Merge branch 'fix-sset-metadata' into 'master'

fix [CSVDatabase]: defining templates metadata.

Closes #191

See merge request !320

src/bob/bio/base/database/csv_database.py

@@ -195,6 +195,7 @@ class CSVDatabase(FileListDatabase, Database):
         protocol: str,
         dataset_protocols_path: Optional[str] = None,
         transformer: Optional[sklearn.pipeline.Pipeline] = None,
+        templates_metadata: Optional[list[str]] = None,
         annotation_type: Optional[str] = None,
         fixed_positions: Optional[dict[str, tuple[float, float]]] = None,
         memory_demanding=False,
@@ -213,6 +214,10 @@ class CSVDatabase(FileListDatabase, Database):
         transformer
             An sklearn pipeline or equivalent transformer that handles some light
             preprocessing of the samples (This will always run locally).
+        templates_metadata
+            Metadata that originate from the samples and must be present in the
+            templates (SampleSet) e.g. ``["gender", "age"]``. This should be metadata
+            that is common to all the samples in a template.
         annotation_type
             A string describing the annotations passed to the annotation loading
             function
@@ -243,6 +248,10 @@ class CSVDatabase(FileListDatabase, Database):
         else:
             self.score_all_vs_all = False
 
+        self.templates_metadata = []
+        if templates_metadata is not None:
+            self.templates_metadata = templates_metadata
+
     def list_file(self, group: str, name: str) -> TextIO:
         """Returns a definition file containing one sample per row.
 
@@ -307,11 +316,16 @@ class CSVDatabase(FileListDatabase, Database):
             # we add that as well.
             samples = list(samples_for_template_id)
             subject_id = samples[0].subject_id
+            metadata = {
+                m: getattr(samples[0], m) for m in self.templates_metadata
+            }
             sample_sets.append(
                 SampleSet(
                     samples,
                     template_id=template_id,
                     subject_id=subject_id,
+                    key=f"template_{template_id}",
+                    **metadata,
                 )
             )
         validate_bio_samples(sample_sets)

tests/data/example_csv_filelist/protocol_only_dev_metadata/dev/for_enrolling.csv

-path,subject_id,template_id,metadata_1,metadata_2
-data/model3_session1_sample1,3,3,F,10
-data/model3_session1_sample2,3,3,F,10
-data/model3_session1_sample3,3,3,F,10
-data/model3_session2_sample1,3,3,F,10
-data/model4_session1_sample1,4,4,M,30
-data/model4_session1_sample2,4,4,M,30
-data/model4_session1_sample3,4,4,M,30
-data/model4_session2_sample1,4,4,M,30
+path,subject_id,template_id,metadata_1,metadata_2,sample_metadata,subject_metadata
+data/model3_session1_sample1,3,3,F,10,1,A
+data/model3_session1_sample2,3,3,F,10,2,A
+data/model3_session1_sample3,3,3,F,10,3,A
+data/model3_session2_sample1,3,3,F,10,4,A
+data/model4_session1_sample1,4,4,M,30,5,B
+data/model4_session1_sample2,4,4,M,30,6,B
+data/model4_session1_sample3,4,4,M,30,7,B
+data/model4_session2_sample1,4,4,M,30,8,B

tests/data/example_csv_filelist/protocol_only_dev_metadata/dev/for_probing.csv

-path,subject_id,template_id,metadata_1,metadata_2
-data/model3_session3_sample1,3,3,F,10
-data/model3_session3_sample2,3,3,F,10
-data/model3_session3_sample3,3,3,F,10
-data/model3_session4_sample1,3,3,F,10
-data/model4_session3_sample1,4,4,M,30
-data/model4_session3_sample2,4,4,M,30
-data/model4_session3_sample1,4,4,M,30
-data/model4_session3_sample2,4,4,M,30
-data/model4_session3_sample3,4,4,M,30
-data/model4_session4_sample1,4,4,M,30
+path,subject_id,template_id,metadata_1,metadata_2,sample_metadata,subject_metadata
+data/model3_session3_sample1,3,3,F,10,1,A
+data/model3_session3_sample2,3,3,F,10,2,A
+data/model3_session3_sample3,3,3,F,10,3,A
+data/model3_session4_sample1,3,3,F,10,4,A
+data/model4_session3_sample1,4,4,M,30,5,B
+data/model4_session3_sample2,4,4,M,30,6,B
+data/model4_session3_sample1,4,4,M,30,7,B
+data/model4_session3_sample2,4,4,M,30,8,B
+data/model4_session3_sample3,4,4,M,30,9,B
+data/model4_session4_sample1,4,4,M,30,10,B

tests/test_database_implementations.py

@@ -8,10 +8,15 @@
 Very simple tests for Implementations
 """
 
+from pathlib import Path
+
 import bob.bio.base
 
 from bob.bio.base.config.dummy.database import database as dummy_database
-from bob.pipelines import DelayedSample
+from bob.bio.base.database import CSVDatabase
+from bob.pipelines import DelayedSample, SampleSet
+
+DATA_DIR = Path(__file__).parent / "data"
 
 
 def test_all_samples():
@@ -27,7 +32,41 @@ def test_atnt():
     database = bob.bio.base.load_resource(
         "atnt", "database", preferred_package="bob.bio.base"
     )
-    assert len(database.background_model_samples()) > 0
-    assert len(database.references()) > 0
-    assert len(database.probes()) > 0
-    assert len(database.all_samples()) > 0
+    train_set = database.background_model_samples()
+    assert len(train_set) > 0
+    assert isinstance(train_set[0], DelayedSample)
+    references = database.references()
+    assert len(references) > 0
+    references_sset = references[0]
+    assert isinstance(references_sset, SampleSet)
+    assert hasattr(references_sset, "key")
+    assert hasattr(references_sset, "subject_id")
+    assert hasattr(references_sset, "template_id")
+    references_sample = references_sset.samples[0]
+    assert isinstance(references_sample, DelayedSample)
+    assert hasattr(references_sample, "key")
+    probes = database.probes()
+    assert len(probes) > 0
+    assert isinstance(probes[0], SampleSet)
+    assert isinstance(probes[0].samples[0], DelayedSample)
+    all_samples = database.all_samples()
+    assert len(all_samples) > 0
+    assert isinstance(all_samples[0], DelayedSample)
+
+
+def test_metadata():
+    local_protocol_definition_path = DATA_DIR / "example_csv_filelist"
+    database = CSVDatabase(
+        name="dummy_metadata",
+        protocol="protocol_only_dev_metadata",
+        dataset_protocols_path=local_protocol_definition_path,
+        templates_metadata=["subject_metadata"],
+    )
+    references_sset = database.references()[0]
+    assert hasattr(references_sset, "subject_metadata")
+    references_sample = references_sset.samples[0]
+    assert hasattr(references_sample, "sample_metadata")
+    probes_sset = database.probes()[0]
+    assert hasattr(probes_sset, "subject_metadata")
+    probes_sample = probes_sset.samples[0]
+    assert hasattr(probes_sample, "sample_metadata")