Merge pull request #152 from AllenNeuralDynamics/release-v0.18.0

Release v0.18.0

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "aind-metadata-mapper"
-description = "Generated from aind-library-template"
+description = "Package to manage mapping of source data into aind-data-schema metadata files."
 license = {text = "MIT"}
 requires-python = ">=3.8"
 authors = [
@@ -17,10 +17,8 @@ readme = "README.md"
 dynamic = ["version"]
 
 dependencies = [
-    "aind-data-schema==1.0.0",
     "aind-data-schema-models>=0.3.2",
-    "pydantic-settings>=2.0",
-    "pydantic<2.9"
+    "pydantic-settings>=2.0"
 ]
 
 [project.optional-dependencies]
@@ -44,12 +42,19 @@ all = [
     "aind-metadata-mapper[u19]",
 ]
 
+schema = [
+    "aind-data-schema>=1.0.0,<2.0",
+    "pydantic<2.9"
+]
+
 bergamo = [
+    "aind-metadata-mapper[schema]",
     "scanimage-tiff-reader==1.4.1.4",
     "numpy >= 1.26.4",
 ]
 
 bruker = [
+    "aind-metadata-mapper[schema]",
     "bruker2nifti==1.0.4",
 ]
 
@@ -60,6 +65,7 @@ mesoscope = [
 ]
 
 openephys = [
+    "aind-metadata-mapper[schema]",
     "h5py >= 3.11.0",
     "np_session >= 0.1.39",
     "npc_ephys >= 0.1.18 ; python_version >= '3.9'",
@@ -69,10 +75,12 @@ openephys = [
 ]
 
 dynamicrouting = [
+    "aind-metadata-mapper[schema]",
     "pyyaml >= 6.0.0",
 ]
 
 u19 = [
+    "aind-metadata-mapper[schema]",
     "pandas >= 2.2.2",
 ]
 

src/aind_metadata_mapper/__init__.py

@@ -1,3 +1,3 @@
 """Init package"""
 
-__version__ = "0.17.1"
+__version__ = "0.18.0"

src/aind_metadata_mapper/bergamo/models.py

@@ -1,29 +1,16 @@
 """Module defining JobSettings for Bergamo ETL"""
 
 from decimal import Decimal
-from pathlib import Path
 from typing import List, Literal, Optional
 
-from pydantic import Field
-
-from aind_metadata_mapper.core import BaseJobSettings
+from aind_metadata_mapper.core_models import BaseJobSettings
 
 
 class JobSettings(BaseJobSettings):
     """Data that needs to be input by user. Can be pulled from env vars with
     BERGAMO prefix or set explicitly."""
 
     job_settings_name: Literal["Bergamo"] = "Bergamo"
-    input_source: Path = Field(
-        ..., description="Directory of files that need to be parsed."
-    )
-    output_directory: Optional[Path] = Field(
-        default=None,
-        description=(
-            "Directory where to save the json file to. If None, then json"
-            " contents will be returned in the Response message."
-        ),
-    )
     # mandatory fields:
     experimenter_full_name: List[str]
     subject_id: str

src/aind_metadata_mapper/bergamo/session.py

@@ -49,7 +49,8 @@
 from ScanImageTiffReader import ScanImageTiffReader
 
 from aind_metadata_mapper.bergamo.models import JobSettings
-from aind_metadata_mapper.core import GenericEtl, JobResponse
+from aind_metadata_mapper.core import GenericEtl
+from aind_metadata_mapper.core_models import JobResponse
 
 
 # This class makes it easier to flag which tif files are which expected type
@@ -80,17 +81,16 @@ class RawImageInfo:
 class BergamoEtl(GenericEtl[JobSettings]):
     """Class to manage transforming bergamo data files into a Session object"""
 
-    def __init__(
-        self,
-        job_settings: Union[JobSettings, str],
-    ):
+    # TODO: Deprecate this constructor. Use GenericEtl constructor instead
+    def __init__(self, job_settings: Union[JobSettings, str]):
         """
         Class constructor for Base etl class.
         Parameters
         ----------
         job_settings: Union[JobSettings, str]
           Variables for a particular session
         """
+
         if isinstance(job_settings, str):
             job_settings_model = JobSettings.model_validate_json(job_settings)
         else:
@@ -1169,5 +1169,6 @@ def from_args(cls, args: list):
 
 if __name__ == "__main__":
     sys_args = sys.argv[1:]
-    etl = BergamoEtl.from_args(sys_args)
+    main_job_settings = JobSettings.from_args(sys_args)
+    etl = BergamoEtl(job_settings=main_job_settings)
     etl.run_job()

src/aind_metadata_mapper/bruker/models.py

@@ -1,28 +1,18 @@
 """Module defining JobSettings for Bruker ETL"""
-
 from pathlib import Path
-from typing import List, Literal, Optional
+from typing import List, Literal, Optional, Union
 
-from aind_data_schema.components.devices import (
-    MagneticStrength,
-    ScannerLocation,
-)
 from pydantic import Field
 
-from aind_metadata_mapper.core import BaseJobSettings
+from aind_metadata_mapper.core_models import BaseJobSettings
 
 
 class JobSettings(BaseJobSettings):
     """Data that needs to be input by user."""
 
     job_settings_name: Literal["Bruker"] = "Bruker"
-    data_path: Path
-    output_directory: Optional[Path] = Field(
-        default=None,
-        description=(
-            "Directory where to save the json file to. If None, then json"
-            " contents will be returned in the Response message."
-        ),
+    data_path: Optional[Union[Path, str]] = Field(
+        default=None, description=("Deprecated, use input_source instead.")
     )
     experimenter_full_name: List[str]
     protocol_id: str = Field(default="", description="Protocol ID")
@@ -34,8 +24,8 @@ class JobSettings(BaseJobSettings):
     primary_scan_number: int
     setup_scan_number: int
     scanner_name: str
-    scan_location: ScannerLocation
-    magnetic_strength: MagneticStrength
+    scan_location: str
+    magnetic_strength: int
     subject_id: str
     iacuc_protocol: str
     session_notes: str

src/aind_metadata_mapper/bruker/session.py

@@ -1,5 +1,4 @@
 """Sets up the MRI ingest ETL"""
-
 import argparse
 import json
 import logging
@@ -15,7 +14,11 @@
     Scale3dTransform,
     Translation3dTransform,
 )
-from aind_data_schema.components.devices import Scanner
+from aind_data_schema.components.devices import (
+    MagneticStrength,
+    Scanner,
+    ScannerLocation,
+)
 from aind_data_schema.core.session import (
     MRIScan,
     MriScanSequence,
@@ -29,7 +32,8 @@
 from bruker2nifti._metadata import BrukerMetadata
 
 from aind_metadata_mapper.bruker.models import JobSettings
-from aind_metadata_mapper.core import GenericEtl, JobResponse
+from aind_metadata_mapper.core import GenericEtl
+from aind_metadata_mapper.core_models import JobResponse
 
 DATETIME_FORMAT = "%H:%M:%S %d %b %Y"
 LENGTH_FORMAT = "%Hh%Mm%Ss%fms"
@@ -38,6 +42,7 @@
 class MRIEtl(GenericEtl[JobSettings]):
     """Class for MRI ETL process."""
 
+    # TODO: Deprecate this constructor. Use GenericEtl constructor instead
     def __init__(self, job_settings: Union[JobSettings, str]):
         """
         Class constructor for Base etl class.
@@ -51,8 +56,14 @@ def __init__(self, job_settings: Union[JobSettings, str]):
             job_settings_model = JobSettings.model_validate_json(job_settings)
         else:
             job_settings_model = job_settings
+        if (
+            job_settings_model.data_path is not None
+            and job_settings_model.input_source is None
+        ):
+            job_settings_model.input_source = job_settings_model.data_path
         super().__init__(job_settings=job_settings_model)
 
+    # TODO: deprecate method
     @classmethod
     def from_args(cls, args: list):
         """
@@ -62,7 +73,10 @@ def from_args(cls, args: list):
         args : list
         A list of command line arguments to parse.
         """
-
+        logging.warning(
+            "This method will be removed in future versions. "
+            "Please use JobSettings.from_args instead."
+        )
         parser = argparse.ArgumentParser()
         parser.add_argument(
             "-u",
@@ -90,7 +104,7 @@ def from_args(cls, args: list):
     def _extract(self) -> BrukerMetadata:
         """Extract the data from the bruker files."""
 
-        metadata = BrukerMetadata(self.job_settings.data_path)
+        metadata = BrukerMetadata(self.job_settings.input_source)
         metadata.parse_scans()
         metadata.parse_subject()
 
@@ -273,15 +287,20 @@ def make_model_from_scan(
 
         scale = self.get_scale(cur_method)
 
+        scan_location = ScannerLocation(self.job_settings.scan_location)
+        magnetic_strength = MagneticStrength(
+            self.job_settings.magnetic_strength
+        )
+
         try:
             return MRIScan(
                 scan_index=scan_index,
                 scan_type=ScanType(scan_type),
                 primary_scan=primary_scan,
                 mri_scanner=Scanner(
                     name=self.job_settings.scanner_name,
-                    scanner_location=self.job_settings.scan_location,
-                    magnetic_strength=self.job_settings.magnetic_strength,
+                    scanner_location=scan_location,
+                    magnetic_strength=magnetic_strength,
                     magnetic_strength_unit="T",
                 ),
                 scan_sequence_type=scan_sequence,
@@ -306,5 +325,6 @@ def make_model_from_scan(
 
 if __name__ == "__main__":
     sys_args = sys.argv[1:]
-    etl = MRIEtl.from_args(sys_args)
+    main_job_settings = JobSettings.from_args(sys_args)
+    etl = MRIEtl(job_settings=main_job_settings)
     etl.run_job()