nmdc-runtime 2.0.1__py3-none-any.whl → 2.1.0__py3-none-any.whl

nmdc_runtime/site/ops.py

@@ -1134,6 +1134,7 @@ def materialize_alldocs(context) -> int:
     # /data_objects/study/{study_id} endpoint
     mdb.alldocs.create_index("has_input")
     mdb.alldocs.create_index("has_output")
+    mdb.alldocs.create_index("was_informed_by")
     context.log.info(
         f"refreshed {mdb.alldocs} collection with {mdb.alldocs.estimated_document_count()} docs."
     )

nmdc_runtime/site/translation/gold_translator.py

@@ -108,7 +108,9 @@ class GoldStudyTranslator(Translator):
             for id in self._project_ids_by_biosample_id[gold_biosample_id]
         )
         return [
-            self._get_curie("biosample", project["ncbiBioSampleAccession"])
+            self._ensure_curie(
+                project["ncbiBioSampleAccession"], default_prefix="biosample"
+            )
             for project in biosample_projects
             if project["ncbiBioSampleAccession"]
         ]
@@ -471,7 +473,9 @@ class GoldStudyTranslator(Translator):
         """
         return nmdc.Study(
             description=gold_study.get("description"),
-            gold_study_identifiers=self._get_curie("gold", gold_study["studyGoldId"]),
+            gold_study_identifiers=self._ensure_curie(
+                gold_study["studyGoldId"], default_prefix="gold"
+            ),
             id=nmdc_study_id,
             name=gold_study.get("studyName"),
             principal_investigator=self._get_pi(gold_study),
@@ -522,7 +526,9 @@ class GoldStudyTranslator(Translator):
             env_local_scale=self._get_env_term_value(gold_biosample, "envoLocalScale"),
             env_medium=self._get_env_term_value(gold_biosample, "envoMedium"),
             geo_loc_name=self._get_text_value(gold_biosample, "geoLocation"),
-            gold_biosample_identifiers=self._get_curie("gold", gold_biosample_id),
+            gold_biosample_identifiers=self._ensure_curie(
+                gold_biosample_id, default_prefix="gold"
+            ),
             habitat=gold_biosample.get("habitat"),
             host_name=gold_biosample.get("hostName"),
             host_taxid=self._get_host_taxid(gold_biosample),
@@ -579,8 +585,8 @@ class GoldStudyTranslator(Translator):
         return nmdc.NucleotideSequencing(
             id=nmdc_nucleotide_sequencing_id,
             name=gold_project.get("projectName"),
-            gold_sequencing_project_identifiers=self._get_curie(
-                "gold", gold_project_id
+            gold_sequencing_project_identifiers=self._ensure_curie(
+                gold_project_id, default_prefix="gold"
             ),
             ncbi_project_name=gold_project.get("projectName"),
             type="nmdc:NucleotideSequencing",

nmdc_runtime/site/translation/submission_portal_translator.py

@@ -189,7 +189,21 @@ class SubmissionPortalTranslator(Translator):
         if not gold_study_id:
             return None
 
-        return [self._get_curie("GOLD", gold_study_id)]
+        return [self._ensure_curie(gold_study_id, default_prefix="gold")]
+
+    def _get_jgi_study_identifiers(
+        self, metadata_submission: JSON_OBJECT
+    ) -> Union[List[str], None]:
+        """Construct a JGI proposal CURIE from the multiomics from data
+
+        :param metadata_submission: submission portal entry
+        :return: JGI proposal CURIE
+        """
+        jgi_study_id = get_in(["multiOmicsForm", "JGIStudyId"], metadata_submission)
+        if not jgi_study_id:
+            return None
+
+        return [self._ensure_curie(jgi_study_id, default_prefix="jgi.proposal")]
 
     def _get_quantity_value(
         self, raw_value: Optional[str], unit: Optional[str] = None
@@ -410,9 +424,6 @@ class SubmissionPortalTranslator(Translator):
         :return: nmdc:Study object
         """
         return nmdc.Study(
-            alternative_identifiers=self._get_from(
-                metadata_submission, ["multiOmicsForm", "JGIStudyId"]
-            ),
             alternative_names=self._get_from(
                 metadata_submission, ["multiOmicsForm", "alternativeNames"]
             ),
@@ -436,6 +447,9 @@ class SubmissionPortalTranslator(Translator):
             insdc_bioproject_identifiers=self._get_from(
                 metadata_submission, ["multiOmicsForm", "NCBIBioProjectId"]
             ),
+            jgi_portal_study_identifiers=self._get_jgi_study_identifiers(
+                metadata_submission
+            ),
             name=self._get_from(metadata_submission, ["studyForm", "studyName"]),
             notes=self._get_from(metadata_submission, ["studyForm", "notes"]),
             principal_investigator=self._get_pi(metadata_submission),

nmdc_runtime/site/translation/translator.py

@@ -14,8 +14,15 @@ class Translator(ABC):
     def _index_by_id(self, collection, id):
         return {item[id]: item for item in collection}
 
-    def _get_curie(self, prefix: str, local: str) -> str:
-        return f"{prefix}:{local}"
+    @staticmethod
+    def _ensure_curie(identifier: str, *, default_prefix: str) -> str:
+        identifier_parts = identifier.split(":", 1)
+
+        # Don't add prefix if identifier is already a CURIE
+        if len(identifier_parts) == 2:
+            return identifier
+
+        return f"{default_prefix}:{identifier_parts[0]}"
 
     @abstractmethod
     def get_database(self) -> nmdc.Database:

{nmdc_runtime-2.0.1.dist-info → nmdc_runtime-2.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nmdc_runtime
-Version: 2.0.1
+Version: 2.1.0
 Summary: A runtime system for NMDC data management and orchestration
 Home-page: https://github.com/microbiomedata/nmdc-runtime
 Author: Donny Winston
@@ -114,7 +114,9 @@ cp .env.example .env
 Create environment variables in your shell session, based upon the contents of the `.env` file.
 
 ```shell
-export $(grep -v '^#' .env | xargs)
+set -a # automatically export all variables
+source .env
+set +a
 ```
 
 If you are connecting to resources that require an SSH tunnel—for example, a MongoDB server that is only accessible on the NERSC network—set up the SSH tunnel.

{nmdc_runtime-2.0.1.dist-info → nmdc_runtime-2.1.0.dist-info}/RECORD

@@ -37,7 +37,7 @@ nmdc_runtime/minter/entrypoints/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm
 nmdc_runtime/minter/entrypoints/fastapi_app.py,sha256=JC4thvzfFwRc1mhWQ-kHy3yvs0SYxF6ktE7LXNCwqlI,4031
 nmdc_runtime/site/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nmdc_runtime/site/graphs.py,sha256=ZHglSPwVHfXzdgR2CGvmbzLLbmsijloU58XvIe9Thjs,13996
-nmdc_runtime/site/ops.py,sha256=EOR4VjRoFoaI2odKRgcgtGxngPsx2U0H45zzWrvQzf8,44603
+nmdc_runtime/site/ops.py,sha256=6P3kn4BygY8LKD_OpfKX2U0AYQKDlB2jw12Yn-hEmD0,44651
 nmdc_runtime/site/repository.py,sha256=rDtwUjozhyOxlkuF9HvaheOQDQWkgZYqVtsB50BcUp4,39121
 nmdc_runtime/site/resources.py,sha256=ZSH1yvA-li0R7Abc22_v0XLbjBYf5igETr2G01J3hnc,17557
 nmdc_runtime/site/util.py,sha256=zAY0oIY7GRf63ecqWelmS27N7PVrAXVwEhtnpescBSw,1415
@@ -59,23 +59,23 @@ nmdc_runtime/site/normalization/gold.py,sha256=iISDD4qs4d6uLhv631WYNeQVOzY5DO201
 nmdc_runtime/site/translation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nmdc_runtime/site/translation/emsl.py,sha256=-aCTJTSCNaK-Koh8BE_4fTf5nyxP1KkquR6lloLEJl0,1245
 nmdc_runtime/site/translation/gold.py,sha256=R3W99sdQb7Pgu_esN7ruIC-tyREQD_idJ4xCzkqWuGw,1622
-nmdc_runtime/site/translation/gold_translator.py,sha256=ln0BXI2v7pTgui3uNbl8tU1iOkfDZH4Z8SCaFLuuTI8,29438
+nmdc_runtime/site/translation/gold_translator.py,sha256=wkl1WwJ45EFwz73l_-t0D9Y3SilctDC1obTieY0eqxM,29600
 nmdc_runtime/site/translation/jgi.py,sha256=qk878KhIw674TkrVfbl2x1QJrKi3zlvE0vesIpe9slM,876
 nmdc_runtime/site/translation/neon_benthic_translator.py,sha256=QIDqYLuf-NlGY9_88gy_5qTswkei3OfgJ5AOFpEXzJo,23985
 nmdc_runtime/site/translation/neon_soil_translator.py,sha256=Rol0g67nVBGSBySUzpfdW4Fwes7bKtvnlv2g5cB0aTI,38550
 nmdc_runtime/site/translation/neon_surface_water_translator.py,sha256=MQgjIfWPgoRe-bhzyfqHSe2mZwFsjcwjdT8tNqpIhlc,27729
 nmdc_runtime/site/translation/neon_utils.py,sha256=d00o7duKKugpLHmsEifNbp4WjeC4GOqcgw0b5qlCg4I,5549
-nmdc_runtime/site/translation/submission_portal_translator.py,sha256=6QkEfz-RN8y_aHckUWLUuQz3DslbvFck2Vyu7funHrQ,29152
-nmdc_runtime/site/translation/translator.py,sha256=xM9dM-nTgSWwu5HFoUVNHf8kqk9iiH4PgWdSx4OKxEk,601
+nmdc_runtime/site/translation/submission_portal_translator.py,sha256=FVBqCvk6NAJIA22IhtFOTyvAQIiFN3KsznHc5zmOG40,29676
+nmdc_runtime/site/translation/translator.py,sha256=V6Aq0y03LoQ4LTL2iHDHxGTh_eMjOmDJJSwNHSrp2wo,837
 nmdc_runtime/site/translation/util.py,sha256=w_l3SiExGsl6cXRqto0a_ssDmHkP64ITvrOVfPxmNpY,4366
 nmdc_runtime/site/validation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nmdc_runtime/site/validation/emsl.py,sha256=OG20mv_3E2rkQqTQtYO0_SVRqFb-Z_zKCiAVbty6Wl0,671
 nmdc_runtime/site/validation/gold.py,sha256=Z5ZzYdjERbrJ2Tu06d0TDTBSfwaFdL1Z23Rl-YkZ2Ow,803
 nmdc_runtime/site/validation/jgi.py,sha256=LdJfhqBVHWCDp0Kzyk8eJZMwEI5NQ-zuTda31BcGwOA,1299
 nmdc_runtime/site/validation/util.py,sha256=GGbMDSwR090sr_E_fHffCN418gpYESaiot6XghS7OYk,3349
-nmdc_runtime-2.0.1.dist-info/LICENSE,sha256=VWiv65r7gHGjgtr3jMJYVmQny5GRpQ6H-W9sScb1x70,2408
-nmdc_runtime-2.0.1.dist-info/METADATA,sha256=zYs4CYP5zWAJdBu6go5MvGlWUZ1ZD14uFLIDFr6iHss,7302
-nmdc_runtime-2.0.1.dist-info/WHEEL,sha256=OVMc5UfuAQiSplgO0_WdW7vXVGAt9Hdd6qtN4HotdyA,91
-nmdc_runtime-2.0.1.dist-info/entry_points.txt,sha256=JxdvOnvxHK_8046cwlvE30s_fV0-k-eTpQtkKYA69nQ,224
-nmdc_runtime-2.0.1.dist-info/top_level.txt,sha256=b0K1s09L_iHH49ueBKaLrB5-lh6cyrSv9vL6x4Qvyz8,13
-nmdc_runtime-2.0.1.dist-info/RECORD,,
+nmdc_runtime-2.1.0.dist-info/LICENSE,sha256=VWiv65r7gHGjgtr3jMJYVmQny5GRpQ6H-W9sScb1x70,2408
+nmdc_runtime-2.1.0.dist-info/METADATA,sha256=AfyiRup3egPo7_udHDVKq4gyna0z4C2pfWMz33NeDXw,7329
+nmdc_runtime-2.1.0.dist-info/WHEEL,sha256=R06PA3UVYHThwHvxuRWMqaGcr-PuniXahwjmQRFMEkY,91
+nmdc_runtime-2.1.0.dist-info/entry_points.txt,sha256=JxdvOnvxHK_8046cwlvE30s_fV0-k-eTpQtkKYA69nQ,224
+nmdc_runtime-2.1.0.dist-info/top_level.txt,sha256=b0K1s09L_iHH49ueBKaLrB5-lh6cyrSv9vL6x4Qvyz8,13
+nmdc_runtime-2.1.0.dist-info/RECORD,,

{nmdc_runtime-2.0.1.dist-info → nmdc_runtime-2.1.0.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.2.0)
+Generator: setuptools (75.5.0)
 Root-Is-Purelib: true
 Tag: py3-none-any