identifier=identifier, name=text.replace('"', "")),
-            specificity=specificity,
-        )
+    def _get_references(self) -> dict[str, set[Reference]]:
+        """Get all references used by the typedef."""
+        rv: defaultdict[str, set[Reference]] = defaultdict(set)
+        rv[self.reference.prefix].add(self.reference)
+        if self.specificity is not None:
+            # weird syntax, but this just gets the synonym scope
+            # predicate as a pyobo reference
+            r = v._c(_cv.synonym_scopes[self.specificity])
+            rv[r.prefix].add(r)
+        return dict(rv)
 
 
 DEFAULT_SYNONYM_TYPE = SynonymTypeDef(
-    reference=Reference(prefix="oboInOwl", identifier="SynonymType", name="Synonym"),
+    reference=Reference(prefix="oboInOwl", identifier="SynonymType", name="synonym type"),
 )
 abbreviation = SynonymTypeDef(
     reference=Reference(prefix="OMO", identifier="0003000", name="abbreviation")
 )
 acronym = SynonymTypeDef(reference=Reference(prefix="omo", identifier="0003012", name="acronym"))
-
-
-ReferenceHint = Union[Reference, "Term", tuple[str, str], str]
-
-
-def _ensure_ref(reference: ReferenceHint) -> Reference:
-    if reference is None:
-        raise ValueError("can not append null reference")
-    if isinstance(reference, Term):
-        return reference.reference
-    if isinstance(reference, str):
-        _rv = Reference.from_curie(reference)
-        if _rv is None:
-            raise ValueError(f"could not parse CURIE from {reference}")
-        return _rv
-    if isinstance(reference, tuple):
-        return Reference(prefix=reference[0], identifier=reference[1])
-    if isinstance(reference, Reference):
-        return reference
-    raise TypeError(f"invalid type given for a reference ({type(reference)}): {reference}")
+uk_spelling = SynonymTypeDef(
+    reference=Reference(prefix="omo", identifier="0003005", name="UK spelling synonym")
+)
+default_synonym_typedefs: dict[ReferenceTuple, SynonymTypeDef] = {
+    abbreviation.pair: abbreviation,
+    acronym.pair: acronym,
+    uk_spelling.pair: uk_spelling,
+}
 
 
 @dataclass
-class Term(Referenced):
+class Term(Stanza):
     """A term in OBO."""
 
     #: The primary reference for the entity
     reference: Reference
 
     #: A description of the entity
-    definition: Optional[str] = None
+    definition: str | None = None
 
-    #: References to articles in which the term appears
-    provenance: list[Reference] = field(default_factory=list)
+    #: Object properties
+    relationships: RelationsHint = field(default_factory=lambda: defaultdict(list))
 
-    #: Relationships defined by [Typedef] stanzas
-    relationships: dict[TypeDef, list[Reference]] = field(default_factory=lambda: defaultdict(list))
+    _axioms: AnnotationsDict = field(default_factory=lambda: defaultdict(list))
 
-    #: Properties, which are not defined with Typedef and have scalar values instead of references.
-    properties: dict[str, list[str]] = field(default_factory=lambda: defaultdict(list))
+    properties: PropertiesHint = field(default_factory=lambda: defaultdict(list))
 
     #: Relationships with the default "is_a"
     parents: list[Reference] = field(default_factory=list)
 
+    intersection_of: IntersectionOfHint = field(default_factory=list)
+    union_of: UnionOfHint = field(default_factory=list)
+    equivalent_to: list[Reference] = field(default_factory=list)
+    disjoint_from: list[Reference] = field(default_factory=list)
+
     #: Synonyms of this term
     synonyms: list[Synonym] = field(default_factory=list)
 
-    #: Equivalent references
+    #: Database cross-references, see :func:`get_mappings` for
+    #: access to all mappings in an SSSOM-like interface
     xrefs: list[Reference] = field(default_factory=list)
-    xref_types: list[Reference] = field(default_factory=list)
-
-    #: Alternate Identifiers
-    alt_ids: list[Reference] = field(default_factory=list)
 
     #: The sub-namespace within the ontology
-    namespace: Optional[str] = None
+    namespace: str | None = None
 
     #: An annotation for obsolescence. By default, is None, but this means that it is not obsolete.
-    is_obsolete: Optional[bool] = None
+    is_obsolete: bool | None = None
+
+    type: StanzaType = "Term"
 
-    type: Literal["Term", "Instance"] = "Term"
+    builtin: bool | None = None
+    is_anonymous: bool | None = None
+    subsets: list[Reference] = field(default_factory=list)
 
-    def __hash__(self):
+    def __hash__(self) -> int:
+        # have to re-define hash because of the @dataclass
         return hash((self.__class__, self.prefix, self.identifier))
 
     @classmethod
@@ -227,10 +312,10 @@ class Term(Referenced):
         cls,
         prefix: str,
         identifier: str,
-        name: Optional[str] = None,
-        definition: Optional[str] = None,
+        name: str | None = None,
+        definition: str | None = None,
         **kwargs,
-    ) -> "Term":
+    ) -> Term:
         """Create a term from a reference."""
         return cls(
             reference=Reference(prefix=prefix, identifier=identifier, name=name),
@@ -239,245 +324,198 @@ class Term(Referenced):
         )
 
     @classmethod
-    def auto(
-        cls,
-        prefix: str,
-        identifier: str,
-    ) -> "Term":
-        """Create a term from a reference."""
-        from ..api import get_definition
-
-        return cls(
-            reference=Reference.auto(prefix=prefix, identifier=identifier),
-            definition=get_definition(prefix, identifier),
-        )
-
-    @classmethod
-    def from_curie(cls, curie: str, name: Optional[str] = None) -> "Term":
-        """Create a term directly from a CURIE and optional name."""
-        prefix, identifier = normalize_curie(curie)
-        if prefix is None or identifier is None:
-            raise ValueError
-        return cls.from_triple(prefix=prefix, identifier=identifier, name=name)
-
-    def append_provenance(self, reference: ReferenceHint) -> None:
-        """Add a provenance reference."""
-        self.provenance.append(_ensure_ref(reference))
-
-    def append_synonym(
-        self,
-        synonym: Union[str, Synonym],
-        *,
-        type: Optional[SynonymTypeDef] = None,
-        specificity: Optional[SynonymSpecificity] = None,
-    ) -> None:
-        """Add a synonym."""
-        if isinstance(synonym, str):
-            synonym = Synonym(
-                synonym, type=type or DEFAULT_SYNONYM_TYPE, specificity=specificity or "EXACT"
-            )
-        self.synonyms.append(synonym)
-
-    def append_alt(self, alt: Union[str, Reference]) -> None:
-        """Add an alternative identifier."""
-        if isinstance(alt, str):
-            alt = Reference(prefix=self.prefix, identifier=alt)
-        self.alt_ids.append(alt)
+    def default(cls, prefix, identifier, name=None) -> Self:
+        """Create a default term."""
+        return cls(reference=default_reference(prefix=prefix, identifier=identifier, name=name))
 
-    def append_see_also(self, reference: ReferenceHint) -> "Term":
-        """Add a see also relationship."""
-        self.relationships[see_also].append(_ensure_ref(reference))
-        return self
-
-    def append_comment(self, value: str) -> "Term":
-        """Add a comment relationship."""
-        self.append_property(comment.curie, value)
-        return self
-
-    def append_replaced_by(self, reference: ReferenceHint) -> "Term":
-        """Add a replaced by relationship."""
-        self.append_relationship(term_replaced_by, reference)
-        return self
-
-    def append_parent(self, reference: ReferenceHint) -> "Term":
-        """Add a parent to this entity."""
-        reference = _ensure_ref(reference)
-        if reference not in self.parents:
-            self.parents.append(reference)
-        return self
+    def append_see_also_uri(self, uri: str) -> Self:
+        """Add a see also property."""
+        return self.annotate_uri(v.see_also, uri)
 
     def extend_parents(self, references: Collection[Reference]) -> None:
         """Add a collection of parents to this entity."""
+        warnings.warn("use append_parent", DeprecationWarning, stacklevel=2)
         if any(x is None for x in references):
             raise ValueError("can not append a collection of parents containing a null parent")
         self.parents.extend(references)
 
-    def get_properties(self, prop) -> list[str]:
+    def get_property_literals(self, prop: ReferenceHint) -> list[str]:
         """Get properties from the given key."""
-        return self.properties[prop]
+        return [reference_or_literal_to_str(t) for t in self.properties.get(_ensure_ref(prop), [])]
 
-    def get_property(self, prop) -> Optional[str]:
+    def get_property(self, prop: ReferenceHint) -> str | None:
         """Get a single property of the given key."""
-        r = self.get_properties(prop)
-        if not r:
-            return None
-        if len(r) != 1:
-            raise ValueError
-        return r[0]
-
-    def get_relationship(self, typedef: TypeDef) -> Optional[Reference]:
-        """Get a single relationship of the given type."""
-        r = self.get_relationships(typedef)
+        r = self.get_property_literals(prop)
         if not r:
             return None
         if len(r) != 1:
             raise ValueError
         return r[0]
 
-    def get_relationships(self, typedef: TypeDef) -> list[Reference]:
-        """Get relationships from the given type."""
-        return self.relationships[typedef]
-
-    def append_exact_match(self, reference: ReferenceHint):
+    def append_exact_match(
+        self,
+        reference: ReferenceHint,
+        *,
+        mapping_justification: Reference | None = None,
+        confidence: float | None = None,
+        contributor: Reference | None = None,
+    ) -> Self:
         """Append an exact match, also adding an xref."""
         reference = _ensure_ref(reference)
-        self.append_relationship(exact_match, reference)
-        self.append_xref(reference)
+        axioms = self._prepare_mapping_annotations(
+            mapping_justification=mapping_justification,
+            confidence=confidence,
+            contributor=contributor,
+        )
+        self.annotate_object(v.exact_match, reference, annotations=axioms)
         return self
 
-    def append_xref(self, reference: ReferenceHint) -> None:
-        """Append an xref."""
-        self.xrefs.append(_ensure_ref(reference))
-
-    def append_relationship(self, typedef: TypeDef, reference: ReferenceHint) -> None:
-        """Append a relationship."""
-        self.relationships[typedef].append(_ensure_ref(reference))
-
-    def set_species(self, identifier: str, name: Optional[str] = None):
+    def set_species(self, identifier: str, name: str | None = None) -> Self:
         """Append the from_species relation."""
         if name is None:
             from pyobo.resources.ncbitaxon import get_ncbitaxon_name
 
             name = get_ncbitaxon_name(identifier)
-        self.append_relationship(
-            from_species, Reference(prefix=NCBITAXON_PREFIX, identifier=identifier, name=name)
+        return self.append_relationship(
+            v.from_species, Reference(prefix=NCBITAXON_PREFIX, identifier=identifier, name=name)
         )
 
-    def get_species(self, prefix: str = NCBITAXON_PREFIX) -> Optional[Reference]:
+    def get_species(self, prefix: str = NCBITAXON_PREFIX) -> Reference | None:
         """Get the species if it exists.
 
         :param prefix: The prefix to use in case the term has several species annotations.
         """
-        for species in self.relationships.get(from_species, []):
+        for species in self.get_relationships(v.from_species):
             if species.prefix == prefix:
                 return species
         return None
 
-    def extend_relationship(self, typedef: TypeDef, references: Iterable[Reference]) -> None:
+    def extend_relationship(self, typedef: ReferenceHint, references: Iterable[Reference]) -> None:
         """Append several relationships."""
+        warnings.warn("use append_relationship", DeprecationWarning, stacklevel=2)
         if any(x is None for x in references):
             raise ValueError("can not extend a collection that includes a null reference")
+        typedef = _ensure_ref(typedef)
         self.relationships[typedef].extend(references)
 
-    def append_property(
-        self, prop: Union[str, Reference, Referenced], value: Union[str, Reference, Referenced]
-    ) -> None:
-        """Append a property."""
-        if isinstance(prop, (Reference, Referenced)):
-            prop = prop.preferred_curie
-        if isinstance(value, (Reference, Referenced)):
-            value = value.preferred_curie
-        self.properties[prop].append(value)
-
-    def _definition_fp(self) -> str:
-        if self.definition is None:
-            raise AssertionError
-        return f'"{obo_escape_slim(self.definition)}" [{comma_separate(self.provenance)}]'
-
-    def iterate_relations(self) -> Iterable[tuple[TypeDef, Reference]]:
-        """Iterate over pairs of typedefs and targets."""
-        for typedef, targets in sorted(self.relationships.items(), key=_sort_relations):
-            for target in sorted(targets, key=lambda ref: ref.preferred_curie):
-                yield typedef, target
-
-    def iterate_properties(self) -> Iterable[tuple[str, str]]:
-        """Iterate over pairs of property and values."""
-        for prop, values in sorted(self.properties.items()):
-            for value in sorted(values):
-                yield prop, value
-
-    def iterate_obo_lines(self, *, ontology, typedefs) -> Iterable[str]:
+    def iterate_obo_lines(
+        self,
+        *,
+        ontology_prefix: str,
+        typedefs: Mapping[ReferenceTuple, TypeDef],
+        synonym_typedefs: Mapping[ReferenceTuple, SynonymTypeDef] | None = None,
+        emit_object_properties: bool = True,
+        emit_annotation_properties: bool = True,
+    ) -> Iterable[str]:
         """Iterate over the lines to write in an OBO file."""
         yield f"\n[{self.type}]"
-        yield f"id: {self.preferred_curie}"
-        if self.is_obsolete:
-            yield "is_obsolete: true"
+        # 1
+        yield f"id: {self._reference(self.reference, ontology_prefix)}"
+        # 2
+        yield from _boolean_tag("is_anonymous", self.is_anonymous)
+        # 3
         if self.name:
             yield f"name: {obo_escape_slim(self.name)}"
+        # 4
         if self.namespace and self.namespace != "?":
             namespace_normalized = (
                 self.namespace.replace(" ", "_").replace("-", "_").replace("(", "").replace(")", "")
             )
             yield f"namespace: {namespace_normalized}"
-
+        # 5
+        for alt in sorted(self.alt_ids):
+            yield f"alt_id: {self._reference(alt, ontology_prefix, add_name_comment=True)}"
+        # 6
         if self.definition:
             yield f"def: {self._definition_fp()}"
-
-        for xref in sorted(self.xrefs, key=attrgetter("prefix", "identifier")):
-            yield f"xref: {xref}"  # __str__ bakes in the ! name
-
+        # 7
+        for x in self.get_property_values(v.comment):
+            if isinstance(x, OBOLiteral):
+                yield f'comment: "{x.value}"'
+        # 8
+        yield from _reference_list_tag("subset", self.subsets, ontology_prefix)
+        # 9
+        for synonym in sorted(self.synonyms):
+            yield synonym.to_obo(ontology_prefix=ontology_prefix, synonym_typedefs=synonym_typedefs)
+        # 10
+        yield from self._iterate_xref_obo(ontology_prefix=ontology_prefix)
+        # 11
+        yield from _boolean_tag("builtin", self.builtin)
+        # 12
+        if emit_annotation_properties:
+            yield from self._iterate_obo_properties(
+                ontology_prefix=ontology_prefix,
+                skip_predicate_objects=v.SKIP_PROPERTY_PREDICATES_OBJECTS,
+                skip_predicate_literals=v.SKIP_PROPERTY_PREDICATES_LITERAL,
+                typedefs=typedefs,
+            )
+        # 13
         parent_tag = "is_a" if self.type == "Term" else "instance_of"
-        for parent in sorted(self.parents, key=attrgetter("prefix", "identifier")):
-            yield f"{parent_tag}: {parent}"  # __str__ bakes in the ! name
-
-        for typedef, references in sorted(self.relationships.items(), key=_sort_relations):
-            if (not typedefs or typedef not in typedefs) and (
-                ontology,
-                typedef.curie,
-            ) not in _TYPEDEF_WARNINGS:
-                logger.warning(f"[{ontology}] typedef not defined in OBO: {typedef.curie}")
-                _TYPEDEF_WARNINGS.add((ontology, typedef.curie))
-
-            typedef_preferred_curie = typedef.preferred_curie
-            for reference in sorted(references, key=attrgetter("prefix", "identifier")):
-                s = f"relationship: {typedef_preferred_curie} {reference.preferred_curie}"
-                if typedef.name or reference.name:
-                    s += " !"
-                if typedef.name:
-                    s += f" {typedef.name}"
-                if reference.name:
-                    s += f" {reference.name}"
-                yield s
-
-        for prop, value in sorted(self.iterate_properties(), key=_sort_properties):
-            # TODO deal with typedefs for properties
-            yield f'property_value: {prop} "{value}" xsd:string'  # TODO deal with types later
-
-        for synonym in sorted(self.synonyms, key=attrgetter("name")):
-            yield synonym.to_obo()
-
-    @staticmethod
-    def _escape(s) -> str:
-        return s.replace("\n", "\\n").replace('"', '\\"')
+        yield from _reference_list_tag(parent_tag, self.parents, ontology_prefix)
+        # 14
+        yield from self._iterate_intersection_of_obo(ontology_prefix=ontology_prefix)
+        # 15
+        yield from _reference_list_tag("union_of", self.union_of, ontology_prefix=ontology_prefix)
+        # 16
+        yield from _reference_list_tag(
+            "equivalent_to", self.equivalent_to, ontology_prefix=ontology_prefix
+        )
+        # 17
+        yield from _reference_list_tag(
+            "disjoint_from", self.disjoint_from, ontology_prefix=ontology_prefix
+        )
+        # 18
+        if emit_object_properties:
+            yield from self._iterate_obo_relations(
+                ontology_prefix=ontology_prefix, typedefs=typedefs
+            )
+        # 19 TODO created_by
+        # 20
+        for x in self.get_property_values(v.obo_creation_date):
+            if isinstance(x, OBOLiteral):
+                yield f"creation_date: {x.value}"
+        # 21
+        yield from _boolean_tag("is_obsolete", self.is_obsolete)
+        # 22
+        yield from _tag_property_targets(
+            "replaced_by", self, v.term_replaced_by, ontology_prefix=ontology_prefix
+        )
+        # 23
+        yield from _tag_property_targets(
+            "consider", self, v.see_also, ontology_prefix=ontology_prefix
+        )
 
 
 #: A set of warnings, used to make sure we don't show the same one over and over
-_TYPEDEF_WARNINGS: set[tuple[str, str]] = set()
-
+_SYNONYM_TYPEDEF_WARNINGS: set[tuple[str, Reference]] = set()
 
-def _sort_relations(r):
-    typedef, _references = r
-    return typedef.preferred_curie
 
-
-def _sort_properties(r):
-    o = r[1]
-    if isinstance(o, str):
-        return o
-    elif isinstance(o, Term):
-        return o.curie
-    else:
-        raise TypeError(f"What {type(r)}: {r}")
+def _synonym_typedef_warn(
+    prefix: str,
+    predicate: Reference | None,
+    synonym_typedefs: Mapping[ReferenceTuple, SynonymTypeDef],
+) -> SynonymTypeDef | None:
+    if predicate is None or predicate.pair == DEFAULT_SYNONYM_TYPE.pair:
+        return None
+    if predicate.pair in default_synonym_typedefs:
+        return default_synonym_typedefs[predicate.pair]
+    if predicate.pair in synonym_typedefs:
+        return synonym_typedefs[predicate.pair]
+    key = prefix, predicate
+    if key not in _SYNONYM_TYPEDEF_WARNINGS:
+        _SYNONYM_TYPEDEF_WARNINGS.add(key)
+        predicate_preferred_curie = get_preferred_curie(predicate)
+        if predicate.prefix == "obo":
+            # Throw our hands up in the air. By using `obo` as the prefix,
+            # we already threw using "real" definitions out the window
+            logger.warning(
+                f"[{prefix}] synonym typedef with OBO prefix not defined: {predicate_preferred_curie}."
+                f"\n\tThis might be because you used an unqualified prefix in an OBO file, "
+                f"which automatically gets an OBO prefix."
+            )
+        else:
+            logger.warning(f"[{prefix}] synonym typedef not defined: {predicate_preferred_curie}")
+    return None
 
 
 class BioregistryError(ValueError):
@@ -495,6 +533,9 @@ class BioregistryError(ValueError):
         )
 
 
+LOGGED_MISSING_URI: set[tuple[str, str]] = set()
+
+
 @dataclass
 class Obo:
     """An OBO document."""
@@ -506,22 +547,19 @@ class Obo:
     check_bioregistry_prefix: ClassVar[bool] = True
 
     #: The name of the ontology. If not given, tries looking up with the Bioregistry.
-    name: ClassVar[Optional[str]] = None
-
-    #: The OBO format
-    format_version: ClassVar[str] = "1.2"
+    name: ClassVar[str | None] = None
 
     #: Type definitions
-    typedefs: ClassVar[Optional[list[TypeDef]]] = None
+    typedefs: ClassVar[list[TypeDef] | None] = None
 
     #: Synonym type definitions
-    synonym_typedefs: ClassVar[Optional[list[SynonymTypeDef]]] = None
+    synonym_typedefs: ClassVar[list[SynonymTypeDef] | None] = None
 
     #: An annotation about how an ontology was generated
-    auto_generated_by: ClassVar[Optional[str]] = None
+    auto_generated_by: ClassVar[str | None] = None
 
     #: The idspaces used in the document
-    idspaces: ClassVar[Optional[Mapping[str, str]]] = None
+    idspaces: ClassVar[Mapping[str, str] | None] = None
 
     #: For super-sized datasets that shouldn't be read into memory
     iter_only: ClassVar[bool] = False
@@ -530,28 +568,32 @@ class Obo:
     dynamic_version: ClassVar[bool] = False
 
     #: Set to a static version for the resource (i.e., the resource is not itself versioned)
-    static_version: ClassVar[Optional[str]] = None
+    static_version: ClassVar[str | None] = None
 
-    bioversions_key: ClassVar[Optional[str]] = None
+    bioversions_key: ClassVar[str | None] = None
 
     #: Root terms to use for the ontology
-    root_terms: ClassVar[Optional[list[Reference]]] = None
+    root_terms: ClassVar[list[Reference] | None] = None
 
     #: The date the ontology was generated
-    date: Optional[datetime] = field(default_factory=datetime.today)
+    date: datetime.datetime | None = field(default_factory=datetime.datetime.today)
 
     #: The ontology version
-    data_version: Optional[str] = None
+    data_version: str | None = None
 
     #: Should this ontology be reloaded?
     force: bool = False
 
     #: The hierarchy of terms
-    _hierarchy: Optional[nx.DiGraph] = field(init=False, default=None, repr=False)
+    _hierarchy: nx.DiGraph | None = field(init=False, default=None, repr=False)
     #: A cache of terms
-    _items: Optional[list[Term]] = field(init=False, default=None, repr=False)
+    _items: list[Term] | None = field(init=False, default=None, repr=False)
+
+    subsetdefs: ClassVar[list[tuple[Reference, str]] | None] = None
 
-    term_sort_key: ClassVar[Optional[Callable[["Obo", Term], int]]] = None
+    property_values: ClassVar[list[Annotation] | None] = None
+
+    imports: ClassVar[list[str] | None] = None
 
     def __post_init__(self):
         """Run post-init checks."""
@@ -576,9 +618,85 @@ class Obo:
             elif "/" in self.data_version:
                 raise ValueError(f"{self.ontology} has a slash in version: {self.data_version}")
         if self.auto_generated_by is None:
-            self.auto_generated_by = f"bio2obo:{self.ontology}"  # type:ignore
+            self.auto_generated_by = f"PyOBO v{get_pyobo_version(with_git_hash=True)} on {datetime.datetime.now().isoformat()}"  # type:ignore
+
+    def _get_clean_idspaces(self) -> dict[str, str]:
+        """Get normalized idspace dictionary."""
+        rv = dict(
+            ChainMap(
+                # Add reasonable defaults, most of which are
+                # mandated by the OWL spec anyway (except skos?)
+                DEFAULT_PREFIX_MAP,
+                dict(self.idspaces or {}),
+                # automatically detect all prefixes in reference in the ontology,
+                # then look up Bioregistry-approved URI prefixes
+                self._infer_prefix_map(),
+            )
+        )
+        return rv
 
-    def _get_version(self) -> Optional[str]:
+    def _infer_prefix_map(self) -> dict[str, str]:
+        """Get a prefix map including all prefixes used in the ontology."""
+        rv = {}
+        for prefix in sorted(self._get_prefixes(), key=str.casefold):
+            resource = bioregistry.get_resource(prefix)
+            if resource is None:
+                raise ValueError
+            uri_prefix = resource.get_rdf_uri_prefix()
+            if uri_prefix is None:
+                uri_prefix = resource.get_uri_prefix()
+            if uri_prefix is None:
+                # This allows us an escape hatch, since some
+                # prefixes don't have an associated URI prefix
+                uri_prefix = f"https://bioregistry.io/{prefix}:"
+                if (self.ontology, prefix) not in LOGGED_MISSING_URI:
+                    LOGGED_MISSING_URI.add((self.ontology, prefix))
+                    logger.warning(
+                        "[%s] uses prefix with no URI format: %s. Auto-generating Bioregistry link: %s",
+                        self.ontology,
+                        prefix,
+                        uri_prefix,
+                    )
+
+            pp = bioregistry.get_preferred_prefix(prefix) or str(prefix)
+            rv[pp] = uri_prefix
+        return rv
+
+    def _get_prefixes(self) -> set[str]:
+        """Get all prefixes used by the ontology."""
+        prefixes: set[str] = set(DEFAULT_PREFIX_MAP)
+        for stanza in self._iter_stanzas():
+            prefixes.update(stanza._get_prefixes())
+        for synonym_typedef in self.synonym_typedefs or []:
+            prefixes.update(synonym_typedef._get_prefixes())
+        prefixes.update(subset.prefix for subset, _ in self.subsetdefs or [])
+        # _iterate_property_pairs covers metadata, root terms,
+        # and properties in self.property_values
+        prefixes.update(_get_prefixes_from_annotations(self._iterate_property_pairs()))
+        if self.auto_generated_by:
+            prefixes.add("oboInOwl")
+        return prefixes
+
+    def _get_references(self) -> dict[str, set[Reference]]:
+        """Get all references used by the ontology."""
+        rv: defaultdict[str, set[Reference]] = defaultdict(set)
+
+        for rr in itt.chain(self, self.typedefs or [], self.synonym_typedefs or []):
+            for prefix, references in rr._get_references().items():
+                rv[prefix].update(references)
+        for subset, _ in self.subsetdefs or []:
+            rv[subset.prefix].add(subset)
+        # _iterate_property_pairs covers metadata, root terms,
+        # and properties in self.property_values
+        for prefix, references in _get_references_from_annotations(
+            self._iterate_property_pairs()
+        ).items():
+            rv[prefix].update(references)
+        if self.auto_generated_by:
+            rv[v.obo_autogenerated_by.prefix].add(v.obo_autogenerated_by)
+        return dict(rv)
+
+    def _get_version(self) -> str | None:
         if self.bioversions_key:
             try:
                 return get_version(self.bioversions_key)
@@ -594,6 +712,13 @@ class Obo:
             raise ValueError(f"There is no version available for {self.ontology}")
         return self.data_version
 
+    @property
+    def _prefix_version(self) -> str:
+        """Get the prefix and version (for logging)."""
+        if self.data_version:
+            return f"{self.ontology} {self.data_version}"
+        return self.ontology
+
     def iter_terms(self, force: bool = False) -> Iterable[Term]:
         """Iterate over terms in this ontology."""
         raise NotImplementedError
@@ -604,273 +729,481 @@ class Obo:
 
         return graph_from_obo(self)
 
-    def write_obograph(self, path: Path) -> None:
+    def write_obograph(self, path: str | Path) -> None:
         """Write OBO Graph json."""
         graph = self.get_graph()
-        path.write_text(graph.model_dump_json(indent=2, exclude_none=True, exclude_unset=True))
+        with safe_open(path, read=False) as file:
+            file.write(graph.model_dump_json(indent=2, exclude_none=True, exclude_unset=True))
 
     @classmethod
-    def cli(cls) -> None:
+    def cli(cls, *args, default_rewrite: bool = False) -> Any:
         """Run the CLI for this class."""
-        cli = cls.get_cls_cli()
-        cli()
+        cli = cls.get_cls_cli(default_rewrite=default_rewrite)
+        return cli(*args)
 
     @classmethod
-    def get_cls_cli(cls) -> click.Command:
+    def get_cls_cli(cls, *, default_rewrite: bool = False) -> click.Command:
         """Get the CLI for this class."""
 
         @click.command()
         @verbose_option
         @force_option
-        @click.option("--rewrite", "-r", is_flag=True)
+        @click.option(
+            "--rewrite/--no-rewrite",
+            "-r",
+            default=False,
+            is_flag=True,
+            help="Re-process the data, but don't download it again.",
+        )
         @click.option("--owl", is_flag=True, help="Write OWL via ROBOT")
-        @click.option("--nodes", is_flag=True, help="Write nodes file")
+        @click.option("--ofn", is_flag=True, help="Write Functional OWL (OFN)")
+        @click.option("--ttl", is_flag=True, help="Write turtle RDF via OFN")
         @click.option(
             "--version", help="Specify data version to get. Use this if bioversions is acting up."
         )
-        def _main(force: bool, owl: bool, nodes: bool, version: Optional[str], rewrite: bool):
+        def _main(force: bool, owl: bool, ofn: bool, ttl: bool, version: str | None, rewrite: bool):
+            rewrite = True
             try:
                 inst = cls(force=force, data_version=version)
             except Exception as e:
                 click.secho(f"[{cls.ontology}] Got an exception during instantiation - {type(e)}")
                 sys.exit(1)
-
-            try:
-                inst.write_default(
-                    write_obograph=True,
-                    write_obo=True,
-                    write_owl=owl,
-                    write_nodes=nodes,
-                    force=force or rewrite,
-                    use_tqdm=True,
-                )
-            except Exception as e:
-                click.secho(f"[{cls.ontology}] Got an exception during OBO writing {type(e)}")
-                sys.exit(1)
+            inst.write_default(
+                write_obograph=False,
+                write_obo=False,
+                write_owl=owl,
+                write_ofn=ofn,
+                write_ttl=ttl,
+                write_nodes=True,
+                force=force or rewrite,
+                use_tqdm=True,
+            )
 
         return _main
 
     @property
     def date_formatted(self) -> str:
         """Get the date as a formatted string."""
-        return (self.date if self.date else datetime.now()).strftime(DATE_FORMAT)
+        return (self.date if self.date else datetime.datetime.now()).strftime(DATE_FORMAT)
+
+    def _iter_terms_safe(self) -> Iterator[Term]:
+        if self.iter_only:
+            return iter(self.iter_terms(force=self.force))
+        return iter(self._items_accessor)
 
     def _iter_terms(self, use_tqdm: bool = False, desc: str = "terms") -> Iterable[Term]:
+        yv = self._iter_terms_safe()
         if use_tqdm:
-            total: Optional[int]
+            total: int | None
             try:
                 total = len(self._items_accessor)
             except TypeError:
                 total = None
-            yield from tqdm(self, desc=desc, unit_scale=True, unit="term", total=total)
-        else:
-            yield from self
-
-    def iterate_obo_lines(self) -> Iterable[str]:
-        """Iterate over the lines to write in an OBO file."""
-        yield f"format-version: {self.format_version}"
+            yv = tqdm(yv, desc=desc, unit_scale=True, unit="term", total=total)
+        yield from yv
 
-        if self.auto_generated_by is not None:
-            yield f"auto-generated-by: {self.auto_generated_by}"
+    def _iter_stanzas(self, use_tqdm: bool = False, desc: str = "terms") -> Iterable[Stanza]:
+        yield from self._iter_terms(use_tqdm=use_tqdm, desc=desc)
+        yield from self.typedefs or []
 
-        if self.data_version is not None:
+    def iterate_obo_lines(
+        self,
+        emit_object_properties: bool = True,
+        emit_annotation_properties: bool = True,
+    ) -> Iterable[str]:
+        """Iterate over the lines to write in an OBO file.
+
+        Here's the order:
+
+        1. format-version (technically, this is the only required field)
+        2. data-version
+        3. date
+        4. saved-by
+        5. auto-generated-by
+        6. import
+        7. subsetdef
+        8. synonymtypedef
+        9. default-namespace
+        10. namespace-id-rule
+        11. idspace
+        12. treat-xrefs-as-equivalent
+        13. treat-xrefs-as-genus-differentia
+        14. treat-xrefs-as-relationship
+        15. treat-xrefs-as-is_a
+        16. remark
+        17. ontology
+        """
+        # 1
+        yield f"format-version: {FORMAT_VERSION}"
+        # 2
+        if self.data_version:
             yield f"data-version: {self.data_version}"
-        else:
-            yield f"date: {self.date_formatted}"
+        # 3
+        if self.date:
+            f"date: {self.date_formatted}"
+        # 4 TODO saved-by
+        # 5
+        if self.auto_generated_by:
+            yield f"auto-generated-by: {self.auto_generated_by}"
+        # 6
+        for imp in self.imports or []:
+            yield f"import: {imp}"
+        # 7
+        for subset, subset_remark in self.subsetdefs or []:
+            yield f'subsetdef: {reference_escape(subset, ontology_prefix=self.ontology)} "{subset_remark}"'
+        # 8
+        for synonym_typedef in sorted(self.synonym_typedefs or []):
+            if synonym_typedef.curie == DEFAULT_SYNONYM_TYPE.curie:
+                continue
+            yield synonym_typedef.to_obo(ontology_prefix=self.ontology)
+        # 9 TODO default-namespace
+        # 10 TODO namespace-id-rule
+        # 11
+        for prefix, url in sorted(self._get_clean_idspaces().items()):
+            if prefix in DEFAULT_PREFIX_MAP:
+                # we don't need to write out the 4 default prefixes from
+                # table 2 in https://www.w3.org/TR/owl2-syntax/#IRIs since
+                # they're considered to always be builtin
+                continue
 
-        for prefix, url in sorted((self.idspaces or {}).items()):
-            yield f"idspace: {prefix} {url}"
+            # additional assumptions about built in
+            if prefix in {"obo", "oboInOwl"}:
+                continue
 
-        for synonym_typedef in sorted((self.synonym_typedefs or []), key=attrgetter("curie")):
-            if synonym_typedef.curie == DEFAULT_SYNONYM_TYPE.curie:
+            # ROBOT assumes that all OBO foundry prefixes are builtin,
+            # so don't re-declare them
+            if bioregistry.is_obo_foundry(prefix):
                 continue
-            yield synonym_typedef.to_obo()
 
+            yv = f"idspace: {prefix} {url}"
+            if _yv_name := bioregistry.get_name(prefix):
+                yv += f' "{_yv_name}"'
+            yield yv
+        # 12-15 are handled only during reading, and
+        # PyOBO unmacros things before outputting
+        # 12 treat-xrefs-as-equivalent
+        # 13 treat-xrefs-as-genus-differentia
+        # 14 treat-xrefs-as-relationship
+        # 15 treat-xrefs-as-is_a
+        # 16 TODO remark
+        # 17
         yield f"ontology: {self.ontology}"
+        # 18 (secret)
+        yield from self._iterate_properties()
+
+        typedefs = self._index_typedefs()
+        synonym_typedefs = self._index_synonym_typedefs()
+
+        # PROPERTIES
+        for typedef in sorted(self.typedefs or []):
+            yield from typedef.iterate_obo_lines(
+                ontology_prefix=self.ontology,
+                typedefs=typedefs,
+                synonym_typedefs=synonym_typedefs,
+            )
 
-        if self.name is None:
-            raise ValueError("ontology is missing name")
-        yield f'property_value: http://purl.org/dc/elements/1.1/title "{self.name}" xsd:string'
-        license_spdx_id = bioregistry.get_license(self.ontology)
-        if license_spdx_id:
-            # TODO add SPDX to idspaces and use as a CURIE?
-            yield f'property_value: http://purl.org/dc/terms/license "{license_spdx_id}" xsd:string'
-        description = bioregistry.get_description(self.ontology)
-        if description:
+        # TERMS AND INSTANCES
+        for term in self._iter_terms():
+            yield from term.iterate_obo_lines(
+                ontology_prefix=self.ontology,
+                typedefs=typedefs,
+                synonym_typedefs=synonym_typedefs,
+                emit_object_properties=emit_object_properties,
+                emit_annotation_properties=emit_annotation_properties,
+            )
+
+    def _iterate_properties(self) -> Iterable[str]:
+        for predicate, value in self._iterate_property_pairs():
+            match value:
+                case OBOLiteral():
+                    end = f'"{obo_escape_slim(value.value)}" {reference_escape(value.datatype, ontology_prefix=self.ontology)}'
+                case Reference():
+                    end = reference_escape(value, ontology_prefix=self.ontology)
+            yield f"property_value: {reference_escape(predicate, ontology_prefix=self.ontology)} {end}"
+
+    def _iterate_property_pairs(self) -> Iterable[Annotation]:
+        # Title
+        if self.name:
+            yield Annotation(v.has_title, OBOLiteral.string(self.name))
+
+        # License
+        # TODO add SPDX to idspaces and use as a CURIE?
+        if license_spdx_id := bioregistry.get_license(self.ontology):
+            if license_spdx_id.startswith("http"):
+                license_literal = OBOLiteral.uri(license_spdx_id)
+            else:
+                license_literal = OBOLiteral.string(license_spdx_id)
+            yield Annotation(v.has_license, license_literal)
+
+        # Description
+        if description := bioregistry.get_description(self.ontology):
             description = obo_escape_slim(description.strip())
-            yield f'property_value: http://purl.org/dc/elements/1.1/description "{description}" xsd:string'
+            yield Annotation(v.has_description, OBOLiteral.string(description.strip()))
 
+        # Root terms
         for root_term in self.root_terms or []:
-            yield f"property_value: {has_ontology_root_term.preferred_curie} {root_term.preferred_curie}"
+            yield Annotation(v.has_ontology_root_term, root_term)
+
+        # Extras
+        if self.property_values:
+            yield from self.property_values
 
-        for typedef in sorted(self.typedefs or [], key=attrgetter("curie")):
-            yield from typedef.iterate_obo_lines()
+    def _index_typedefs(self) -> Mapping[ReferenceTuple, TypeDef]:
+        from .typedef import default_typedefs
 
-        for term in self:
-            yield from term.iterate_obo_lines(ontology=self.ontology, typedefs=self.typedefs)
+        return ChainMap(
+            {t.pair: t for t in self.typedefs or []},
+            default_typedefs,
+        )
+
+    def _index_synonym_typedefs(self) -> Mapping[ReferenceTuple, SynonymTypeDef]:
+        return ChainMap(
+            {t.pair: t for t in self.synonym_typedefs or []},
+            default_synonym_typedefs,
+        )
 
     def write_obo(
-        self, file: Union[None, str, TextIO, Path] = None, use_tqdm: bool = False
+        self,
+        file: None | str | TextIO | Path = None,
+        *,
+        use_tqdm: bool = False,
+        emit_object_properties: bool = True,
+        emit_annotation_properties: bool = True,
     ) -> None:
         """Write the OBO to a file."""
-        it = self.iterate_obo_lines()
+        it = self.iterate_obo_lines(
+            emit_object_properties=emit_object_properties,
+            emit_annotation_properties=emit_annotation_properties,
+        )
         if use_tqdm:
-            it = tqdm(it, desc=f"Writing {self.ontology}", unit_scale=True, unit="line")
-        if isinstance(file, (str, Path, os.PathLike)):
-            with open(file, "w") as fh:
+            it = tqdm(
+                it,
+                desc=f"[{self._prefix_version}] writing OBO",
+                unit_scale=True,
+                unit="line",
+            )
+        if isinstance(file, str | Path | os.PathLike):
+            with safe_open(file, read=False) as fh:
                 self._write_lines(it, fh)
         else:
             self._write_lines(it, file)
 
     @staticmethod
-    def _write_lines(it, file: Optional[TextIO]):
+    def _write_lines(it, file: TextIO | None):
         for line in it:
             print(line, file=file)
 
-    def write_obonet_gz(self, path: Union[str, Path]) -> None:
+    def write_obonet_gz(self, path: str | Path) -> None:
         """Write the OBO to a gzipped dump in Obonet JSON."""
         graph = self.to_obonet()
-        with gzip.open(path, "wt") as file:
-            json.dump(nx.node_link_data(graph), file)
+        write_gzipped_graph(path=path, graph=graph)
 
-    def _path(self, *parts: str, name: Optional[str] = None) -> Path:
-        return prefix_directory_join(self.ontology, *parts, name=name, version=self.data_version)
+    def write_ofn(self, path: str | Path) -> None:
+        """Write as Functional OWL (OFN)."""
+        from .functional.obo_to_functional import get_ofn_from_obo
 
-    def _cache(self, *parts: str, name: Optional[str] = None) -> Path:
-        return self._path("cache", *parts, name=name)
+        ofn = get_ofn_from_obo(self)
+        ofn.write_funowl(path)
 
-    @property
-    def _names_path(self) -> Path:
-        return self._cache(name="names.tsv")
+    def write_rdf(self, path: str | Path) -> None:
+        """Write as Turtle RDF."""
+        from .functional.obo_to_functional import get_ofn_from_obo
 
-    @property
-    def _definitions_path(self) -> Path:
-        return self._cache(name="definitions.tsv")
+        ofn = get_ofn_from_obo(self)
+        ofn.write_rdf(path)
 
-    @property
-    def _species_path(self) -> Path:
-        return self._cache(name="species.tsv")
+    def write_nodes(self, path: str | Path) -> None:
+        """Write a nodes TSV file."""
+        write_iterable_tsv(
+            path=path,
+            header=self.nodes_header,
+            it=self.iterate_edge_rows(),
+        )
 
     @property
-    def _synonyms_path(self) -> Path:
-        return self._cache(name="synonyms.tsv")
+    def nodes_header(self) -> Sequence[str]:
+        """Get the header for nodes."""
+        return [
+            "curie:ID",
+            "name:string",
+            "synonyms:string[]",
+            "synonym_predicates:string[]",
+            "synonym_types:string[]",
+            "definition:string",
+            "deprecated:boolean",
+            "type:string",
+            "provenance:string[]",
+            "alts:string[]",
+            "replaced_by:string[]",
+            "mapping_objects:string[]",
+            "mapping_predicates:string[]",
+            "version:string",
+        ]
 
-    @property
-    def _alts_path(self):
-        return self._cache(name="alt_ids.tsv")
+    def _get_node_row(self, node: Term, sep: str, version: str) -> Sequence[str]:
+        synonym_predicate_curies, synonym_type_curies, synonyms = [], [], []
+        for synonym in node.synonyms:
+            synonym_predicate_curies.append(synonym.predicate.curie)
+            synonym_type_curies.append(synonym.type.curie if synonym.type else "")
+            synonyms.append(synonym.name)
+        mapping_predicate_curies, mapping_target_curies = [], []
+        for predicate, obj in node.get_mappings(include_xrefs=True, add_context=False):
+            mapping_predicate_curies.append(predicate.curie)
+            mapping_target_curies.append(obj.curie)
+        return (
+            node.curie,
+            node.name or "",
+            sep.join(synonyms),
+            sep.join(synonym_predicate_curies),
+            sep.join(synonym_type_curies),
+            node.definition or "",
+            "true" if node.is_obsolete else "false",
+            node.type,
+            sep.join(
+                reference.curie for reference in node.provenance if isinstance(reference, Reference)
+            ),
+            sep.join(alt_reference.curie for alt_reference in node.alt_ids),
+            sep.join(ref.curie for ref in node.get_replaced_by()),
+            sep.join(mapping_target_curies),
+            sep.join(mapping_predicate_curies),
+            version,
+        )
 
-    @property
-    def _typedefs_path(self) -> Path:
-        return self._cache(name="typedefs.tsv")
+    def iterate_node_rows(self, sep: str = ";") -> Iterable[Sequence[str]]:
+        """Get a nodes iterator appropriate for serialization."""
+        version = self.data_version or ""
+        for node in self.iter_terms():
+            if node.prefix != self.ontology:
+                continue
+            yield self._get_node_row(node, sep=sep, version=version)
+
+    def write_edges(self, path: str | Path) -> None:
+        """Write a edges TSV file."""
+        # node, this is actually taken care of as part of the cache configuration
+        write_iterable_tsv(
+            path=path,
+            header=self.edges_header,
+            it=self.iterate_edge_rows(),
+        )
 
-    @property
-    def _xrefs_path(self) -> Path:
-        return self._cache(name="xrefs.tsv")
+    def _path(self, *parts: str, name: str | None = None) -> Path:
+        return prefix_directory_join(self.ontology, *parts, name=name, version=self.data_version)
 
-    @property
-    def _relations_path(self) -> Path:
-        return self._cache(name="relations.tsv")
-
-    @property
-    def _properties_path(self) -> Path:
-        return self._cache(name="properties.tsv")
+    def _get_cache_path(self, name: CacheArtifact) -> Path:
+        return get_cache_path(self.ontology, name=name, version=self.data_version)
 
     @property
     def _root_metadata_path(self) -> Path:
         return prefix_directory_join(self.ontology, name="metadata.json")
 
-    @property
-    def _versioned_metadata_path(self) -> Path:
-        return self._cache(name="metadata.json")
-
     @property
     def _obo_path(self) -> Path:
-        return get_prefix_obo_path(self.ontology, version=self.data_version)
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.obo.gz")
 
     @property
     def _obograph_path(self) -> Path:
-        return self._path(name=f"{self.ontology}.json")
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.json.gz")
 
     @property
     def _owl_path(self) -> Path:
-        return self._path(name=f"{self.ontology}.owl")
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.owl.gz")
 
     @property
     def _obonet_gz_path(self) -> Path:
-        return self._path(name=f"{self.ontology}.obonet.json.gz")
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.obonet.json.gz")
 
     @property
-    def _nodes_path(self) -> Path:
-        return self._path(name=f"{self.ontology}.nodes.tsv")
-
-    def write_default(
-        self,
-        use_tqdm: bool = False,
-        force: bool = False,
-        write_obo: bool = False,
-        write_obonet: bool = False,
-        write_obograph: bool = False,
-        write_owl: bool = False,
-        write_nodes: bool = False,
-    ) -> None:
-        """Write the OBO to the default path."""
-        metadata = self.get_metadata()
-        for path in (self._root_metadata_path, self._versioned_metadata_path):
-            logger.debug("[%s v%s] caching metadata to %s", self.ontology, self.data_version, path)
-            with path.open("w") as file:
-                json.dump(metadata, file, indent=2)
+    def _ofn_path(self) -> Path:
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.ofn.gz")
 
-        logger.debug(
-            "[%s v%s] caching typedefs to %s", self.ontology, self.data_version, self._typedefs_path
-        )
-        typedef_df: pd.DataFrame = self.get_typedef_df()
-        typedef_df.sort_values(list(typedef_df.columns), inplace=True)
-        typedef_df.to_csv(self._typedefs_path, sep="\t", index=False)
+    @property
+    def _ttl_path(self) -> Path:
+        return self._path(BUILD_SUBDIRECTORY_NAME, name=f"{self.ontology}.ttl")
 
-        for label, path, header, fn in [
-            ("names", self._names_path, [f"{self.ontology}_id", "name"], self.iterate_id_name),
+    def _get_cache_config(self) -> list[tuple[CacheArtifact, Sequence[str], Callable]]:
+        return [
+            (CacheArtifact.names, [f"{self.ontology}_id", "name"], self.iterate_id_name),
             (
-                "definitions",
-                self._definitions_path,
+                CacheArtifact.definitions,
                 [f"{self.ontology}_id", "definition"],
                 self.iterate_id_definition,
             ),
             (
-                "species",
-                self._species_path,
+                CacheArtifact.species,
                 [f"{self.ontology}_id", "taxonomy_id"],
                 self.iterate_id_species,
             ),
+            (CacheArtifact.alts, [f"{self.ontology}_id", "alt_id"], self.iterate_alt_rows),
+            (CacheArtifact.mappings, SSSOM_DF_COLUMNS, self.iterate_mapping_rows),
+            (CacheArtifact.relations, self.relations_header, self.iter_relation_rows),
+            (CacheArtifact.edges, self.edges_header, self.iterate_edge_rows),
+            (
+                CacheArtifact.object_properties,
+                self.object_properties_header,
+                self.iter_object_properties,
+            ),
+            (
+                CacheArtifact.literal_properties,
+                self.literal_properties_header,
+                self.iter_literal_properties,
+            ),
             (
-                "synonyms",
-                self._synonyms_path,
-                [f"{self.ontology}_id", "synonym"],
-                self.iterate_synonym_rows,
+                CacheArtifact.literal_mappings,
+                ssslm.LiteralMappingTuple._fields,
+                self.iterate_literal_mapping_rows,
             ),
-            ("alts", self._alts_path, [f"{self.ontology}_id", "alt_id"], self.iterate_alt_rows),
-            ("xrefs", self._xrefs_path, self.xrefs_header, self.iterate_xref_rows),
-            ("relations", self._relations_path, self.relations_header, self.iter_relation_rows),
-            ("properties", self._properties_path, self.properties_header, self.iter_property_rows),
-        ]:
-            if path.exists() and not force:
+        ]
+
+    def write_metadata(self) -> None:
+        """Write the metadata JSON file."""
+        metadata = self.get_metadata()
+        for path in (self._root_metadata_path, self._get_cache_path(CacheArtifact.metadata)):
+            logger.debug("[%s] caching metadata to %s", self._prefix_version, path)
+            with safe_open(path, read=False) as file:
+                json.dump(metadata, file, indent=2)
+
+    def write_prefix_map(self) -> None:
+        """Write a prefix map file that includes all prefixes used in this ontology."""
+        with self._get_cache_path(CacheArtifact.prefixes).open("w") as file:
+            json.dump(self._get_clean_idspaces(), file, indent=2)
+
+    def write_cache(self, *, force: bool = False) -> None:
+        """Write cache parts."""
+        typedefs_path = self._get_cache_path(CacheArtifact.typedefs)
+        logger.debug(
+            "[%s] caching typedefs to %s",
+            self._prefix_version,
+            typedefs_path,
+        )
+        typedef_df: pd.DataFrame = self.get_typedef_df()
+        typedef_df.sort_values(list(typedef_df.columns), inplace=True)
+        typedef_df.to_csv(typedefs_path, sep="\t", index=False)
+
+        for cache_artifact, header, fn in self._get_cache_config():
+            path = self._get_cache_path(cache_artifact)
+            if path.is_file() and not force:
                 continue
-            logger.debug("[%s v%s] caching %s to %s", self.ontology, self.data_version, label, path)
+            tqdm.write(
+                f"[{self._prefix_version}] writing {cache_artifact.name} to {path}",
+            )
             write_iterable_tsv(
                 path=path,
                 header=header,
                 it=fn(),  # type:ignore
             )
 
-        for relation in (is_a, has_part, part_of, from_species, orthologous):
-            if relation is not is_a and self.typedefs is not None and relation not in self.typedefs:
+        typedefs = self._index_typedefs()
+        for relation in (v.is_a, v.has_part, v.part_of, v.from_species, v.orthologous):
+            if relation is not v.is_a and relation.pair not in typedefs:
                 continue
-            relations_path = self._cache("relations", name=f"{relation.curie}.tsv")
-            if relations_path.exists() and not force:
+            relations_path = get_relation_cache_path(
+                self.ontology, reference=relation, version=self.data_version
+            )
+            if relations_path.is_file() and not force:
                 continue
             logger.debug(
-                "[%s v%s] caching relation %s ! %s",
-                self.ontology,
-                self.data_version,
+                "[%s] caching relation %s ! %s",
+                self._prefix_version,
                 relation.curie,
                 relation.name,
             )
@@ -880,36 +1213,82 @@ class Obo:
             relation_df.sort_values(list(relation_df.columns), inplace=True)
             relation_df.to_csv(relations_path, sep="\t", index=False)
 
-        if (write_obo or write_owl) and (not self._obo_path.exists() or force):
+    def write_default(
+        self,
+        use_tqdm: bool = False,
+        force: bool = False,
+        write_obo: bool = False,
+        write_obonet: bool = False,
+        write_obograph: bool = False,
+        write_owl: bool = False,
+        write_ofn: bool = False,
+        write_ttl: bool = False,
+        write_nodes: bool = False,
+        obograph_use_internal: bool = False,
+        write_cache: bool = True,
+    ) -> None:
+        """Write the OBO to the default path."""
+        self.write_metadata()
+        self.write_prefix_map()
+        if write_cache:
+            self.write_cache(force=force)
+        if write_obo and (not self._obo_path.is_file() or force):
+            tqdm.write(f"[{self._prefix_version}] writing OBO to {self._obo_path}")
             self.write_obo(self._obo_path, use_tqdm=use_tqdm)
-        if write_obograph and (not self._obograph_path.exists() or force):
-            self.write_obograph(self._obograph_path)
-        if write_owl and (not self._owl_path.exists() or force):
-            obo_to_owl(self._obo_path, self._owl_path)
-        if write_obonet and (not self._obonet_gz_path.exists() or force):
-            logger.debug("writing obonet to %s", self._obonet_gz_path)
+        if (write_ofn or write_owl or write_obograph) and (not self._ofn_path.is_file() or force):
+            tqdm.write(f"[{self._prefix_version}] writing OFN to {self._ofn_path}")
+            self.write_ofn(self._ofn_path)
+        if write_obograph and (not self._obograph_path.is_file() or force):
+            if obograph_use_internal:
+                tqdm.write(f"[{self._prefix_version}] writing OBO Graph to {self._obograph_path}")
+                self.write_obograph(self._obograph_path)
+            else:
+                import bioontologies.robot
+
+                tqdm.write(
+                    f"[{self.ontology}] converting OFN to OBO Graph at {self._obograph_path}"
+                )
+                bioontologies.robot.convert(
+                    self._ofn_path, self._obograph_path, debug=True, merge=False, reason=False
+                )
+        if write_owl and (not self._owl_path.is_file() or force):
+            tqdm.write(f"[{self._prefix_version}] writing OWL to {self._owl_path}")
+            import bioontologies.robot
+
+            bioontologies.robot.convert(
+                self._ofn_path, self._owl_path, debug=True, merge=False, reason=False
+            )
+        if write_ttl and (not self._ttl_path.is_file() or force):
+            tqdm.write(f"[{self._prefix_version}] writing Turtle to {self._ttl_path}")
+            self.write_rdf(self._ttl_path)
+        if write_obonet and (not self._obonet_gz_path.is_file() or force):
+            tqdm.write(f"[{self._prefix_version}] writing obonet to {self._obonet_gz_path}")
             self.write_obonet_gz(self._obonet_gz_path)
         if write_nodes:
-            self.get_graph().get_nodes_df().to_csv(self._nodes_path, sep="\t", index=False)
+            nodes_path = self._get_cache_path(CacheArtifact.nodes)
+            tqdm.write(f"[{self._prefix_version}] writing nodes TSV to {nodes_path}")
+            self.write_nodes(nodes_path)
 
     @property
-    def _items_accessor(self):
+    def _items_accessor(self) -> list[Term]:
         if self._items is None:
-            key = self.term_sort_key or attrgetter("curie")
-            self._items = sorted(self.iter_terms(force=self.force), key=key)
+            # if the term sort key is None, then the terms get sorted by their reference
+            self._items = sorted(
+                self.iter_terms(force=self.force),
+            )
         return self._items
 
-    def __iter__(self) -> Iterator["Term"]:
-        if self.iter_only:
-            return iter(self.iter_terms(force=self.force))
-        return iter(self._items_accessor)
+    def __iter__(self) -> Iterator[Term]:
+        yield from self._iter_terms_safe()
 
     def ancestors(self, identifier: str) -> set[str]:
         """Return a set of identifiers for parents of the given identifier."""
+        # FIXME switch to references
         return nx.descendants(self.hierarchy, identifier)  # note this is backwards
 
     def descendants(self, identifier: str) -> set[str]:
         """Return a set of identifiers for the children of the given identifier."""
+        # FIXME switch to references
         return nx.ancestors(self.hierarchy, identifier)  # note this is backwards
 
     def is_descendant(self, descendant: str, ancestor: str) -> bool:
@@ -917,9 +1296,9 @@ class Obo:
 
         .. code-block:: python
 
-            from pyobo import get_obo
+            from pyobo import get_ontology
 
-            obo = get_obo("go")
+            obo = get_ontology("go")
 
             interleukin_10_complex = "1905571"  # interleukin-10 receptor complex
             all_complexes = "0032991"
@@ -935,21 +1314,22 @@ class Obo:
 
         .. code-block:: python
 
-            from pyobo import get_obo
+            from pyobo import get_ontology
 
-            obo = get_obo("go")
+            obo = get_ontology("go")
 
             identifier = "1905571"  # interleukin-10 receptor complex
             is_complex = "0032991" in nx.descendants(obo.hierarchy, identifier)  # should be true
         """
         if self._hierarchy is None:
             self._hierarchy = nx.DiGraph()
-            for term in self._iter_terms(desc=f"[{self.ontology}] getting hierarchy"):
-                for parent in term.parents:
-                    self._hierarchy.add_edge(term.identifier, parent.identifier)
+            for stanza in self._iter_stanzas(desc=f"[{self.ontology}] getting hierarchy"):
+                for parent in stanza.parents:
+                    # FIXME add referneces
+                    self._hierarchy.add_edge(stanza.identifier, parent.identifier)
         return self._hierarchy
 
-    def to_obonet(self: "Obo", *, use_tqdm: bool = False) -> nx.MultiDiGraph:
+    def to_obonet(self: Obo, *, use_tqdm: bool = False) -> nx.MultiDiGraph:
         """Export as a :mod`obonet` style graph."""
         rv = nx.MultiDiGraph()
         rv.graph.update(
@@ -957,55 +1337,64 @@ class Obo:
                 "name": self.name,
                 "ontology": self.ontology,
                 "auto-generated-by": self.auto_generated_by,
-                "typedefs": _convert_typedefs(self.typedefs),
-                "format-version": self.format_version,
+                "format-version": FORMAT_VERSION,
                 "data-version": self.data_version,
-                "synonymtypedef": _convert_synonym_typedefs(self.synonym_typedefs),
                 "date": self.date_formatted,
+                "typedefs": [typedef.reference.model_dump() for typedef in self.typedefs or []],
+                "synonymtypedef": [
+                    synonym_typedef.to_obo(ontology_prefix=self.ontology)
+                    for synonym_typedef in self.synonym_typedefs or []
+                ],
             }
         )
 
         nodes = {}
+        #: a list of 3-tuples u,v,k
         links = []
-        for term in self._iter_terms(use_tqdm=use_tqdm):
+        typedefs = self._index_typedefs()
+        synonym_typedefs = self._index_synonym_typedefs()
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm):
             parents = []
-            for parent in term.parents:
+            for parent in stanza.parents:
                 if parent is None:
                     raise ValueError("parent should not be none!")
-                links.append((term.curie, "is_a", parent.curie))
+                links.append((stanza.curie, "is_a", parent.curie))
                 parents.append(parent.curie)
 
             relations = []
-            for typedef, target in term.iterate_relations():
-                if target is None:
-                    raise ValueError("target should not be none!")
+            for typedef, target in stanza.iterate_relations():
                 relations.append(f"{typedef.curie} {target.curie}")
-                links.append((term.curie, typedef.curie, target.curie))
+                links.append((stanza.curie, typedef.curie, target.curie))
+
+            for typedef, targets in sorted(stanza.properties.items()):
+                for target_or_literal in targets:
+                    if isinstance(target_or_literal, curies.Reference):
+                        links.append((stanza.curie, typedef.curie, target_or_literal.curie))
 
             d = {
-                "id": term.curie,
-                "name": term.name,
-                "def": term.definition and term._definition_fp(),
-                "xref": [xref.curie for xref in term.xrefs],
+                "id": stanza.curie,
+                "name": stanza.name,
+                "def": stanza.definition and stanza._definition_fp(),
+                "xref": [xref.curie for xref in stanza.xrefs],
                 "is_a": parents,
                 "relationship": relations,
-                "synonym": [synonym._fp() for synonym in term.synonyms],
-                "property_value": [
-                    f"{prop} {value}"
-                    for prop, values in term.properties.items()
-                    for value in values
+                "synonym": [
+                    synonym._fp(ontology_prefix=self.ontology, synonym_typedefs=synonym_typedefs)
+                    for synonym in stanza.synonyms
                 ],
+                "property_value": list(
+                    stanza._iterate_obo_properties(ontology_prefix=self.ontology, typedefs=typedefs)
+                ),
             }
-            nodes[term.curie] = {k: v for k, v in d.items() if v}
+            nodes[stanza.curie] = {k: v for k, v in d.items() if v}
 
         rv.add_nodes_from(nodes.items())
         for _source, _key, _target in links:
             rv.add_edge(_source, _target, key=_key)
 
         logger.info(
-            "[%s v%s] exported graph with %d nodes",
-            self.ontology,
-            self.data_version,
+            "[%s] exported graph with %d nodes",
+            self._prefix_version,
             rv.number_of_nodes(),
         )
         return rv
@@ -1017,11 +1406,21 @@ class Obo:
             "date": self.date and self.date.isoformat(),
         }
 
+    def iterate_references(self, *, use_tqdm: bool = False) -> Iterable[Reference]:
+        """Iterate over identifiers."""
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting identifiers"
+        ):
+            if self._in_ontology(stanza.reference):
+                yield stanza.reference
+
     def iterate_ids(self, *, use_tqdm: bool = False) -> Iterable[str]:
         """Iterate over identifiers."""
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting names"):
-            if term.prefix == self.ontology:
-                yield term.identifier
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting identifiers"
+        ):
+            if self._in_ontology_strict(stanza.reference):
+                yield stanza.identifier
 
     def get_ids(self, *, use_tqdm: bool = False) -> set[str]:
         """Get the set of identifiers."""
@@ -1029,9 +1428,11 @@ class Obo:
 
     def iterate_id_name(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str]]:
         """Iterate identifier name pairs."""
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting names"):
-            if term.prefix == self.ontology and term.name:
-                yield term.identifier, term.name
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting names"
+        ):
+            if self._in_ontology(stanza.reference) and stanza.name:
+                yield stanza.identifier, stanza.name
 
     def get_id_name_mapping(self, *, use_tqdm: bool = False) -> Mapping[str, str]:
         """Get a mapping from identifiers to names."""
@@ -1039,11 +1440,13 @@ class Obo:
 
     def iterate_id_definition(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str]]:
         """Iterate over pairs of terms' identifiers and their respective definitions."""
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting names"):
-            if term.identifier and term.definition:
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting names"
+        ):
+            if stanza.identifier and stanza.definition:
                 yield (
-                    term.identifier,
-                    term.definition.strip('"')
+                    stanza.identifier,
+                    stanza.definition.strip('"')
                     .replace("\n", " ")
                     .replace("\t", " ")
                     .replace("  ", " "),
@@ -1056,11 +1459,11 @@ class Obo:
     def get_obsolete(self, *, use_tqdm: bool = False) -> set[str]:
         """Get the set of obsolete identifiers."""
         return {
-            term.identifier
-            for term in self._iter_terms(
+            stanza.identifier
+            for stanza in self._iter_stanzas(
                 use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting obsolete"
             )
-            if term.identifier and term.is_obsolete
+            if stanza.identifier and stanza.is_obsolete
         }
 
     ############
@@ -1068,18 +1471,19 @@ class Obo:
     ############
 
     def iterate_id_species(
-        self, *, prefix: Optional[str] = None, use_tqdm: bool = False
+        self, *, prefix: str | None = None, use_tqdm: bool = False
     ) -> Iterable[tuple[str, str]]:
         """Iterate over terms' identifiers and respective species (if available)."""
         if prefix is None:
             prefix = NCBITAXON_PREFIX
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting species"):
-            species = term.get_species(prefix=prefix)
-            if species:
-                yield term.identifier, species.identifier
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting species"
+        ):
+            if isinstance(stanza, Term) and (species := stanza.get_species(prefix=prefix)):
+                yield stanza.identifier, species.identifier
 
     def get_id_species_mapping(
-        self, *, prefix: Optional[str] = None, use_tqdm: bool = False
+        self, *, prefix: str | None = None, use_tqdm: bool = False
     ) -> Mapping[str, str]:
         """Get a mapping from identifiers to species."""
         return dict(self.iterate_id_species(prefix=prefix, use_tqdm=use_tqdm))
@@ -1109,42 +1513,103 @@ class Obo:
     # PROPS #
     #########
 
-    def iterate_properties(self, *, use_tqdm: bool = False) -> Iterable[tuple[Term, str, str]]:
+    def iterate_properties(self, *, use_tqdm: bool = False) -> Iterable[tuple[Stanza, Annotation]]:
         """Iterate over tuples of terms, properties, and their values."""
-        # TODO if property_prefix is set, try removing that as a prefix from all prop strings.
-        for term in self._iter_terms(
+        for stanza in self._iter_stanzas(
             use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting properties"
         ):
-            for prop, value in term.iterate_properties():
-                yield term, prop, value
+            for property_tuple in stanza.get_property_annotations():
+                yield stanza, property_tuple
 
     @property
     def properties_header(self):
         """Property dataframe header."""
-        return [f"{self.ontology}_id", "property", "value"]
+        return [f"{self.ontology}_id", "property", "value", "datatype", "language"]
 
-    def iter_property_rows(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str, str]]:
+    @property
+    def object_properties_header(self):
+        """Property dataframe header."""
+        return ["source", "predicate", "target"]
+
+    @property
+    def literal_properties_header(self):
+        """Property dataframe header."""
+        return ["source", "predicate", "target", "datatype", "language"]
+
+    def _iter_property_rows(
+        self, *, use_tqdm: bool = False
+    ) -> Iterable[tuple[str, str, str, str, str]]:
         """Iterate property rows."""
-        for term, prop, value in self.iterate_properties(use_tqdm=use_tqdm):
-            yield term.identifier, prop, value
+        for term, t in self.iterate_properties(use_tqdm=use_tqdm):
+            pred = term._reference(t.predicate, ontology_prefix=self.ontology)
+            match t.value:
+                case OBOLiteral(value, datatype, language):
+                    yield (
+                        term.identifier,
+                        pred,
+                        value,
+                        get_preferred_curie(datatype),
+                        language or "",
+                    )
+                case Reference() as obj:
+                    yield term.identifier, pred, get_preferred_curie(obj), "", ""
+                case _:
+                    raise TypeError(f"got: {type(t)} - {t}")
+
+    def get_properties_df(self, *, use_tqdm: bool = False, drop_na: bool = True) -> pd.DataFrame:
+        """Get all properties as a dataframe."""
+        df = pd.DataFrame(
+            self._iter_property_rows(use_tqdm=use_tqdm),
+            columns=self.properties_header,
+        )
+        if drop_na:
+            df.dropna(inplace=True)
+        return df
+
+    def iter_object_properties(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str, str]]:
+        """Iterate over object property triples."""
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm):
+            for predicate, target in stanza.iterate_object_properties():
+                yield stanza.curie, predicate.curie, target.curie
 
-    def get_properties_df(self, *, use_tqdm: bool = False) -> pd.DataFrame:
+    def get_object_properties_df(self, *, use_tqdm: bool = False) -> pd.DataFrame:
         """Get all properties as a dataframe."""
         return pd.DataFrame(
-            list(self.iter_property_rows(use_tqdm=use_tqdm)),
-            columns=self.properties_header,
+            self.iter_object_properties(use_tqdm=use_tqdm), columns=self.object_properties_header
         )
 
+    def iter_literal_properties(
+        self, *, use_tqdm: bool = False
+    ) -> Iterable[tuple[str, str, str, str, str]]:
+        """Iterate over literal properties quads."""
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm):
+            for predicate, target in stanza.iterate_literal_properties():
+                yield (
+                    stanza.curie,
+                    predicate.curie,
+                    target.value,
+                    target.datatype.curie,
+                    target.language or "",
+                )
+
+    def get_literal_properties_df(self, *, use_tqdm: bool = False) -> pd.DataFrame:
+        """Get all properties as a dataframe."""
+        return pd.DataFrame(self.iter_literal_properties(), columns=self.literal_properties_header)
+
     def iterate_filtered_properties(
-        self, prop: str, *, use_tqdm: bool = False
-    ) -> Iterable[tuple[Term, str]]:
+        self, prop: ReferenceHint, *, use_tqdm: bool = False
+    ) -> Iterable[tuple[Stanza, str]]:
         """Iterate over tuples of terms and the values for the given property."""
-        for term in self._iter_terms(use_tqdm=use_tqdm):
-            for _prop, value in term.iterate_properties():
-                if _prop == prop:
-                    yield term, value
-
-    def get_filtered_properties_df(self, prop: str, *, use_tqdm: bool = False) -> pd.DataFrame:
+        prop = _ensure_ref(prop)
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm):
+            for t in stanza.get_property_annotations():
+                if t.predicate != prop:
+                    continue
+                yield stanza, reference_or_literal_to_str(t.value)
+
+    def get_filtered_properties_df(
+        self, prop: ReferenceHint, *, use_tqdm: bool = False
+    ) -> pd.DataFrame:
         """Get a dataframe of terms' identifiers to the given property's values."""
         return pd.DataFrame(
             list(self.get_filtered_properties_mapping(prop, use_tqdm=use_tqdm).items()),
@@ -1152,7 +1617,7 @@ class Obo:
         )
 
     def get_filtered_properties_mapping(
-        self, prop: str, *, use_tqdm: bool = False
+        self, prop: ReferenceHint, *, use_tqdm: bool = False
     ) -> Mapping[str, str]:
         """Get a mapping from a term's identifier to the property.
 
@@ -1164,7 +1629,7 @@ class Obo:
         }
 
     def get_filtered_properties_multimapping(
-        self, prop: str, *, use_tqdm: bool = False
+        self, prop: ReferenceHint, *, use_tqdm: bool = False
     ) -> Mapping[str, list[str]]:
         """Get a mapping from a term's identifier to the property values."""
         return multidict(
@@ -1176,22 +1641,63 @@ class Obo:
     # RELATIONS #
     #############
 
+    def iterate_edges(
+        self, *, use_tqdm: bool = False
+    ) -> Iterable[tuple[Stanza, TypeDef, Reference]]:
+        """Iterate over triples of terms, relations, and their targets."""
+        _warned: set[ReferenceTuple] = set()
+        typedefs = self._index_typedefs()
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm, desc=f"[{self.ontology}] edge"):
+            for predicate, reference in stanza._iter_edges():
+                if td := self._get_typedef(stanza, predicate, _warned, typedefs):
+                    yield stanza, td, reference
+
+    @property
+    def edges_header(self) -> Sequence[str]:
+        """Header for the edges dataframe."""
+        return [":START_ID", ":TYPE", ":END_ID"]
+
     def iterate_relations(
         self, *, use_tqdm: bool = False
-    ) -> Iterable[tuple[Term, TypeDef, Reference]]:
-        """Iterate over tuples of terms, relations, and their targets."""
-        for term in self._iter_terms(
-            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting relations"
-        ):
-            for parent in term.parents:
-                yield term, is_a, parent
-            for typedef, reference in term.iterate_relations():
-                if (self.typedefs is None or typedef not in self.typedefs) and (
-                    typedef.prefix,
-                    typedef.identifier,
-                ) not in default_typedefs:
-                    raise ValueError(f"Undefined typedef: {typedef.curie} ! {typedef.name}")
-                yield term, typedef, reference
+    ) -> Iterable[tuple[Stanza, TypeDef, Reference]]:
+        """Iterate over tuples of terms, relations, and their targets.
+
+        This only outputs stuff from the `relationship:` tag, not
+        all possible triples. For that, see :func:`iterate_edges`.
+        """
+        _warned: set[ReferenceTuple] = set()
+        typedefs = self._index_typedefs()
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm, desc=f"[{self.ontology}] relation"):
+            for predicate, reference in stanza.iterate_relations():
+                if td := self._get_typedef(stanza, predicate, _warned, typedefs):
+                    yield stanza, td, reference
+
+    def get_edges_df(self, *, use_tqdm: bool = False) -> pd.DataFrame:
+        """Get an edges dataframe."""
+        return pd.DataFrame(self.iterate_edge_rows(use_tqdm=use_tqdm), columns=self.edges_header)
+
+    def iterate_edge_rows(self, use_tqdm: bool = False) -> Iterable[tuple[str, str, str]]:
+        """Iterate the edge rows."""
+        for term, typedef, reference in self.iterate_edges(use_tqdm=use_tqdm):
+            yield term.curie, typedef.curie, reference.curie
+
+    def _get_typedef(
+        self,
+        term: Stanza,
+        predicate: Reference,
+        _warned: set[ReferenceTuple],
+        typedefs: Mapping[ReferenceTuple, TypeDef],
+    ) -> TypeDef | None:
+        pp = predicate.pair
+        if pp in typedefs:
+            return typedefs[pp]
+        if pp not in _warned:
+            _warn_string = f"[{term.curie}] undefined typedef: {pp}"
+            if predicate.name:
+                _warn_string += f" ({predicate.name})"
+            logger.warning(_warn_string)
+            _warned.add(pp)
+        return None
 
     def iter_relation_rows(
         self, use_tqdm: bool = False
@@ -1208,14 +1714,14 @@ class Obo:
 
     def iterate_filtered_relations(
         self,
-        relation: RelationHint,
+        relation: ReferenceHint,
         *,
         use_tqdm: bool = False,
-    ) -> Iterable[tuple[Term, Reference]]:
+    ) -> Iterable[tuple[Stanza, Reference]]:
         """Iterate over tuples of terms and ther targets for the given relation."""
-        _target_prefix, _target_identifier = get_reference_tuple(relation)
-        for term, typedef, reference in self.iterate_relations(use_tqdm=use_tqdm):
-            if typedef.prefix == _target_prefix and typedef.identifier == _target_identifier:
+        _pair = _ensure_ref(relation, ontology_prefix=self.ontology).pair
+        for term, predicate, reference in self.iterate_relations(use_tqdm=use_tqdm):
+            if _pair == predicate.pair:
                 yield term, reference
 
     @property
@@ -1232,7 +1738,7 @@ class Obo:
 
     def get_filtered_relations_df(
         self,
-        relation: RelationHint,
+        relation: ReferenceHint,
         *,
         use_tqdm: bool = False,
     ) -> pd.DataFrame:
@@ -1247,11 +1753,11 @@ class Obo:
 
     def iterate_filtered_relations_filtered_targets(
         self,
-        relation: RelationHint,
+        relation: ReferenceHint,
         target_prefix: str,
         *,
         use_tqdm: bool = False,
-    ) -> Iterable[tuple[Term, Reference]]:
+    ) -> Iterable[tuple[Stanza, Reference]]:
         """Iterate over relationships between one identifier and another."""
         for term, reference in self.iterate_filtered_relations(
             relation=relation, use_tqdm=use_tqdm
@@ -1261,7 +1767,7 @@ class Obo:
 
     def get_relation_mapping(
         self,
-        relation: RelationHint,
+        relation: ReferenceHint,
         target_prefix: str,
         *,
         use_tqdm: bool = False,
@@ -1272,8 +1778,8 @@ class Obo:
 
          Example usage: get homology between HGNC and MGI:
 
-        >>> from pyobo.sources.hgnc import get_obo
-        >>> obo = get_obo()
+        >>> from pyobo.sources.hgnc import HGNCGetter
+        >>> obo = HGNCGetter()
         >>> human_mapt_hgnc_id = "6893"
         >>> mouse_mapt_mgi_id = "97180"
         >>> hgnc_mgi_orthology_mapping = obo.get_relation_mapping("ro:HOM0000017", "mgi")
@@ -1291,15 +1797,15 @@ class Obo:
     def get_relation(
         self,
         source_identifier: str,
-        relation: RelationHint,
+        relation: ReferenceHint,
         target_prefix: str,
         *,
         use_tqdm: bool = False,
-    ) -> Optional[str]:
+    ) -> str | None:
         """Get the value for a bijective relation mapping between this resource and a target resource.
 
-        >>> from pyobo.sources.hgnc import get_obo
-        >>> obo = get_obo()
+        >>> from pyobo.sources.hgnc import HGNCGetter
+        >>> obo = HGNCGetter()
         >>> human_mapt_hgnc_id = "6893"
         >>> mouse_mapt_mgi_id = "97180"
         >>> assert mouse_mapt_mgi_id == obo.get_relation(human_mapt_hgnc_id, "ro:HOM0000017", "mgi")
@@ -1311,7 +1817,7 @@ class Obo:
 
     def get_relation_multimapping(
         self,
-        relation: RelationHint,
+        relation: ReferenceHint,
         target_prefix: str,
         *,
         use_tqdm: bool = False,
@@ -1334,22 +1840,24 @@ class Obo:
     ) -> Mapping[str, list[Reference]]:
         """Get a mapping from identifiers to a list of all references for the given relation."""
         return multidict(
-            (term.identifier, reference)
-            for term in self._iter_terms(
+            (stanza.identifier, reference)
+            for stanza in self._iter_stanzas(
                 use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting {typedef.curie}"
             )
-            for reference in term.get_relationships(typedef)
+            for reference in stanza.get_relationships(typedef)
         )
 
     ############
     # SYNONYMS #
     ############
 
-    def iterate_synonyms(self, *, use_tqdm: bool = False) -> Iterable[tuple[Term, Synonym]]:
+    def iterate_synonyms(self, *, use_tqdm: bool = False) -> Iterable[tuple[Stanza, Synonym]]:
         """Iterate over pairs of term and synonym object."""
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting synonyms"):
-            for synonym in sorted(term.synonyms, key=attrgetter("name")):
-                yield term, synonym
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting synonyms"
+        ):
+            for synonym in sorted(stanza.synonyms):
+                yield stanza, synonym
 
     def iterate_synonym_rows(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str]]:
         """Iterate over pairs of identifier and synonym text."""
@@ -1360,40 +1868,95 @@ class Obo:
         """Get a mapping from identifiers to a list of sorted synonym strings."""
         return multidict(self.iterate_synonym_rows(use_tqdm=use_tqdm))
 
+    def get_literal_mappings(self) -> Iterable[ssslm.LiteralMapping]:
+        """Get literal mappings in a standard data model."""
+        stanzas: Iterable[Stanza] = itt.chain(self, self.typedefs or [])
+        yield from itt.chain.from_iterable(
+            stanza.get_literal_mappings()
+            for stanza in stanzas
+            if self._in_ontology(stanza.reference)
+        )
+
+    def _in_ontology(self, reference: Reference | Referenced) -> bool:
+        return self._in_ontology_strict(reference) or self._in_ontology_aux(reference)
+
+    def _in_ontology_strict(self, reference: Reference | Referenced) -> bool:
+        return reference.prefix == self.ontology
+
+    def _in_ontology_aux(self, reference: Reference | Referenced) -> bool:
+        return reference.prefix == "obo" and reference.identifier.startswith(self.ontology + "#")
+
     #########
     # XREFS #
     #########
 
-    def iterate_xrefs(self, *, use_tqdm: bool = False) -> Iterable[tuple[Term, Reference]]:
+    def iterate_xrefs(self, *, use_tqdm: bool = False) -> Iterable[tuple[Stanza, Reference]]:
         """Iterate over xrefs."""
-        for term in self._iter_terms(use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting xrefs"):
-            for xref in term.xrefs:
-                yield term, xref
+        for stanza in self._iter_stanzas(
+            use_tqdm=use_tqdm, desc=f"[{self.ontology}] getting xrefs"
+        ):
+            xrefs = {xref for _, xref in stanza.get_mappings(add_context=False)}
+            for xref in sorted(xrefs):
+                yield stanza, xref
 
     def iterate_filtered_xrefs(
         self, prefix: str, *, use_tqdm: bool = False
-    ) -> Iterable[tuple[Term, Reference]]:
+    ) -> Iterable[tuple[Stanza, Reference]]:
         """Iterate over xrefs to a given prefix."""
         for term, xref in self.iterate_xrefs(use_tqdm=use_tqdm):
             if xref.prefix == prefix:
                 yield term, xref
 
-    def iterate_xref_rows(self, *, use_tqdm: bool = False) -> Iterable[tuple[str, str, str]]:
-        """Iterate over terms' identifiers, xref prefixes, and xref identifiers."""
-        for term, xref in self.iterate_xrefs(use_tqdm=use_tqdm):
-            yield term.identifier, xref.prefix, xref.identifier
+    def iterate_literal_mapping_rows(self) -> Iterable[ssslm.LiteralMappingTuple]:
+        """Iterate over literal mapping rows."""
+        for synonym in self.get_literal_mappings():
+            yield synonym._as_row()
 
-    @property
-    def xrefs_header(self):
-        """The header for the xref dataframe."""
-        return [f"{self.ontology}_id", TARGET_PREFIX, TARGET_ID]
+    def get_literal_mappings_df(self) -> pd.DataFrame:
+        """Get a literal mappings dataframe."""
+        return ssslm.literal_mappings_to_df(self.get_literal_mappings())
 
-    def get_xrefs_df(self, *, use_tqdm: bool = False) -> pd.DataFrame:
-        """Get a dataframe of all xrefs extracted from the OBO document."""
-        return pd.DataFrame(
-            list(self.iterate_xref_rows(use_tqdm=use_tqdm)),
-            columns=[f"{self.ontology}_id", TARGET_PREFIX, TARGET_ID],
-        ).drop_duplicates()
+    def iterate_mapping_rows(
+        self, *, use_tqdm: bool = False
+    ) -> Iterable[tuple[str, str, str, str, str, float | None, str | None]]:
+        """Iterate over SSSOM rows for mappings."""
+        for stanza in self._iter_stanzas(use_tqdm=use_tqdm):
+            for predicate, obj_ref, context in stanza.get_mappings(
+                include_xrefs=True, add_context=True
+            ):
+                yield (
+                    get_preferred_curie(stanza),
+                    stanza.name,
+                    get_preferred_curie(obj_ref),
+                    get_preferred_curie(predicate),
+                    get_preferred_curie(context.justification),
+                    context.confidence if context.confidence is not None else None,
+                    get_preferred_curie(context.contributor) if context.contributor else None,
+                )
+
+    def get_mappings_df(
+        self,
+        *,
+        use_tqdm: bool = False,
+        include_subject_labels: bool = False,
+        include_mapping_source_column: bool = False,
+    ) -> pd.DataFrame:
+        """Get a dataframe with SSSOM extracted from the OBO document."""
+        df = pd.DataFrame(self.iterate_mapping_rows(use_tqdm=use_tqdm), columns=SSSOM_DF_COLUMNS)
+        if not include_subject_labels:
+            del df["subject_label"]
+
+        # if no confidences/contributor, remove that column
+        for c in ["confidence", "contributor"]:
+            if df[c].isna().all():
+                del df[c]
+
+        # append on the mapping_source
+        # (https://mapping-commons.github.io/sssom/mapping_source/)
+        if include_mapping_source_column:
+            df["mapping_source"] = self.ontology
+
+        return df
 
     def get_filtered_xrefs_mapping(
         self, prefix: str, *, use_tqdm: bool = False
@@ -1417,11 +1980,12 @@ class Obo:
     # ALTS #
     ########
 
-    def iterate_alts(self) -> Iterable[tuple[Term, Reference]]:
+    def iterate_alts(self) -> Iterable[tuple[Stanza, Reference]]:
         """Iterate over alternative identifiers."""
-        for term in self:
-            for alt in term.alt_ids:
-                yield term, alt
+        for stanza in self._iter_stanzas():
+            if self._in_ontology(stanza):
+                for alt in stanza.alt_ids:
+                    yield stanza, alt
 
     def iterate_alt_rows(self) -> Iterable[tuple[str, str]]:
         """Iterate over pairs of terms' primary identifiers and alternate identifiers."""
@@ -1433,33 +1997,315 @@ class Obo:
         return multidict((term.identifier, alt.identifier) for term, alt in self.iterate_alts())
 
 
+@dataclass
+class TypeDef(Stanza):
+    """A type definition in OBO.
+
+    See the subsection of https://owlcollab.github.io/oboformat/doc/GO.format.obo-1_4.html#S.2.2.
+    """
+
+    reference: Annotated[Reference, 1]
+    is_anonymous: Annotated[bool | None, 2] = None
+    # 3 - name is covered by reference
+    namespace: Annotated[str | None, 4] = None
+    # 5 alt_id is part of proerties
+    definition: Annotated[str | None, 6] = None
+    comment: Annotated[str | None, 7] = None
+    subsets: Annotated[list[Reference], 8] = field(default_factory=list)
+    synonyms: Annotated[list[Synonym], 9] = field(default_factory=list)
+    xrefs: Annotated[list[Reference], 10] = field(default_factory=list)
+    _axioms: AnnotationsDict = field(default_factory=lambda: defaultdict(list))
+    properties: Annotated[PropertiesHint, 11] = field(default_factory=lambda: defaultdict(list))
+    domain: Annotated[Reference | None, 12, "typedef-only"] = None
+    range: Annotated[Reference | None, 13, "typedef-only"] = None
+    builtin: Annotated[bool | None, 14] = None
+    holds_over_chain: Annotated[list[list[Reference]], 15, "typedef-only"] = field(
+        default_factory=list
+    )
+    is_anti_symmetric: Annotated[bool | None, 16, "typedef-only"] = None
+    is_cyclic: Annotated[bool | None, 17, "typedef-only"] = None
+    is_reflexive: Annotated[bool | None, 18, "typedef-only"] = None
+    is_symmetric: Annotated[bool | None, 19, "typedef-only"] = None
+    is_transitive: Annotated[bool | None, 20, "typedef-only"] = None
+    is_functional: Annotated[bool | None, 21, "typedef-only"] = None
+    is_inverse_functional: Annotated[bool | None, 22, "typedef-only"] = None
+    parents: Annotated[list[Reference], 23] = field(default_factory=list)
+    intersection_of: Annotated[IntersectionOfHint, 24] = field(default_factory=list)
+    union_of: Annotated[list[Reference], 25] = field(default_factory=list)
+    equivalent_to: Annotated[list[Reference], 26] = field(default_factory=list)
+    disjoint_from: Annotated[list[Reference], 27] = field(default_factory=list)
+    # TODO inverse should be inverse_of, cardinality any
+    inverse: Annotated[Reference | None, 28, "typedef-only"] = None
+    # TODO check if there are any examples of this being multiple
+    transitive_over: Annotated[list[Reference], 29, "typedef-only"] = field(default_factory=list)
+    equivalent_to_chain: Annotated[list[list[Reference]], 30, "typedef-only"] = field(
+        default_factory=list
+    )
+    #: From the OBO spec:
+    #:
+    #:   For example: spatially_disconnected_from is disjoint_over part_of, in that two
+    #:   disconnected entities have no parts in common. This can be translated to OWL as:
+    #:   ``disjoint_over(R S), R(A B) ==> (S some A) disjointFrom (S some B)``
+    disjoint_over: Annotated[list[Reference], 31] = field(default_factory=list)
+    relationships: Annotated[RelationsHint, 32] = field(default_factory=lambda: defaultdict(list))
+    is_obsolete: Annotated[bool | None, 33] = None
+    created_by: Annotated[str | None, 34] = None
+    creation_date: Annotated[datetime.datetime | None, 35] = None
+    # TODO expand_assertion_to
+    # TODO expand_expression_to
+    #: Whether this relationship is a metadata tag. Properties that are marked as metadata tags are
+    #: used to record object metadata. Object metadata is additional information about an object
+    #: that is useful to track, but does not impact the definition of the object or how it should
+    #: be treated by a reasoner. Metadata tags might be used to record special term synonyms or
+    #: structured notes about a term, for example.
+    is_metadata_tag: Annotated[bool | None, 40, "typedef-only"] = None
+    is_class_level: Annotated[bool | None, 41] = None
+
+    type: StanzaType = "TypeDef"
+
+    def __hash__(self) -> int:
+        # have to re-define hash because of the @dataclass
+        return hash((self.__class__, self.prefix, self.identifier))
+
+    def _get_references(self) -> dict[str, set[Reference]]:
+        rv = super()._get_references()
+
+        def _add(r: Reference) -> None:
+            rv[r.prefix].add(r)
+
+        if self.domain:
+            _add(self.domain)
+        if self.range:
+            _add(self.range)
+        if self.inverse:
+            _add(self.inverse)
+
+        # TODO all of the properties, which are from oboInOwl
+        for rr in itt.chain(self.transitive_over, self.disjoint_over):
+            _add(rr)
+        for part in itt.chain(self.holds_over_chain, self.equivalent_to_chain):
+            for rr in part:
+                _add(rr)
+        return dict(rv)
+
+    def iterate_obo_lines(
+        self,
+        ontology_prefix: str,
+        synonym_typedefs: Mapping[ReferenceTuple, SynonymTypeDef] | None = None,
+        typedefs: Mapping[ReferenceTuple, TypeDef] | None = None,
+    ) -> Iterable[str]:
+        """Iterate over the lines to write in an OBO file.
+
+        :param ontology_prefix:
+            The prefix of the ontology into which the type definition is being written.
+            This is used for compressing builtin identifiers
+        :yield:
+            The lines to write to an OBO file
+
+        `S.3.5.5 <https://owlcollab.github.io/oboformat/doc/GO.format.obo-1_4.html#S.3.5.5>`_
+        of the OBO Flat File Specification v1.4 says tags should appear in the following order:
+
+        1. id
+        2. is_anonymous
+        3. name
+        4. namespace
+        5. alt_id
+        6. def
+        7. comment
+        8. subset
+        9. synonym
+        10. xref
+        11. property_value
+        12. domain
+        13. range
+        14. builtin
+        15. holds_over_chain
+        16. is_anti_symmetric
+        17. is_cyclic
+        18. is_reflexive
+        19. is_symmetric
+        20. is_transitive
+        21. is_functional
+        22. is_inverse_functional
+        23. is_a
+        24. intersection_of
+        25. union_of
+        26. equivalent_to
+        27. disjoint_from
+        28. inverse_of
+        29. transitive_over
+        30. equivalent_to_chain
+        31. disjoint_over
+        32. relationship
+        33. is_obsolete
+        34. created_by
+        35. creation_date
+        36. replaced_by
+        37. consider
+        38. expand_assertion_to
+        39. expand_expression_to
+        40. is_metadata_tag
+        41. is_class_level
+        """
+        if synonym_typedefs is None:
+            synonym_typedefs = {}
+        if typedefs is None:
+            typedefs = {}
+
+        yield "\n[Typedef]"
+        # 1
+        yield f"id: {reference_escape(self.reference, ontology_prefix=ontology_prefix)}"
+        # 2
+        yield from _boolean_tag("is_anonymous", self.is_anonymous)
+        # 3
+        if self.name:
+            yield f"name: {self.name}"
+        # 4
+        if self.namespace:
+            yield f"namespace: {self.namespace}"
+        # 5
+        yield from _reference_list_tag("alt_id", self.alt_ids, ontology_prefix)
+        # 6
+        if self.definition:
+            yield f"def: {self._definition_fp()}"
+        # 7
+        if self.comment:
+            yield f"comment: {self.comment}"
+        # 8
+        yield from _reference_list_tag("subset", self.subsets, ontology_prefix)
+        # 9
+        for synonym in self.synonyms:
+            yield synonym.to_obo(ontology_prefix=ontology_prefix, synonym_typedefs=synonym_typedefs)
+        # 10
+        yield from self._iterate_xref_obo(ontology_prefix=ontology_prefix)
+        # 11
+        yield from self._iterate_obo_properties(
+            ontology_prefix=ontology_prefix,
+            skip_predicate_objects=v.SKIP_PROPERTY_PREDICATES_OBJECTS,
+            skip_predicate_literals=v.SKIP_PROPERTY_PREDICATES_LITERAL,
+            typedefs=typedefs,
+        )
+        # 12
+        if self.domain:
+            yield f"domain: {reference_escape(self.domain, ontology_prefix=ontology_prefix, add_name_comment=True)}"
+        # 13
+        if self.range:
+            yield f"range: {reference_escape(self.range, ontology_prefix=ontology_prefix, add_name_comment=True)}"
+        # 14
+        yield from _boolean_tag("builtin", self.builtin)
+        # 15
+        yield from _chain_tag("holds_over_chain", self.holds_over_chain, ontology_prefix)
+        # 16
+        yield from _boolean_tag("is_anti_symmetric", self.is_anti_symmetric)
+        # 17
+        yield from _boolean_tag("is_cyclic", self.is_cyclic)
+        # 18
+        yield from _boolean_tag("is_reflexive", self.is_reflexive)
+        # 19
+        yield from _boolean_tag("is_symmetric", self.is_symmetric)
+        # 20
+        yield from _boolean_tag("is_transitive", self.is_transitive)
+        # 21
+        yield from _boolean_tag("is_functional", self.is_functional)
+        # 22
+        yield from _boolean_tag("is_inverse_functional", self.is_inverse_functional)
+        # 23
+        yield from _reference_list_tag("is_a", self.parents, ontology_prefix)
+        # 24
+        yield from self._iterate_intersection_of_obo(ontology_prefix=ontology_prefix)
+        # 25
+        yield from _reference_list_tag("union_of", self.union_of, ontology_prefix)
+        # 26
+        yield from _reference_list_tag("equivalent_to", self.equivalent_to, ontology_prefix)
+        # 27
+        yield from _reference_list_tag("disjoint_from", self.disjoint_from, ontology_prefix)
+        # 28
+        if self.inverse:
+            yield f"inverse_of: {reference_escape(self.inverse, ontology_prefix=ontology_prefix, add_name_comment=True)}"
+        # 29
+        yield from _reference_list_tag("transitive_over", self.transitive_over, ontology_prefix)
+        # 30
+        yield from _chain_tag("equivalent_to_chain", self.equivalent_to_chain, ontology_prefix)
+        # 31 disjoint_over, see https://github.com/search?q=%22disjoint_over%3A%22+path%3A*.obo&type=code
+        yield from _reference_list_tag(
+            "disjoint_over", self.disjoint_over, ontology_prefix=ontology_prefix
+        )
+        # 32
+        yield from self._iterate_obo_relations(ontology_prefix=ontology_prefix, typedefs=typedefs)
+        # 33
+        yield from _boolean_tag("is_obsolete", self.is_obsolete)
+        # 34
+        if self.created_by:
+            yield f"created_by: {self.created_by}"
+        # 35
+        if self.creation_date is not None:
+            yield f"creation_date: {self.creation_date.isoformat()}"
+        # 36
+        yield from _tag_property_targets(
+            "replaced_by", self, v.term_replaced_by, ontology_prefix=ontology_prefix
+        )
+        # 37
+        yield from _tag_property_targets(
+            "consider", self, v.see_also, ontology_prefix=ontology_prefix
+        )
+        # 38 TODO expand_assertion_to
+        # 39 TODO expand_expression_to
+        # 40
+        yield from _boolean_tag("is_metadata_tag", self.is_metadata_tag)
+        # 41
+        yield from _boolean_tag("is_class_level", self.is_class_level)
+
+    @classmethod
+    def from_triple(cls, prefix: str, identifier: str, name: str | None = None) -> TypeDef:
+        """Create a typedef from a reference."""
+        return cls(reference=Reference(prefix=prefix, identifier=identifier, name=name))
+
+    @classmethod
+    def default(
+        cls, prefix: str, identifier: str, *, name: str | None = None, is_metadata_tag: bool
+    ) -> Self:
+        """Construct a default type definition from within the OBO namespace."""
+        return cls(
+            reference=default_reference(prefix, identifier, name=name),
+            is_metadata_tag=is_metadata_tag,
+        )
+
+
+class AdHocOntologyBase(Obo):
+    """A base class for ad-hoc ontologies."""
+
+
 def make_ad_hoc_ontology(
     _ontology: str,
-    _name: str,
-    _auto_generated_by: Optional[str] = None,
-    _format_version: str = "1.2",
-    _typedefs: Optional[list[TypeDef]] = None,
-    _synonym_typedefs: Optional[list[SynonymTypeDef]] = None,
-    _date: Optional[datetime] = None,
-    _data_version: Optional[str] = None,
-    _idspaces: Optional[Mapping[str, str]] = None,
-    _root_terms: Optional[list[Reference]] = None,
+    _name: str | None = None,
+    _auto_generated_by: str | None = None,
+    _typedefs: list[TypeDef] | None = None,
+    _synonym_typedefs: list[SynonymTypeDef] | None = None,
+    _date: datetime.datetime | None = None,
+    _data_version: str | None = None,
+    _idspaces: Mapping[str, str] | None = None,
+    _root_terms: list[Reference] | None = None,
+    _subsetdefs: list[tuple[Reference, str]] | None = None,
+    _property_values: list[Annotation] | None = None,
+    _imports: list[str] | None = None,
     *,
-    terms: list[Term],
-) -> "Obo":
+    terms: list[Term] | None = None,
+) -> Obo:
     """Make an ad-hoc ontology."""
 
-    class AdHocOntology(Obo):
+    class AdHocOntology(AdHocOntologyBase):
         """An ad hoc ontology created from an OBO file."""
 
         ontology = _ontology
         name = _name
         auto_generated_by = _auto_generated_by
-        format_version = _format_version
         typedefs = _typedefs
         synonym_typedefs = _synonym_typedefs
         idspaces = _idspaces
         root_terms = _root_terms
+        subsetdefs = _subsetdefs
+        property_values = _property_values
+        imports = _imports
 
         def __post_init__(self):
             self.date = _date
@@ -1467,30 +2313,11 @@ def make_ad_hoc_ontology(
 
         def iter_terms(self, force: bool = False) -> Iterable[Term]:
             """Iterate over terms in the ad hoc ontology."""
-            return terms
+            return terms or []
 
     return AdHocOntology()
 
 
-def _convert_typedefs(typedefs: Optional[Iterable[TypeDef]]) -> list[Mapping[str, Any]]:
-    """Convert the type defs."""
-    if not typedefs:
-        return []
-    return [_convert_typedef(typedef) for typedef in typedefs]
-
-
-def _convert_typedef(typedef: TypeDef) -> Mapping[str, Any]:
-    """Convert a type def."""
-    # TODO add more later
-    return typedef.reference.model_dump()
-
-
-def _convert_synonym_typedefs(synonym_typedefs: Optional[Iterable[SynonymTypeDef]]) -> list[str]:
-    """Convert the synonym type defs."""
-    if not synonym_typedefs:
-        return []
-    return [_convert_synonym_typedef(synonym_typedef) for synonym_typedef in synonym_typedefs]
-
-
-def _convert_synonym_typedef(synonym_typedef: SynonymTypeDef) -> str:
-    return f'{synonym_typedef.preferred_curie} "{synonym_typedef.name}"'
+HUMAN_TERM = Term(reference=v.HUMAN)
+CHARLIE_TERM = Term(reference=v.CHARLIE, type="Instance").append_parent(HUMAN_TERM)
+PYOBO_INJECTED = "Injected by PyOBO"