[db] implementing MEDS database in the new interface

bob/bio/face/database/__init__.py

@@ -12,13 +12,13 @@ from .multipie import MultipieBioDatabase
 from .ijbc import IJBCBioDatabase
 from .replaymobile import ReplayMobileBioDatabase
 from .fargo import FargoBioDatabase
-
+from .meds import MEDSDatabase
 
 # gets sphinx autodoc done right - don't remove it
 
 
 def __appropriate__(*args):
-  """Says object was actually declared here, and not in the import module.
+    """Says object was actually declared here, and not in the import module.
   Fixing sphinx warnings of not being able to find classes, when path is shortened.
   Parameters:
 
@@ -28,20 +28,21 @@ def __appropriate__(*args):
   <https://github.com/sphinx-doc/sphinx/issues/3048>`
   """
 
-  for obj in args:
-    obj.__module__ = __name__
+    for obj in args:
+        obj.__module__ = __name__
+
 
 __appropriate__(
-  FaceBioFile,
-  MobioBioDatabase,
-  ReplayBioDatabase,
-  AtntBioDatabase,
-  GBUBioDatabase,
-  ARFaceBioDatabase,
-  LFWBioDatabase,
-  MultipieBioDatabase,
-  IJBCBioDatabase,
-  ReplayMobileBioDatabase,
-  FargoBioDatabase
+    FaceBioFile,
+    MobioBioDatabase,
+    ReplayBioDatabase,
+    AtntBioDatabase,
+    GBUBioDatabase,
+    ARFaceBioDatabase,
+    LFWBioDatabase,
+    MultipieBioDatabase,
+    IJBCBioDatabase,
+    ReplayMobileBioDatabase,
+    FargoBioDatabase,
 )
-__all__ = [_ for _ in dir() if not _.startswith('_')]
+__all__ = [_ for _ in dir() if not _.startswith("_")]

bob/bio/face/database/meds.py

+#!/usr/bin/env python
+# vim: set fileencoding=utf-8 :
+# Tiago de Freitas Pereira <tiago.pereira@idiap.ch>
+
+"""
+  MEDS database implementation 
+"""
+
+from bob.bio.base.database import CSVDatasetDevEval, CSVToSampleLoader
+from bob.extension import rc
+import bob.io.base
+from bob.bio.face.database.sample_loaders import EyesAnnotationsLoader
+
+
+# TODO: POINT TO THE `.bob/meds``
+dataset_protocol_path = "/idiap/user/tpereira/gitlab/bob/bob.nightlies/meds"
+
+
+class MEDSDatabase(CSVDatasetDevEval):
+    def __init__(
+        self,
+        protocol,
+        dataset_protocol_path=dataset_protocol_path,
+        csv_to_sample_loader=CSVToSampleLoader(
+            data_loader=bob.io.base.load,
+            metadata_loader=EyesAnnotationsLoader(),
+            dataset_original_directory=rc["bob.db.meds.directory"],
+            extension=".jpg",
+        ),
+    ):
+
+        # TODO: IMPLEMENT THE DOWNLOAD MECHANISM
+
+        super().__init__(dataset_protocol_path, protocol, csv_to_sample_loader)
+

bob/bio/face/database/sample_loaders.py

 #!/usr/bin/env python
 # vim: set fileencoding=utf-8 :
 
-"""  Sample loader"""
+"""  Sample and Metatada loader"""
 
 
-from bob.bio.base.database import CSVToSampleLoader
-from bob.pipelines import Sample, DelayedSample, SampleSet
-import functools
-import os
-
-
-class CSVToSampleLoaderEyesAnnotations(CSVToSampleLoader):
+class EyesAnnotationsLoader:
     """
-    Convert CSV files in the format below to either a list of
-    :any:`bob.pipelines.DelayedSample` or :any:`bob.pipelines.SampleSet`
-
     Convert  leye_x, leye_y, reye_x, reye_y attributes to `annotations = (leye, reye)`
-
     """
 
-    def convert_row_to_sample(self, row, header):
-        path = row[0]
-        subject = row[1]
-        kwargs = dict([[h, r] for h, r in zip(header[2:], row[2:])])
-
-        annotations = {
-            "leye": (kwargs["leye_x"], kwargs["leye_y"]),
-            "reye": (kwargs["reye_x"], kwargs["reye_y"]),
+    def __call__(self, row, header=None):
+        def find_attribute(attribute):
+            for i, a in enumerate(header):
+                if a == attribute:
+                    return i
+            else:
+                ValueError(f"Attribute not found in the dataset: {a}")
+
+        eyes = {
+            "leye": (row[find_attribute("leye_x")], row[find_attribute("leye_y")]),
+            "reye": (row[find_attribute("reye_x")], row[find_attribute("reye_y")]),
         }
 
-        kwargs.pop("leye_x")
-        kwargs.pop("leye_y")
-        kwargs.pop("reye_x")
-        kwargs.pop("reye_y")
-
-        return DelayedSample(
-            functools.partial(
-                self.data_loader,
-                os.path.join(self.dataset_original_directory, path + self.extension),
-            ),
-            key=path,
-            subject=subject,
-            annotations=annotations,
-            **kwargs,
-        )
-
-
-"""
-class CSVToSampleLoaderEyesAnnotations(CSVToSampleLoader):
-    def __call__(self, filename):
-        import ipdb
-
-        ipdb.set_trace()
-        samples = super(CSVToSampleLoaderEyesAnnotations, self).__call__(filename)
-
-        def generate_annotations(sample):
-            
-            Convert  leye_x, leye_y, reye_x, reye_y attributes to
-            `annotations = (leye, reye)`
-
-            
-
-            check_keys = [
-                a in (sample.__dict__.keys())
-                for a in ["leye_x", "leye_y", "reye_x", "reye_y"]
-            ]
-
-            if not check_keys:
-                raise ValueError(
-                    "Sample needs to contain the following annotations: 'leye_x', 'leye_y', 'reye_x', 'reye_y'"
-                )
-
-            annotations = {
-                "leye": (sample.leye_x, sample.leye_y),
-                "reye": (sample.reye_x, sample.reye_y),
-            }
-
-            # Changing the state of samples for efficiency
-            # We might have a gigantic amount of datasets
-            sample.__dict__.pop("leye_x")
-            sample.__dict__.pop("leye_y")
-            sample.__dict__.pop("reye_x")
-            sample.__dict__.pop("reye_y")
-            sample.annotations = annotations
-
-        for sample in samples:
-            generate_annotations(sample)
+        annotation = {"annotations": eyes}
 
-        return samples
-"""
+        return annotation