nmdc-runtime 2.9.0__py3-none-any.whl → 2.10.0__py3-none-any.whl

nmdc_runtime/minter/domain/model.py

@@ -1,9 +1,11 @@
 from enum import Enum
+import re
 from typing import Optional
 
+from base32_lib import base32
 from pydantic import BaseModel, PositiveInt
 
-from nmdc_runtime.minter.config import schema_classes
+from nmdc_runtime.minter.config import schema_classes, typecodes
 
 
 class Entity(BaseModel):
@@ -71,3 +73,35 @@ class Identifier(Entity):
 class Typecode(Entity):
     schema_class: str
     name: str
+
+
+id_prefix_pattern = rf"(?P<prefix>nmdc)"
+id_typecode_pattern = rf"(?P<typecode>[a-z]{{1,6}})"
+id_shoulder_pattern = rf"(?P<shoulder>[0-9][a-z]{{0,6}}[0-9])"
+id_blade_pattern = rf"(?P<blade>[A-Za-z0-9]+)"
+id_version_pattern = rf"(?P<version>(\.[A-Za-z0-9]+)*)"
+id_locus_pattern = rf"(?P<locus>_[A-Za-z0-9_\.-]+)?"
+id_pattern = (
+    rf"^{id_prefix_pattern}:{id_typecode_pattern}-{id_shoulder_pattern}-"
+    rf"{id_blade_pattern}{id_version_pattern}{id_locus_pattern}$"
+)
+ID_TYPECODE_VALUES = [t["name"] for t in typecodes()]
+id_typecode_pattern_strict = rf"(?P<typecode_strict>({'|'.join(ID_TYPECODE_VALUES)}))"
+id_blade_pattern_strict = rf"(?P<blade_strict>[{base32.ENCODING_CHARS}]+)"
+id_pattern_strict = (
+    rf"^{id_prefix_pattern}:{id_typecode_pattern_strict}-{id_shoulder_pattern}-"
+    rf"{id_blade_pattern_strict}{id_version_pattern}{id_locus_pattern}$"
+)
+id_pattern_strict_compiled = re.compile(id_pattern_strict)
+
+
+def check_valid_ids(ids: list[str]):
+    for id_ in ids:
+        if not re.match(id_pattern, id_):
+            raise ValueError(
+                (
+                    f"Invalid ID format for given ID: '{id_}'.\n\nAn ID must match the pattern: '{id_pattern}'.\n\n"
+                    "See: <https://microbiomedata.github.io/nmdc-schema/identifiers/#ids-minted-for-use-within-nmdc>"
+                )
+            )
+    return ids

nmdc_runtime/minter/entrypoints/fastapi_app.py

@@ -8,7 +8,7 @@ from nmdc_runtime.api.core.util import raise404_if_none
 from nmdc_runtime.api.db.mongo import get_mongo_db
 from nmdc_runtime.api.models.site import get_current_client_site, Site
 from nmdc_runtime.minter.adapters.repository import MongoIDStore, MinterError
-from nmdc_runtime.minter.config import minting_service_id, schema_classes
+from nmdc_runtime.minter.config import minting_service_id
 from nmdc_runtime.minter.domain.model import (
     Identifier,
     AuthenticatedMintingRequest,

nmdc_runtime/mongo_util.py

@@ -1,7 +1,6 @@
-from pymongo import MongoClient
 from pymongo.database import Database
 from pymongo.collection import Collection
-from typing import Any, Mapping, Optional, Type, Callable
+from typing import Any, Optional
 from pymongo.client_session import ClientSession
 import inspect
 

nmdc_runtime/site/backup/nmdcdb_mongodump.py

@@ -6,7 +6,7 @@ $ nmdcdb-mongodump
 
 import os
 import subprocess
-from datetime import datetime, timezone
+from datetime import datetime
 from pathlib import Path
 from zoneinfo import ZoneInfo
 

nmdc_runtime/site/backup/nmdcdb_mongoexport.py

@@ -16,9 +16,7 @@ from toolz import assoc
 
 from nmdc_runtime.api.core.util import pick
 from nmdc_runtime.api.db.mongo import get_mongo_db
-from nmdc_runtime.site.repository import run_config_frozen__normal_env
-from nmdc_runtime.site.resources import get_mongo
-from nmdc_runtime.util import nmdc_jsonschema, schema_collection_names_with_id_field
+from nmdc_runtime.util import schema_collection_names_with_id_field
 
 
 def collection_stats(mdb: MongoDatabase):

nmdc_runtime/site/export/ncbi_xml.py

@@ -4,7 +4,7 @@ import datetime
 import xml.etree.ElementTree as ET
 import xml.dom.minidom
 
-from typing import Any, List, Union
+from typing import Any, List
 from urllib.parse import urlparse
 from nmdc_runtime.site.export.ncbi_xml_utils import (
     handle_controlled_identified_term_value,
@@ -16,7 +16,6 @@ from nmdc_runtime.site.export.ncbi_xml_utils import (
     handle_float_value,
     handle_string_value,
     load_mappings,
-    validate_xml,
 )
 
 

nmdc_runtime/site/export/ncbi_xml_utils.py

@@ -1,5 +1,5 @@
 from io import BytesIO, StringIO
-from typing import Any, Dict, List, Union
+from typing import Any, Dict, List
 
 from nmdc_runtime.api.endpoints.util import strip_oid
 from nmdc_runtime.minter.config import typecodes

nmdc_runtime/site/graphs.py

@@ -1,7 +1,6 @@
-from dagster import graph, GraphIn
+from dagster import graph
 
 from nmdc_runtime.site.ops import (
-    build_merged_db,
     generate_biosample_set_for_nmdc_study_from_gold,
     nmdc_schema_database_export_filename,
     nmdc_schema_database_from_gold_study,
@@ -12,8 +11,6 @@ from nmdc_runtime.site.ops import (
     gold_projects_by_study,
     gold_study,
     poll_for_run_completion,
-    run_etl,
-    local_file_to_api_object,
     get_operation,
     produce_curated_db,
     delete_operations,
@@ -70,24 +67,6 @@ from nmdc_runtime.site.ops import (
 from nmdc_runtime.site.export.study_metadata import get_biosamples_by_study_id
 
 
-@graph
-def gold_translation():
-    """
-    Translating an export of the JGI GOLD [1] SQL database to the NMDC database JSON schema.
-
-    [1] Genomes OnLine Database (GOLD) <https://gold.jgi.doe.gov/>.
-    """
-    local_file_to_api_object(run_etl(build_merged_db()))
-
-
-@graph()
-def gold_translation_curation():
-    # TODO
-    #   - have produce_curated_db do actual curation (see notebook), persisting to db.
-    #   - more steps in pipeline? Or handoff via run_status_sensor on DagsterRunStatus.SUCCESS.
-    produce_curated_db(get_operation())
-
-
 @graph()
 def create_objects_from_site_object_puts():
     delete_operations(

nmdc_runtime/site/ops.py

@@ -4,19 +4,18 @@ import logging
 import mimetypes
 import os
 import subprocess
-import tempfile
 from collections import defaultdict
 from datetime import datetime, timezone
-from io import BytesIO, StringIO
+from io import BytesIO
 from pprint import pformat
 from toolz.dicttoolz import keyfilter
-from typing import Tuple, Set, Union
+from typing import Tuple, Set
 from zipfile import ZipFile
 from itertools import chain
 from ontology_loader.ontology_load_controller import OntologyLoaderController
 import pandas as pd
 import requests
-
+from refscan.lib.helpers import get_names_of_classes_in_effective_range_of_slot
 
 from bson import ObjectId, json_util
 from dagster import (
@@ -44,7 +43,7 @@ from dagster import (
 from gridfs import GridFS
 from linkml_runtime.utils.dictutils import as_simple_dict
 from linkml_runtime.utils.yamlutils import YAMLRoot
-from nmdc_runtime.api.db.mongo import get_mongo_db, validate_json
+from nmdc_runtime.api.db.mongo import validate_json
 from nmdc_runtime.api.core.idgen import generate_one_id
 from nmdc_runtime.api.core.metadata import (
     _validate_changesheet,
@@ -103,22 +102,19 @@ from nmdc_runtime.site.util import (
 )
 from nmdc_runtime.util import (
     drs_object_in_for,
-    get_names_of_classes_in_effective_range_of_slot,
     pluralize,
     put_object,
     specialize_activity_set_docs,
     collection_name_to_class_names,
-    class_hierarchy_as_list,
     nmdc_schema_view,
     populated_schema_collection_names_with_id_field,
 )
 from nmdc_schema import nmdc
-from nmdc_schema.nmdc import Database as NMDCDatabase
-from pydantic import BaseModel
 from pymongo import InsertOne, UpdateOne
 from pymongo.database import Database as MongoDatabase
 from starlette import status
-from toolz import assoc, dissoc, get_in, valfilter, identity
+from toolz import get_in, valfilter, identity
+
 
 # batch size for writing documents to alldocs
 BULK_WRITE_BATCH_SIZE = 2000
@@ -153,99 +149,6 @@ def mongo_stats(context) -> List[str]:
     return collection_names
 
 
-@op(
-    required_resource_keys={"mongo", "runtime_api_site_client"},
-    retry_policy=RetryPolicy(max_retries=2),
-)
-def local_file_to_api_object(context, file_info):
-    client: RuntimeApiSiteClient = context.resources.runtime_api_site_client
-    storage_path: str = file_info["storage_path"]
-    mime_type = file_info.get("mime_type")
-    if mime_type is None:
-        mime_type = mimetypes.guess_type(storage_path)[0]
-    rv = client.put_object_in_site(
-        {"mime_type": mime_type, "name": storage_path.rpartition("/")[-1]}
-    )
-    if not rv.status_code == status.HTTP_200_OK:
-        raise Failure(description=f"put_object_in_site failed: {rv.content}")
-    op = rv.json()
-    context.log.info(f"put_object_in_site: {op}")
-    rv = put_object(storage_path, op["metadata"]["url"])
-    if not rv.status_code == status.HTTP_200_OK:
-        raise Failure(description=f"put_object failed: {rv.content}")
-    op_patch = {"done": True, "result": drs_object_in_for(storage_path, op)}
-    rv = client.update_operation(op["id"], op_patch)
-    if not rv.status_code == status.HTTP_200_OK:
-        raise Failure(description="update_operation failed")
-    op = rv.json()
-    context.log.info(f"update_operation: {op}")
-    rv = client.create_object_from_op(op)
-    if rv.status_code != status.HTTP_201_CREATED:
-        raise Failure("create_object_from_op failed")
-    obj = rv.json()
-    context.log.info(f'Created /objects/{obj["id"]}')
-    mdb = context.resources.mongo.db
-    rv = mdb.operations.delete_one({"id": op["id"]})
-    if rv.deleted_count != 1:
-        context.log.error("deleting op failed")
-    yield AssetMaterialization(
-        asset_key=AssetKey(["object", obj["name"]]),
-        description="output of metadata-translation run_etl",
-        metadata={"object_id": MetadataValue.text(obj["id"])},
-    )
-    yield Output(obj)
-
-
-@op(
-    out={
-        "merged_data_path": Out(
-            str,
-            description="path to TSV merging of source metadata",
-        )
-    }
-)
-def build_merged_db(context) -> str:
-    context.log.info("metadata-translation: running `make build-merged-db`")
-    run_and_log(
-        "cd /opt/dagster/lib/metadata-translation/ && make build-merged-db", context
-    )
-    storage_path = (
-        "/opt/dagster/lib/metadata-translation/src/data/nmdc_merged_data.tsv.zip"
-    )
-    yield AssetMaterialization(
-        asset_key=AssetKey(["gold_translation", "merged_data.tsv.zip"]),
-        description="input to metadata-translation run_etl",
-        metadata={"path": MetadataValue.path(storage_path)},
-    )
-    yield Output(storage_path, "merged_data_path")
-
-
-@op(
-    required_resource_keys={"runtime_api_site_client"},
-)
-def run_etl(context, merged_data_path: str):
-    context.log.info("metadata-translation: running `make run-etl`")
-    if not os.path.exists(merged_data_path):
-        raise Failure(description=f"merged_db not present at {merged_data_path}")
-    run_and_log("cd /opt/dagster/lib/metadata-translation/ && make run-etl", context)
-    storage_path = (
-        "/opt/dagster/lib/metadata-translation/src/data/nmdc_database.json.zip"
-    )
-    with ZipFile(storage_path) as zf:
-        name = zf.namelist()[0]
-        with zf.open(name) as f:
-            rv = json.load(f)
-    context.log.info(f"nmdc_database.json keys: {list(rv.keys())}")
-    yield AssetMaterialization(
-        asset_key=AssetKey(["gold_translation", "database.json.zip"]),
-        description="output of metadata-translation run_etl",
-        metadata={
-            "path": MetadataValue.path(storage_path),
-        },
-    )
-    yield Output({"storage_path": storage_path})
-
-
 @op(required_resource_keys={"mongo"})
 def get_operation(context):
     mdb = context.resources.mongo.db
@@ -1043,15 +946,17 @@ def load_ontology(context: OpExecutionContext):
     context.log.info(f"Ontology load for {source_ontology} completed successfully!")
 
 
-def _add_related_ids_to_alldocs(
+def _add_linked_instances_to_alldocs(
     temp_collection, context, document_reference_ranged_slots_by_type
 ) -> None:
     """
-    Adds {`_inbound`,`_outbound`} fields to each document in the temporary alldocs collection.
+    Adds {`_upstream`,`_downstream`} fields to each document in the temporary alldocs collection.
 
-    The {`_inbound`,`_outbound`} fields each contain an array of subdocuments, each with fields `id` and `type`.
-    Each subdocument represents a link to any other document that either links to or is linked from
-    the document via document-reference-ranged slots.
+    The {`_upstream`,`_downstream`} fields each contain an array of subdocuments, each with fields `id` and `type`.
+    Each subdocument represents a link to another document that either links to or is linked from the document via
+    document-reference-ranged slots. If document A links to document B, document A is not necessarily "upstream of"
+    document B. Rather, "upstream" and "downstream" are defined by domain semantics. For example, a Study is
+    considered upstream of a Biosample even though the link `associated_studies` goes from a Biosample to a Study.
 
     Args:
         temp_collection: The temporary MongoDB collection to process
@@ -1063,7 +968,7 @@ def _add_related_ids_to_alldocs(
     """
 
     context.log.info(
-        "Building relationships and adding `_inbound` and `_outbound` fields..."
+        "Building relationships and adding `_upstream` and `_downstream` fields..."
     )
 
     # document ID -> type (with "nmdc:" prefix preserved)
@@ -1078,6 +983,7 @@ def _add_related_ids_to_alldocs(
         # Store the full type with prefix intact
         doc_type = doc["type"]
         # For looking up reference slots, we still need the type without prefix
+        # FIXME `document_reference_ranged_slots_by_type` should key on `doc_type`
         doc_type_no_prefix = doc_type[5:] if doc_type.startswith("nmdc:") else doc_type
 
         # Record ID to type mapping - preserve the original type with prefix
@@ -1103,34 +1009,32 @@ def _add_related_ids_to_alldocs(
         f"{len({d for (d, _, _) in relationship_triples})} containing references"
     )
 
-    # The bifurcation of document-reference-ranged slots as "inbound" and "outbound" is essential
+    # The bifurcation of document-reference-ranged slots as "upstream" and "downstream" is essential
     # in order to perform graph traversal and collect all entities "related" to a given entity without
     # recursion "exploding".
     #
     # Note: We are hard-coding this "direction" information here in the Runtime
     #       because the NMDC schema does not currently contain or expose it.
     #
-    # An "inbound" slot is one for which an entity in the domain "was influenced by" (formally,
-    # <https://www.w3.org/ns/prov#wasInfluencedBy>, with typical CURIE prov:wasInfluencedBy) an entity in the range.
-    inbound_document_reference_ranged_slots = [
-        "collected_from",  # a `nmdc:Biosample` was influenced by the `nmdc:Site` from which it was collected.
-        "has_chromatography_configuration",  # a `nmdc:PlannedProcess` was influenced by its `nmdc:Configuration`.
-        "has_input",  # a `nmdc:PlannedProcess` was influenced by a `nmdc:NamedThing`.
-        "has_mass_spectrometry_configuration",  # a `nmdc:PlannedProcess` was influenced by its `nmdc:Configuration`.
-        "instrument_used",  # a `nmdc:PlannedProcess` was influenced by a used `nmdc:Instrument`.
-        "uses_calibration",  # a `nmdc:PlannedProcess` was influenced by `nmdc:CalibrationInformation`.
-        "was_generated_by",  # prov:wasGeneratedBy rdfs:subPropertyOf prov:wasInfluencedBy.
-        "was_informed_by",  # prov:wasInformedBy rdfs:subPropertyOf prov:wasInfluencedBy.
+    # An "upstream" slot is such that the range entity originated, or helped produce, the domain entity.
+    upstream_document_reference_ranged_slots = [
+        "associated_studies",  # when a `nmdc:Study` is upstream of a `nmdc:Biosample`.
+        "collected_from",  # when a `nmdc:Site` is upstream of a `nmdc:Biosample`.
+        "has_chromatography_configuration",  # when a `nmdc:Configuration` is upstream of a `nmdc:PlannedProcess`.
+        "has_input",  # when a `nmdc:NamedThing` is upstream of a `nmdc:PlannedProcess`.
+        "has_mass_spectrometry_configuration",  # when a `nmdc:Configuration` is upstream of a `nmdc:PlannedProcess`.
+        "instrument_used",  # when a `nmdc:Instrument` is upstream of a `nmdc:PlannedProcess`.
+        "part_of",  # when a `nmdc:NamedThing` is upstream of a `nmdc:NamedThing`.
+        "was_generated_by",  # when a `nmdc:DataEmitterProcess` is upstream of a `nmdc:DataObject`.
+        "was_informed_by",  # when a  `nmdc:DataGeneration` is upstream of a `nmdc:WorkflowExecution`.
     ]
-    # An "outbound" slot is one for which an entity in the domain "influences"
-    # (i.e., [owl:inverseOf prov:wasInfluencedBy]) an entity in the range.
-    outbound_document_reference_ranged_slots = [
-        "associated_studies",  # a `nmdc:Biosample` influences a `nmdc:Study`.
-        "calibration_object",  # `nmdc:CalibrationInformation` generates a `nmdc:DataObject`.
-        "generates_calibration",  # a `nmdc:PlannedProcess` generates `nmdc:CalibrationInformation`.
-        "has_output",  # a `nmdc:PlannedProcess` generates a `nmdc:NamedThing`.
-        "in_manifest",  # a `nmdc:DataObject` becomes associated with `nmdc:Manifest`.
-        "part_of",  # a "contained" `nmdc:NamedThing` influences its "container" `nmdc:NamedThing`,
+    # A "downstream" slot is such that the range entity originated from, or is considered part of, the domain entity.
+    downstream_document_reference_ranged_slots = [
+        "calibration_object",  # when a `nmdc:DataObject` is downstream of a `nmdc:CalibrationInformation`.
+        "generates_calibration",  # when a `nmdc:CalibrationInformation` is downstream of a `nmdc:PlannedProcess`.
+        "has_output",  # when a `nmdc:NamedThing` is downstream of a `nmdc:PlannedProcess`.
+        "in_manifest",  # when a `nmdc:Manifest` is downstream of a `nmdc:DataObject`.
+        "uses_calibration",  # when a `nmdc:CalibrationInformation`is part of a `nmdc:PlannedProcess`.
     ]
 
     unique_document_reference_ranged_slot_names = set()
@@ -1138,15 +1042,15 @@ def _add_related_ids_to_alldocs(
         for slot_name in slot_names:
             unique_document_reference_ranged_slot_names.add(slot_name)
     context.log.info(f"{unique_document_reference_ranged_slot_names=}")
-    if len(inbound_document_reference_ranged_slots) + len(
-        outbound_document_reference_ranged_slots
+    if len(upstream_document_reference_ranged_slots) + len(
+        downstream_document_reference_ranged_slots
     ) != len(unique_document_reference_ranged_slot_names):
         raise Failure(
             "Number of detected unique document-reference-ranged slot names does not match "
-            "sum of accounted-for inbound and outbound document-reference-ranged slot names."
+            "sum of accounted-for upstream and downstream document-reference-ranged slot names."
         )
 
-    # Construct, and update documents with, `_incoming` and `_outgoing` field values.
+    # Construct, and update documents with, `_upstream` and `_downstream` field values.
     #
     # manage batching of MongoDB `bulk_write` operations
     bulk_operations, update_count = [], 0
@@ -1154,10 +1058,10 @@ def _add_related_ids_to_alldocs(
 
         # Determine in which respective fields to push this relationship
         # for the subject (doc) and object (ref) of this triple.
-        if slot in inbound_document_reference_ranged_slots:
-            field_for_doc, field_for_ref = "_inbound", "_outbound"
-        elif slot in outbound_document_reference_ranged_slots:
-            field_for_doc, field_for_ref = "_outbound", "_inbound"
+        if slot in upstream_document_reference_ranged_slots:
+            field_for_doc, field_for_ref = "_upstream", "_downstream"
+        elif slot in downstream_document_reference_ranged_slots:
+            field_for_doc, field_for_ref = "_downstream", "_upstream"
         else:
             raise Failure(f"Unknown slot {slot} for document {doc_id}")
 
@@ -1204,14 +1108,6 @@ def _add_related_ids_to_alldocs(
 
     context.log.info(f"Pushed {update_count} updates in total")
 
-    context.log.info("Creating {`_inbound`,`_outbound`} indexes...")
-    temp_collection.create_index("_inbound.id")
-    temp_collection.create_index("_outbound.id")
-    # Create compound indexes to ensure index-covered queries
-    temp_collection.create_index([("_inbound.type", 1), ("_inbound.id", 1)])
-    temp_collection.create_index([("_outbound.type", 1), ("_outbound.id", 1)])
-    context.log.info("Successfully created {`_inbound`,`_outbound`} indexes")
-
 
 # Note: Here, we define a so-called "Nothing dependency," which allows us to (in a graph)
 #       pass an argument to the op (in order to specify the order of the ops in the graph)
@@ -1228,8 +1124,8 @@ def materialize_alldocs(context) -> int:
     2. Create a temporary collection to build the new alldocs collection.
     3. For each document in schema collections, extract `id`, `type`, and document-reference-ranged slot values.
     4. Add a special `_type_and_ancestors` field that contains the class hierarchy for the document's type.
-    5. Add special `_inbound` and `_outbound` fields with subdocuments containing ID and type of related entities.
-    6. Add indexes for `id`, relationship fields, and `{_inbound,_outbound}.type`/`.id` compound indexes.
+    5. Add special `_upstream` and `_downstream` fields with subdocuments containing ID and type of related entities.
+    6. Add indexes for `id`, relationship fields, and `{_upstream,_downstream}{.id,(.type, .id)}` (compound) indexes.
     7. Finally, atomically replace the existing `alldocs` collection with the temporary one.
 
     The `alldocs` collection is scheduled to be updated daily via a scheduled job defined as
@@ -1240,7 +1136,7 @@ def materialize_alldocs(context) -> int:
     `/workflow_executions/{workflow_execution_id}/related_resources` that need to perform graph traversal to find
     related documents. It serves as a denormalized view of the database to make these complex queries more efficient.
 
-    The {`_inbound`,`_outbound`} fields enable efficient index-covered queries to find all entities of specific types
+    The {`_upstream`,`_downstream`} fields enable efficient index-covered queries to find all entities of specific types
     that are related to a given set of source entities, leveraging the `_type_and_ancestors` field for subtype
     expansions.
     """
@@ -1271,6 +1167,9 @@ def materialize_alldocs(context) -> int:
         )
     )
 
+    # FIXME rename to `document_reference_ranged_slots_by_type`
+    # FIXME key on CURIE, e.g. `nmdc:Study`
+    #  (here, not upstream in `cls_slot_map`/`document_referenceable_ranges`, b/c `schema_view` used directly in those)
     document_reference_ranged_slots = defaultdict(list)
     for cls_name, slot_map in cls_slot_map.items():
         for slot_name, slot in slot_map.items():
@@ -1310,12 +1209,12 @@ def materialize_alldocs(context) -> int:
             new_doc = keyfilter(lambda slot: slot in slots_to_include, doc)
 
             new_doc["_type_and_ancestors"] = schema_view.class_ancestors(doc_type)
-            # InsertOne is a method on the py-mongo Client class.
             # Get ancestors without the prefix, but add prefix to each one in the output
             ancestors = schema_view.class_ancestors(doc_type)
             new_doc["_type_and_ancestors"] = [
                 "nmdc:" + a if not a.startswith("nmdc:") else a for a in ancestors
             ]
+            # InsertOne is a pymongo representation of a mongo command.
             write_operations.append(InsertOne(new_doc))
             if len(write_operations) == BULK_WRITE_BATCH_SIZE:
                 _ = temp_alldocs_collection.bulk_write(write_operations, ordered=False)
@@ -1339,19 +1238,28 @@ def materialize_alldocs(context) -> int:
     # so that `temp_alldocs_collection` will be "good to go" on renaming.
     temp_alldocs_collection.create_index("id", unique=True)
     # Add indexes to improve performance of `GET /data_objects/study/{study_id}`:
+    # TODO add indexes on each of `set(document_reference_ranged_slots.values())`.
     slots_to_index = ["has_input", "has_output", "was_informed_by"]
     [temp_alldocs_collection.create_index(slot) for slot in slots_to_index]
     context.log.info(f"created indexes on id, {slots_to_index}.")
 
     # Add related-ids fields to enable efficient relationship traversal
     context.log.info("Adding fields for related ids to documents...")
-    _add_related_ids_to_alldocs(
+    _add_linked_instances_to_alldocs(
         temp_alldocs_collection, context, document_reference_ranged_slots
     )
+    context.log.info("Creating {`_upstream`,`_downstream`} indexes...")
+    temp_alldocs_collection.create_index("_upstream.id")
+    temp_alldocs_collection.create_index("_downstream.id")
+    # Create compound indexes to ensure index-covered queries
+    temp_alldocs_collection.create_index([("_upstream.type", 1), ("_upstream.id", 1)])
+    temp_alldocs_collection.create_index(
+        [("_downstream.type", 1), ("_downstream.id", 1)]
+    )
+    context.log.info("Successfully created {`_upstream`,`_downstream`} indexes")
 
     context.log.info(f"renaming `{temp_alldocs_collection.name}` to `alldocs`...")
     temp_alldocs_collection.rename("alldocs", dropTarget=True)
-
     n_alldocs_documents = mdb.alldocs.estimated_document_count()
     context.log.info(
         f"Rebuilt `alldocs` collection with {n_alldocs_documents} documents."