made process step and ld_container._to_expanded_json more robust

Author: Michael Fritzsche

src/hermes/commands/process/base.py

@@ -11,6 +11,7 @@
 from hermes.commands.base import HermesCommand, HermesPlugin
 from hermes.model.api import SoftwareMetadata
 from hermes.model.context_manager import HermesContext
+from hermes.model.error import HermesContextError
 from hermes.model.merge.container import ld_merge_dict
 
 
@@ -42,7 +43,13 @@ def __call__(self, args: argparse.Namespace) -> None:
         ctx.prepare_step('harvest')
         for harvester in harvester_names:
             self.log.info("## Process data from %s", harvester)
-            merged_doc.update(SoftwareMetadata.load_from_cache(ctx, harvester))
+            try:
+                metadata = SoftwareMetadata.load_from_cache(ctx, harvester)
+            except HermesContextError as e:
+                self.log.error("Error while trying to load data from harvest plugin '%s': %s", harvester, e)
+                self.errors.append(e)
+                continue
+            merged_doc.update(metadata)
         ctx.finalize_step("harvest")
 
         ctx.prepare_step("process")

src/hermes/model/types/ld_container.py

@@ -237,7 +237,7 @@ def _to_expanded_json(
         # while searching build a path such that it leads from the found ld_dicts ld_value to selfs data_dict/ item_list
         parent = self
         path = []
-        while parent.__class__.__name__ not in ("ld_dict", "SoftwareMetadata", "ld_merge_dict"):
+        while not "ld_dict" in [sub_cls.__name__ for sub_cls in type(parent).mro()]:
             if parent.container_type == "@list":
                 path.extend(["@list", 0])
             elif parent.container_type == "@graph":
@@ -250,7 +250,7 @@ def _to_expanded_json(
         # if neither self nor any of its parents is a ld_dict:
         # create a dict with the key of the outer most parent of self and this parents ld_value as a value
         # this dict is stored in an ld_container and simulates the most minimal JSON-LD object possible
-        if parent.__class__.__name__ not in ("ld_dict", "SoftwareMetadata", "ld_merge_dict"):
+        if not "ld_dict" in [sub_cls.__name__ for sub_cls in type(parent).mro()]:
             key = self.ld_proc.expand_iri(parent.active_ctx, parent.key)
             parent = ld_container([{key: parent._data}])
         path.append(0)
@@ -277,7 +277,7 @@ def _to_expanded_json(
                     [(new_key, temp) for new_key in temp.keys() if isinstance(temp[new_key], special_types)]
                 )
             elif isinstance(temp, ld_container):
-                if temp.__class__.__name__ in ("ld_list", "ld_merge_list") and temp.container_type == "@set":
+                if "ld_list" in [sub_cls.__name__ for sub_cls in type(temp).mro()] and temp.container_type == "@set":
                     ref[key] = temp._data
                 else:
                     ref[key] = temp._data[0]