Merge branch 'improve_management_proposal_files' into 'master'

improve handling of bliss proposal file.

See merge request !85

nxtomomill/converter/hdf5/hdf5converter.py

@@ -365,9 +365,18 @@ class _H5ToNxConverter(BaseConverter):
             for group_name in groups:
                 _logger.debug("parse {}".format(group_name))
                 entry = h5d[group_name]
+                # improve handling of External (this is the case of proposal files)
+                if isinstance(h5d.get(group_name, getlink=True), h5py.ExternalLink):
+                    external_link = h5d.get(group_name, getlink=True)
+                    file_path = external_link.filename
+                    data_path = external_link.path
+                else:
+                    file_path = self.configuration.input_file
+                    data_path = entry.name
+
                 url = DataUrl(
-                    file_path=self.configuration.input_file,
-                    data_path=entry.name,
+                    file_path=file_path,
+                    data_path=data_path,
                     scheme="silx",
                     data_slice=None,
                 )
@@ -518,7 +527,6 @@ class _H5ToNxConverter(BaseConverter):
         return False
 
     def write(self):
-        # TODO: should be split into different parts as well
         res = []
         possible_extensions = (".hdf5", ".h5", ".nx", ".nexus")
         output_file_basename = os.path.basename(self.configuration.output_file)
@@ -535,7 +543,7 @@ class _H5ToNxConverter(BaseConverter):
         _logger.debug("\n   - ".join(acq_str))
         if len(self.acquisitions) == 0:
             _logger.warning(
-                "Not valid acquisitions have been found. Maybe no "
+                "No valid acquisitions have been found. Maybe no "
                 "init (zserie) titles have been found. You can "
                 "provide more."
             )
@@ -796,7 +804,11 @@ def from_h5_to_nx(
 
 
 def get_bliss_tomo_entries(input_file_path: str, configuration: TomoHDF5Config):
-    """Util function. Used by tomwer for example"""
+    """Util function.
+    Return the set of entries at root that match bliss entries.
+    Used by tomwer for example.
+    Warning: entries can be external links (in the case of the file beeing a proposal file)
+    """
     if not isinstance(configuration, TomoHDF5Config):
         raise TypeError("configuration is expected to be a HDF5Config")
 
@@ -806,15 +818,22 @@ def get_bliss_tomo_entries(input_file_path: str, configuration: TomoHDF5Config):
         for group_name in h5d.keys():
             _logger.debug("parse %s" % group_name)
             entry = h5d[group_name]
-            data_path = entry.name
-            if not data_path.startswith("/"):
-                data_path = "/" + data_path
-            url = DataUrl(file_path=input_file_path, data_path=data_path)
+            # improve handling of External (this is the case of proposal files)
+            if isinstance(h5d.get(group_name, getlink=True), h5py.ExternalLink):
+                external_link = h5d.get(group_name, getlink=True)
+                file_path = external_link.filename
+                data_path = external_link.path
+            else:
+                file_path = input_file_path
+                data_path = entry.name
+                if not data_path.startswith("/"):
+                    data_path = "/" + data_path
+            url = DataUrl(file_path=file_path, data_path=data_path)
             if configuration.is_using_titles:
                 # if use title take the ones corresponding to init
                 entry_type = get_entry_type(url=url, configuration=configuration)
                 if entry_type is AcquisitionStep.INITIALIZATION:
-                    acquisitions.append(entry.name)
+                    acquisitions.append(group_name)
             else:
                 # check if the entry fit one of the data_frame_grps
                 # with an init status