implemented suggestions and fixed bug

Author: notactuallyfinn

src/hermes/commands/deposit/file.py

@@ -7,12 +7,17 @@
 # SPDX-FileContributor: Stephan Druskat
 
 import json
+import logging
+import os
 
 from pydantic import BaseModel
 
 from hermes.commands.deposit.base import BaseDepositPlugin
 
 
+_log = logging.getLogger("cli.deposit.file")
+
+
 class FileDepositSettings(BaseModel):
     filename: str = 'hermes.json'
 
@@ -31,3 +36,4 @@ def publish(self) -> None:
 
         with open(file_config.filename, 'w') as deposition_file:
             json.dump(self.metadata.compact(), deposition_file, indent=2)
+        _log.info(f"The deposited metadata can be found in {os.path.abspath(file_config.filename)}.")

src/hermes/commands/deposit/invenio.py

@@ -311,13 +311,23 @@ def prepare(self) -> None:
         - update ``self.metadata`` with metadata collected during the checks
         """
 
-        rec_id = self.config.record_id
-        doi = self.config.doi
-
-        codemeta_identifier = self.metadata.get("identifier", None)
-        rec_id, rec_meta = self.resolver.resolve_latest_id(
-            record_id=rec_id, doi=doi, codemeta_identifier=codemeta_identifier
-        )
+        conf_rec_id = self.config.record_id
+        conf_doi = self.config.doi
+
+        codemeta_identifiers = self.metadata.get("identifier", [None])
+        rec_id, rec_meta = None, {}
+        for codemeta_identifier in codemeta_identifiers:
+            if not isinstance(codemeta_identifier, str):
+                # FIXME: Can also be PropertyValue (i.e. ld_dict), that case has to be handled.
+                codemeta_identifier = None
+            tmp_rec_id, tmp_rec_meta = self.resolver.resolve_latest_id(
+                record_id=conf_rec_id, doi=conf_doi, codemeta_identifier=codemeta_identifier
+            )
+            if tmp_rec_id is not None or tmp_rec_meta != {}:
+                if rec_id != tmp_rec_id or rec_meta != tmp_rec_meta:
+                    # FIXME: Maybe finding different record ids is not fatal?
+                    raise HermesValidationError("Found two different record ids or conflicting metadata.")
+                rec_id, rec_meta = tmp_rec_id, tmp_rec_meta
 
         if len(self.metadata.get("version", [])) > 1:
             raise HermesValidationError("Too many licenses for invenio deposit.")

src/hermes/commands/process/base.py

@@ -13,7 +13,6 @@
 from hermes.error import HermesPluginRunError, MisconfigurationError
 from hermes.model.api import SoftwareMetadata
 from hermes.model.context_manager import HermesContext
-from hermes.model.error import HermesValidationError
 from hermes.model.merge.action import MergeAction
 from hermes.model.merge.container import ld_merge_dict
 
@@ -28,6 +27,7 @@ def __call__(self, command: HermesCommand) -> dict[Union[str, None], dict[Union[
 class ProcessSettings(BaseModel):
     """Generic deposition settings."""
 
+    sources: list = []
     plugins: list = []
 
 
@@ -72,19 +72,32 @@ def __call__(self, args: argparse.Namespace) -> None:
 
         self.log.info("## Merge the metadata of the harvesters")
         # Get all harvesters
-        harvester_names = self.root_settings.harvest.sources
+        harvester_names = self.settings.sources if self.settings.sources else self.root_settings.harvest.sources
+        merged_any = False
         for harvester in harvester_names:
             self.log.info(f"## Load data from {harvester} plugin")
             # load data from harvester
             try:
                 metadata = SoftwareMetadata.load_from_cache(ctx, harvester)
-            except Exception as e:
-                self.log.error(f"The data from the harvester {harvester} could not be loaded or is invalid.")
-                raise HermesValidationError(f"The results of the harvest plugin {harvester} is invalid.") from e
+            except Exception:
+                # skip this harvester when the data is invalid
+                self.log.warning(f"The data from the harvester {harvester} could not be loaded or is invalid.")
+                self.log.info(f"## Aborting merge for {harvester}")
+                continue
 
             self.log.info(f"## Merge data from {harvester} plugin")
             # merge data into the merge dict
             merged_doc.update(metadata)
+            merged_any = True
+
+        # error if nothing was merged
+        if not merged_any:
+            self.log.error(
+                f"""No metadata has been merged. {
+                    "No harvesters to merge from were supplied" if not harvester_names else
+                    "The merging failed for all harvesters."
+                }"""
+            )
 
         self.log.info("## Store processed metadata")
         # store processed data