mtbls-mhd-integration 0.0.11__py3-none-any.whl → 0.0.12__py3-none-any.whl

mtbls2mhd/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "v0.0.11"
+__version__ = "v0.0.12"
 
 import pathlib
 import sys

mtbls2mhd/v0_1/legacy/builder.py

@@ -594,7 +594,7 @@ class MhdLegacyDatasetBuilder:
         data: MetabolightsStudyModel,
         config: Mtbls2MhdConfiguration,
     ):
-        result_file_map = {}
+        result_file_map: dict[str, mhd_domain.ResultFile] = {}
         tsv_format = create_cv_term_object(
             type_="descriptor", accession="EDAM:3475", source="EDAM", name="TSV"
         )
@@ -1650,54 +1650,103 @@ class MhdLegacyDatasetBuilder:
         mhd_builder: MhDatasetBuilder,
         mhd_study: mhd_domain.Study,
         data: MetabolightsStudyModel,
+        result_files: dict[str, mhd_domain.ResultFile],
     ):
-        for file_name, maf_file in data.metabolite_assignments.items():
-            if maf_file.table.data.get("metabolite_identification"):
-                identifiers = maf_file.table.data.get("database_identifier")
-                for idx, name in enumerate(
-                    maf_file.table.data["metabolite_identification"]
-                ):
-                    if not name:
-                        continue
-                    met = mhd_domain.Metabolite(
-                        name=name,
-                    )
-                    if identifiers and identifiers[idx]:
-                        value = identifiers[idx]
-                        identifier = None
-                        if value.startswith("CHEBI"):
-                            identifier = create_cv_term_value_object(
-                                type_="metabolite-identifier",
-                                source="CHEMINF",
-                                accession="CHEMINF:000407",
-                                name="ChEBI identifier",
-                                value=value,
-                            )
-                        elif value.startswith("HMDB"):
-                            identifier = create_cv_term_value_object(
-                                type_="metabolite-identifier",
-                                source="CHEMINF",
-                                accession="CHEMINF:000408",
-                                name="HMDB identifier",
-                                value=value.replace(":", ""),
-                            )
+        for maf_filename, maf_file in data.metabolite_assignments.items():
+            if not maf_file.table.data.get("metabolite_identification"):
+                continue
+            result_file = result_files.get(maf_filename)
+            for idx, name in enumerate(
+                maf_file.table.data["metabolite_identification"]
+            ):
+                if not name or not name.strip():
+                    continue
+                met = mhd_domain.Metabolite(name=name)
+                assignments = {}
+                data: dict[str, str] = maf_file.table.data
+                submitted_identifiers = []
+                assigned_chebi_identifiers = []
+                assigned_refmet_identifiers = []
+                if maf_file.table.data.get("database_identifier"):
+                    submitted_identifiers = [
+                        x.strip()
+                        for x in data["database_identifier"][idx].split("|")
+                        if x
+                    ]
+                if maf_file.table.data.get("assigned_chebi_identifier"):
+                    assigned_chebi_identifiers = [
+                        x.strip()
+                        for x in data["assigned_chebi_identifier"][idx].split("|")
+                        if x
+                    ]
+                if maf_file.table.data.get("assigned_refmet_identifier"):
+                    assigned_refmet_identifiers = [
+                        x.strip()
+                        for x in data["assigned_refmet_identifier"][idx].split("|")
+                        if x
+                    ]
 
-                        if identifier:
-                            mhd_builder.add(identifier)
-                            # met.identifier_refs = [identifier.id_]
-                            mhd_builder.link(
-                                met,
-                                "identified-as",
-                                identifier,
-                                reverse_relationship_name="reported-identifier-of",
-                            )
-                    mhd_builder.add(met)
+                for identifiers in [
+                    (submitted_identifiers, ""),
+                    (assigned_chebi_identifiers, "CHEBI"),
+                    (assigned_refmet_identifiers, "REFMET"),
+                ]:
+                    for identifiers, compound_source in assignments:
+                        if not identifiers:
+                            continue
+                        for identifier_value in identifiers:
+                            identifier = None
+                            if (
+                                compound_source == "CHEBI"
+                                or identifier_value.upper().startswith("CHEBI")
+                            ):
+                                identifier = create_cv_term_value_object(
+                                    type_="metabolite-identifier",
+                                    source="CHEMINF",
+                                    accession="CHEMINF:000407",
+                                    name="ChEBI identifier",
+                                    value=identifier_value,
+                                )
+                            elif identifier_value.upper().startswith("HMDB"):
+                                identifier = create_cv_term_value_object(
+                                    type_="metabolite-identifier",
+                                    source="CHEMINF",
+                                    accession="CHEMINF:000408",
+                                    name="HMDB identifier",
+                                    value=identifier_value,
+                                )
+                            elif compound_source == "REFMET":
+                                identifier = create_cv_term_value_object(
+                                    type_="metabolite-identifier",
+                                    source="REFMET",
+                                    accession="",
+                                    name="RefMet identifier",
+                                    value=identifier_value,
+                                )
+
+                            if identifier:
+                                mhd_builder.add(identifier)
+                                mhd_builder.link(
+                                    met,
+                                    "identified-as",
+                                    identifier,
+                                    reverse_relationship_name="reported-identifier-of",
+                                )
+                mhd_builder.add(met)
+                if result_file:
                     mhd_builder.link(
-                        mhd_study,
+                        result_file,
                         "reports",
                         met,
                         reverse_relationship_name="reported-in",
                     )
+                result_file
+                mhd_builder.link(
+                    mhd_study,
+                    "reports",
+                    met,
+                    reverse_relationship_name="reported-in",
+                )
 
     def add_assays(
         self,
@@ -2030,11 +2079,12 @@ class MhdLegacyDatasetBuilder:
             self.add_protocols(mhd_builder, mhd_study, study)
 
             self.add_keywords(mhd_builder, mhd_study, study)
-            self.add_reported_metabolites(mhd_builder, mhd_study, data)
 
             result_files = self.add_result_files(
                 mhd_builder, mhd_study, data, config=config
             )
+            self.add_reported_metabolites(mhd_builder, mhd_study, data, result_files)
+
             files_map = self.add_data_files(
                 mhd_builder,
                 mhd_study,

mtbls2mhd/v0_1/legacy/convertor.py

@@ -38,7 +38,7 @@ class LegacyProfileV01Convertor(BaseMhdConvertor):
         )
         try:
             success, message = mhd_dataset_builder.build(
-                mhd_id=mhd_identifier,
+                mhd_id=None,
                 mtbls_study_id=repository_identifier,
                 mtbls_study_path=mtbls_study_path,
                 mtbls_study_repository_url=mtbls_study_repository_url,

{mtbls_mhd_integration-0.0.11.dist-info → mtbls_mhd_integration-0.0.12.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mtbls-mhd-integration
-Version: 0.0.11
+Version: 0.0.12
 Summary: MetaboLights - MetabolomicsHub Integration
 Author-email: MetaboLights Team <metabolights-help@ebi.ac.uk>
 License-Expression: Apache-2.0
@@ -9,7 +9,7 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: asyncpg>=0.30.0
 Requires-Dist: metabolights-utils>=1.4.16
-Requires-Dist: mhd-model>=0.1.39
+Requires-Dist: mhd-model>=0.1.41
 Requires-Dist: psycopg[binary,pool]>=3.3.2
 Requires-Dist: pydantic>=2.12.4
 Requires-Dist: pydantic-settings>=2.10.1

{mtbls_mhd_integration-0.0.11.dist-info → mtbls_mhd_integration-0.0.12.dist-info}/RECORD

@@ -1,4 +1,4 @@
-mtbls2mhd/__init__.py,sha256=0BFPsjHigJ2_UsW5db4g-LyFf8QbDFBUqvXv4BwGiFw,158
+mtbls2mhd/__init__.py,sha256=lndk-uwz1jRMVCVy3G7BTa0Mz9464n8uyyOOlNSzYSs,158
 mtbls2mhd/config.py,sha256=BjOqAyfDhp9byoFjJz70xh4HRR8pu1yrm_5jweqygSI,2310
 mtbls2mhd/convertor_factory.py,sha256=4loatqIRIvIhcaeIS0cSonJNYJu47o56ZllX6593ypk,1133
 mtbls2mhd/commands/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -8,16 +8,16 @@ mtbls2mhd/commands/create_mhd_file.py,sha256=0sDr-Cm0JhhEB5V1g66uoag3rlcaAnGP8Md
 mtbls2mhd/commands/validate.py,sha256=iwIKegviRxdH0r8scRXbDISlwQUzAq5uVoCHinU7x6Q,473
 mtbls2mhd/v0_1/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 mtbls2mhd/v0_1/legacy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-mtbls2mhd/v0_1/legacy/builder.py,sha256=IN3z8I3VsDkmLv_XMeDaLVkq88d9yl6hRWilNtNKo-U,85914
-mtbls2mhd/v0_1/legacy/convertor.py,sha256=M4-3NB56C0DN8H5vmM0yC0J_qW87ZImjjL1vzDEBc5s,2208
+mtbls2mhd/v0_1/legacy/builder.py,sha256=ou_YOtRNFjlLCvZYJc7JphczEbpOCrX5rqndCMxRFXI,88325
+mtbls2mhd/v0_1/legacy/convertor.py,sha256=Nu6xJvEk8WsRQJQFoxM5eo-y46tVfWV8EkedVoqI9rI,2198
 mtbls2mhd/v0_1/legacy/db_metadata_collector.py,sha256=UGk1AeST1NQ9lWwy_sYZxaaWs0ajgaKELDJtXJ4-Uco,13071
 mtbls2mhd/v0_1/legacy/folder_metadata_collector.py,sha256=QwtXI9rBvdh6pxILQDHymIwYDqzGuMxqdOcqtdAObME,7538
 mtbls2mhd/v0_1/legacy/mtbls_study_schema.py,sha256=gUTbRmI8GfHI5leLiw8dxsmWnV3NnWw5RPX_LQWRFRQ,3162
 mtbls2mhd/v0_1/ms/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 mtbls2mhd/v0_1/ms/convertor.py,sha256=kLIUpxOrH6hcs2Y9Bq1D0Mdvypg40pLyEJpHtGj6H_g,89
-mtbls_mhd_integration-0.0.11.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-mtbls_mhd_integration-0.0.11.dist-info/METADATA,sha256=e8AHuzd119EKD4I_3au3GbJ2zPNGzQ4e6MZhB5HLsz4,688
-mtbls_mhd_integration-0.0.11.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-mtbls_mhd_integration-0.0.11.dist-info/entry_points.txt,sha256=WQjM4flaYMyvHyv9zGKjCVk1i1_FGdNlhTmFVGgLgxs,61
-mtbls_mhd_integration-0.0.11.dist-info/top_level.txt,sha256=b7pI95n6HIQMFXDD0yL1NwldiDc-XdeWql4Iw-uYygQ,10
-mtbls_mhd_integration-0.0.11.dist-info/RECORD,,
+mtbls_mhd_integration-0.0.12.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+mtbls_mhd_integration-0.0.12.dist-info/METADATA,sha256=p307N8kU4TvWHnKv_28VXUSp7N2fi0ZAh5JN_7ILx08,688
+mtbls_mhd_integration-0.0.12.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+mtbls_mhd_integration-0.0.12.dist-info/entry_points.txt,sha256=WQjM4flaYMyvHyv9zGKjCVk1i1_FGdNlhTmFVGgLgxs,61
+mtbls_mhd_integration-0.0.12.dist-info/top_level.txt,sha256=b7pI95n6HIQMFXDD0yL1NwldiDc-XdeWql4Iw-uYygQ,10
+mtbls_mhd_integration-0.0.12.dist-info/RECORD,,