"""Utilities for interacting with data and the schema."""
from __future__ import annotations
import csv
import json
import logging
from collections import defaultdict
from collections.abc import Mapping
from functools import lru_cache
from operator import attrgetter
from pathlib import Path
from curies import Reference
from pydantic import BaseModel, Field
from typing_extensions import Literal
from .constants import (
BIOREGISTRY_PATH,
COLLECTIONS_PATH,
CONTEXTS_PATH,
CURATED_MAPPINGS_PATH,
METAREGISTRY_PATH,
)
from .schema import Collection, Context, Registry, Resource
logger = logging.getLogger(__name__)
def _read_metaregistry(path: str | Path) -> Mapping[str, Registry]:
with open(path, encoding="utf-8") as file:
data = json.load(file)
return {
registry.prefix: registry
for registry in (Registry(**record) for record in data["metaregistry"])
}
[docs]
def registries() -> list[Registry]:
"""Get a list of registries in the Bioregistry."""
return sorted(read_metaregistry().values(), key=attrgetter("prefix"))
[docs]
@lru_cache(maxsize=1)
def read_registry() -> Mapping[str, Resource]:
"""Read the Bioregistry as JSON."""
return _registry_from_path(BIOREGISTRY_PATH)
[docs]
def resources() -> list[Resource]:
"""Get a list of resources in the Bioregistry."""
return sorted(read_registry().values(), key=attrgetter("prefix"))
def _registry_from_path(path: str | Path) -> Mapping[str, Resource]:
with open(path, encoding="utf-8") as file:
data = json.load(file)
for prefix, value in data.items():
value.setdefault("prefix", prefix)
return {prefix: Resource.model_validate(value) for prefix, value in data.items()}
def add_resource(resource: Resource) -> None:
"""Add a resource to the registry.
:param resource: A resource object to write
:raises KeyError: if the prefix is already present in the registry
"""
registry = dict(read_registry())
if resource.prefix in registry:
raise KeyError(f"Tried to add duplicate prefix to the registry: {resource.prefix}")
registry[resource.prefix] = resource
# Clear the cache
read_registry.cache_clear()
write_registry(registry)
class SemanticMapping(BaseModel):
"""A model representing a SSSOM semantic mapping."""
subject: Reference = Field(..., alias="subject_id")
predicate_modifier: Literal["Not"] | None = Field(None)
predicate: Reference = Field(..., alias="predicate_id")
object: Reference = Field(..., alias="object_id")
creator: Reference = Field(..., alias="creator_id")
mapping_justification: Reference = Field(...)
comment: str | None = Field(None)
issue_tracker_item: int | None = Field(
None, description="The PR or issue associated with the change"
)
date: str = Field(
...,
pattern="^\\d{4}-\\d{2}-\\d{2}$",
description="The ISO-8601 date of curation in YYYY-MM-DD",
)
[docs]
def read_mismatches() -> dict[str, dict[str, set[str]]]:
"""Read the mismatches subset of curated mappings as a nested dictionary data structure."""
mismatches: defaultdict[str, defaultdict[str, set[str]]] = defaultdict(lambda: defaultdict(set))
for m in read_mappings():
if m.predicate.curie == "skos:exactMatch" and m.predicate_modifier == "Not":
mismatches[m.subject.identifier][m.object.prefix].add(m.object.identifier)
return {k: dict(v) for k, v in mismatches.items()}
[docs]
@lru_cache(maxsize=1)
def read_mappings() -> list[SemanticMapping]:
"""Read curated mappings as a nested dict data structure."""
return _read_semantic_mappings(CURATED_MAPPINGS_PATH)
def _read_semantic_mappings(path: str | Path) -> list[SemanticMapping]:
"""Read curated mappings as a nested dict data structure."""
with Path(path).expanduser().resolve().open() as file:
return [
SemanticMapping.model_validate({k: v for k, v in record.items() if v})
for record in csv.DictReader(file, delimiter="\t")
]
[docs]
def is_mismatch(bioregistry_prefix: str, external_metaprefix: str, external_prefix: str) -> bool:
"""Return if the triple is a mismatch."""
return external_prefix in read_mismatches().get(bioregistry_prefix, {}).get(
external_metaprefix, {}
)
def write_mappings(mappings: list[SemanticMapping]) -> None:
"""Write mappings into the curated mappings file with appropriate sorting."""
mappings = sorted(
mappings,
key=lambda x: (x.subject, x.object, x.predicate, x.predicate_modifier),
)
header = [
"subject_id",
"predicate_modifier",
"predicate_id",
"object_id",
"creator_id",
"mapping_justification",
"comment",
"issue_tracker_item",
"date",
]
with CURATED_MAPPINGS_PATH.open("w") as file:
writer = csv.writer(file, delimiter="\t", lineterminator="\n")
writer.writerow(header)
writer.writerows(
(
mapping.subject.curie,
mapping.predicate_modifier,
mapping.predicate.curie,
mapping.object.curie,
mapping.creator.curie,
mapping.mapping_justification.curie,
mapping.comment,
mapping.issue_tracker_item,
mapping.date,
)
for mapping in mappings
)
[docs]
@lru_cache(maxsize=1)
def read_collections() -> Mapping[str, Collection]:
"""Read the manually curated collections."""
return _collections_from_path(COLLECTIONS_PATH)
def _collections_from_path(path: str | Path) -> Mapping[str, Collection]:
with open(path, encoding="utf-8") as file:
data = json.load(file)
return {
collection.identifier: collection
for collection in (Collection(**record) for record in data["collections"])
}
def write_collections(collections: Mapping[str, Collection]) -> None:
"""Write the collections."""
values = [v for _, v in sorted(collections.items())]
for collection in values:
collection.resources = sorted(set(collection.resources))
with open(COLLECTIONS_PATH, encoding="utf-8", mode="w") as file:
json.dump(
{"collections": [c.model_dump(exclude_none=True) for c in values]},
file,
indent=2,
sort_keys=True,
ensure_ascii=False,
)
[docs]
def write_registry(registry: Mapping[str, Resource], *, path: Path | None = None) -> None:
"""Write to the Bioregistry."""
if path is None:
path = BIOREGISTRY_PATH
with path.open(mode="w", encoding="utf-8") as file:
json.dump(
{
key: resource.model_dump(exclude_none=True, exclude={"prefix"})
for key, resource in registry.items()
},
file,
indent=2,
sort_keys=True,
ensure_ascii=False,
)
def write_metaregistry(metaregistry: Mapping[str, Registry]) -> None:
"""Write to the metaregistry."""
values = [v for _, v in sorted(metaregistry.items())]
with open(METAREGISTRY_PATH, mode="w", encoding="utf-8") as file:
json.dump(
{"metaregistry": [m.model_dump(exclude_none=True) for m in values]},
fp=file,
indent=2,
sort_keys=True,
ensure_ascii=False,
)
[docs]
def write_contexts(contexts: Mapping[str, Context]) -> None:
"""Write to contexts."""
with open(CONTEXTS_PATH, mode="w", encoding="utf-8") as file:
json.dump(
{key: context.model_dump(exclude_none=True) for key, context in contexts.items()},
fp=file,
indent=2,
sort_keys=True,
ensure_ascii=False,
)
def read_prefix_contributions(registry: Mapping[str, Resource]) -> Mapping[str, set[str]]:
"""Get a mapping from contributor ORCID identifiers to prefixes."""
rv = defaultdict(set)
for prefix, resource in registry.items():
if resource.contributor and resource.contributor.orcid:
rv[resource.contributor.orcid].add(prefix)
for contributor in resource.contributor_extras or []:
if contributor.orcid:
rv[contributor.orcid].add(prefix)
return dict(rv)
def read_prefix_reviews(registry: Mapping[str, Resource]) -> Mapping[str, set[str]]:
"""Get a mapping from reviewer ORCID identifiers to prefixes."""
rv = defaultdict(set)
for prefix, resource in registry.items():
if resource.reviewer and resource.reviewer.orcid:
rv[resource.reviewer.orcid].add(prefix)
for reviewer in resource.reviewer_extras or []:
if reviewer.orcid:
rv[reviewer.orcid].add(prefix)
return dict(rv)
def read_prefix_contacts(registry: Mapping[str, Resource]) -> Mapping[str, set[str]]:
"""Get a mapping from contact ORCID identifiers to prefixes."""
rv = defaultdict(set)
for prefix, resource in registry.items():
contact_orcid = resource.get_contact_orcid()
if contact_orcid:
rv[contact_orcid].add(prefix)
# Add all secondary contacts' ORCIDs
for secondary_contact in resource.contact_extras or []:
if secondary_contact.orcid:
rv[secondary_contact.orcid].add(prefix)
return dict(rv)
def read_collections_contributions(collections: Mapping[str, Collection]) -> Mapping[str, set[str]]:
"""Get a mapping from contributor ORCID identifiers to collections."""
rv = defaultdict(set)
for collection_id, resource in collections.items():
for author in resource.authors or []:
rv[author.orcid].add(collection_id)
return dict(rv)
def read_registry_contributions(metaregistry: Mapping[str, Registry]) -> Mapping[str, set[str]]:
"""Get a mapping from contributor ORCID identifiers to collections."""
rv = defaultdict(set)
for metaprefix, resource in metaregistry.items():
if resource.contact and resource.contact.orcid:
rv[resource.contact.orcid].add(metaprefix)
return dict(rv)
def read_context_contributions(contexts: Mapping[str, Context]) -> Mapping[str, set[str]]:
"""Get a mapping from contributor ORCID identifiers to contexts."""
rv = defaultdict(set)
for context_key, context in contexts.items():
for maintainer in context.maintainers:
rv[maintainer.orcid].add(context_key)
return dict(rv)
[docs]
@lru_cache(1)
def read_contexts() -> Mapping[str, Context]:
"""Get a mapping from context keys to contexts."""
return _contexts_from_path(CONTEXTS_PATH)
def _contexts_from_path(path: str | Path) -> Mapping[str, Context]:
with open(path, encoding="utf-8") as file:
data = json.load(file)
return {key: Context(**data) for key, data in data.items()}