pydantic thing

src/fairscape_cli/commands/build_commands.py

@@ -29,6 +29,7 @@
 )
 
 from fairscape_models.rocrate import ROCrateV1_2, ROCrateMetadataElem
+from fairscape_cli.utils.serialization import prune_none
 from fairscape_models.conversion.converter import ROCToTargetConverter
 from fairscape_models.conversion.mapping.croissant import MAPPING_CONFIGURATION as CROISSANT_MAPPING
 
@@ -535,7 +536,7 @@ def generate_evidence_graph(
 
             # Write the updated metadata back to the file
             with open(metadata_file, 'w') as f:
-                json.dump(metadata, f, indent=2)
+                json.dump(prune_none(metadata), f, indent=2)
 
             click.echo(f"Added hasEvidenceGraph reference to {ark_id} in RO-Crate metadata")
         except Exception as e:
@@ -774,4 +775,4 @@ def validate_merkle_command(ctx, rocrate_path: pathlib.Path, release: bool):
             elif s != c:
                 click.echo(f"    CHANGED: {url}")
 
-        ctx.exit(1)
+        ctx.exit(1)

src/fairscape_cli/data_fetcher/GenomicData.py

@@ -103,6 +103,7 @@ class Outputs(BaseModel):
 from fairscape_cli.data_fetcher.bioproject_fetcher import fetch_bioproject_data
 
 from fairscape_cli.models.rocrate import GenerateROCrate, AppendCrate
+from fairscape_cli.utils.serialization import prune_none
 from fairscape_cli.models.dataset import GenerateDataset
 from fairscape_cli.models.experiment import GenerateExperiment
 from fairscape_cli.models.instrument import GenerateInstrument
@@ -342,7 +343,7 @@ def to_rocrate(
                              root_dataset_node["hasPart"].append({"@id": entity_id})
 
                 f.seek(0)
-                json.dump(crate_json, f, indent=2)
+                json.dump(prune_none(crate_json), f, indent=2)
                 f.truncate()
 
 
@@ -360,7 +361,7 @@ def to_rocrate(
 
             if updated:
                 f.seek(0)
-                json.dump(crate_json_final, f, indent=2)
+                json.dump(prune_none(crate_json_final), f, indent=2)
                 f.truncate()
 
 
@@ -500,4 +501,3 @@ def from_json(cls, data: dict) -> 'GenomicData':
             experiments=Experiments(items=internal_experiments),
             outputs=Outputs(items=outputs_list)
         )
-

src/fairscape_cli/datasheet_builder/rocrate/datasheet_generator.py

@@ -159,6 +159,8 @@ def convert_main_sections(self) -> FairscapeDatasheet:
         distribution = distribution_converter.convert()
 
         subcrate_items = self._process_all_subcrates()
+        if not subcrate_items:
+            subcrate_items = self._build_single_crate_composition()
         composition = CompositionSection(items=subcrate_items) if subcrate_items else None
 
         return FairscapeDatasheet(
@@ -226,6 +228,42 @@ def _process_all_subcrates(self) -> List[SubCrateItem]:
 
         return subcrate_items
 
+    def _build_single_crate_composition(self) -> List[SubCrateItem]:
+        """When no subcrates exist, treat the main crate itself as a single subcrate."""
+        try:
+            converter = ROCToTargetConverter(
+                source_crate=self.main_crate,
+                mapping_configuration=SUBCRATE_MAPPING_CONFIGURATION,
+                global_index=self.global_metadata_index
+            )
+            subcrate_item = converter.convert()
+            subcrate_item.published = self.published
+            subcrate_item.preview_url = ""
+
+            if not subcrate_item.size and self.base_dir.exists():
+                try:
+                    dir_size = get_directory_size(str(self.base_dir))
+                    subcrate_item.size = format_size(dir_size)
+                except Exception:
+                    subcrate_item.size = "Unknown"
+
+            main_root = self.main_crate.metadataGraph[1].model_dump()
+            if not subcrate_item.doi:
+                subcrate_item.doi = main_root.get('identifier')
+            if not subcrate_item.related_publications:
+                pubs = main_root.get('associatedPublication', [])
+                if pubs:
+                    subcrate_item.related_publications = pubs if isinstance(pubs, list) else [pubs]
+
+            self._enhance_subcrate_item(subcrate_item, self.main_crate)
+
+            return [subcrate_item]
+        except Exception as e:
+            print(f"Error building single crate composition: {e}")
+            import traceback
+            traceback.print_exc()
+            return []
+
     def _enhance_subcrate_item(self, subcrate_item: SubCrateItem, subcrate: ROCrateV1_2):
         """Add statistical summary info to subcrate item if present."""
         root_dict = subcrate.metadataGraph[1].model_dump() if len(subcrate.metadataGraph) > 1 else {}

src/fairscape_cli/datasheet_builder/rocrate/summary_generator.py

@@ -11,6 +11,7 @@
 from fairscape_models.rocrate import ROCrateV1_2
 from fairscape_models.conversion.mapping.AIReady import score_rocrate
 from fairscape_models.conversion.models.AIReady import AIReadyScore
+from fairscape_cli.utils.serialization import model_dump_pruned
 
 
 @dataclass
@@ -89,7 +90,7 @@ def extract_summary_data(self, crate: ROCrateV1_2) -> SummaryData:
             formats = []
         formats = [f for f in formats if f and f != "unknown"]
 
-        return SummaryData(
+        summary = SummaryData(
             name=root_data.get("name", "Unnamed Dataset"),
             description=root_data.get("description", ""),
             total_size_formatted=size_str,
@@ -100,6 +101,36 @@ def extract_summary_data(self, crate: ROCrateV1_2) -> SummaryData:
             formats=formats
         )
 
+        # Fallback: compute from graph if evi:* fields are absent (single crate case)
+        if summary.total_entities == 0:
+            formats_set = set()
+            for item in crate.metadataGraph:
+                if item.guid == "ro-crate-metadata.json":
+                    continue
+                item_dict = item.model_dump(by_alias=True)
+                item_type = item_dict.get("@type", "")
+                type_str = " ".join(item_type) if isinstance(item_type, list) else str(item_type)
+
+                if "ROCrate" in type_str or "CreativeWork" in type_str:
+                    continue
+
+                summary.total_entities += 1
+
+                if "Dataset" in type_str:
+                    summary.dataset_count += 1
+                    fmt = item_dict.get("fileFormat")
+                    if fmt and fmt != "unknown":
+                        formats_set.add(fmt)
+                elif "Software" in type_str or "SoftwareSourceCode" in type_str:
+                    summary.software_count += 1
+                elif "Computation" in type_str:
+                    summary.computation_count += 1
+
+            if not summary.formats and formats_set:
+                summary.formats = sorted(formats_set)
+
+        return summary
+
     @staticmethod
     def _format_size(size_bytes: int) -> str:
         """Format bytes to human-readable size."""
@@ -160,7 +191,7 @@ def compute_aiready_score(self, crate: ROCrateV1_2) -> Tuple[AIReadyScoreData, A
 
     def save_aiready_score(self, raw_score: AIReadyScore, output_path: Path) -> None:
         """Save the AI-Ready score to a JSON file."""
-        score_dict = raw_score.model_dump()
+        score_dict = model_dump_pruned(raw_score)
         with open(output_path, 'w') as f:
             json.dump(score_dict, f, indent=2)
 

src/fairscape_cli/datasheet_builder/templates/sections/subcrates.html

@@ -297,9 +297,11 @@ <h4>Content Summary</h4>
         </div>
       </div>
     </div>
+    {% if subcrate.preview_url %}
     <div class="view-full-link">
       <a href="{{ subcrate.preview_url }}">View Full Dataset Details</a>
     </div>
+    {% endif %}
   </div>
   {% endfor %} {% else %}
   <p>No subcrates found.</p>

src/fairscape_cli/entailments/find_outputs.py

@@ -1,6 +1,7 @@
 import pathlib
 import json
 from typing import List, Dict, Tuple, Set, Any
+from fairscape_cli.utils.serialization import prune_none
 
 def extract_datasets_from_graph(graph: List[Dict]) -> List[Tuple[str, bool]]:
     """
@@ -169,7 +170,7 @@ def add_inputs_outputs_to_rocrate(rocrate_path: pathlib.Path) -> Tuple[bool, str
         metadata["@graph"] = graph
 
         with open(metadata_path, 'w') as f:
-            json.dump(metadata, f, indent=2)
+            json.dump(prune_none(metadata), f, indent=2)
 
         input_count = len(inputs)
         output_count = len(outputs)
@@ -179,4 +180,4 @@ def add_inputs_outputs_to_rocrate(rocrate_path: pathlib.Path) -> Tuple[bool, str
     except json.JSONDecodeError as e:
         return False, f"Error parsing JSON: {e}"
     except Exception as e:
-        return False, f"Unexpected error: {e}"
+        return False, f"Unexpected error: {e}"

src/fairscape_cli/entailments/inverse.py

@@ -1,5 +1,6 @@
 import pathlib
 import json
+from fairscape_cli.utils.serialization import prune_none
 from typing import List, Tuple, Dict, Any
 from rdflib import Graph, URIRef
 from rdflib.namespace import OWL
@@ -192,12 +193,12 @@ def augment_rocrate_with_inverses(
     if modified_count > 0:
         try:
             with open(metadata_file_path, 'w') as f:
-                json.dump(json_data, f, indent=2, ensure_ascii=False)
+                json.dump(prune_none(json_data), f, indent=2, ensure_ascii=False)
             print(f"RO-Crate '{metadata_file_path}' augmented with inverse properties. {modified_count} modifications made.")
         except Exception as e:
             print(f"Error saving augmented RO-Crate JSON to {metadata_file_path}: {e}")
             return False
     else:
         print(f"No inverse properties needed to be added or RO-Crate '{metadata_file_path}' is already consistent.")
 
-    return True
+    return True

src/fairscape_cli/models/bagit.py

@@ -10,6 +10,7 @@
 from typing import (
     Optional
 )
+from fairscape_cli.utils.serialization import model_dump_pruned
 
 
 class BagIt(BaseModel):
@@ -94,7 +95,7 @@ def create_bagit_metadata(self):
         bagit_info_path = self.bagit_path / 'bag-info.txt'
 
         with bagit_info_path.open(mode="w") as bag_info_file:                        
-            for key, value in self.model_dump(by_alias=True).items():                              
+            for key, value in model_dump_pruned(self, by_alias=True).items():
                 if key != 'bagit_path' and key != 'rocrate_path':                    
                     bag_info_file.write('%s: %s\n' % (key, value))
 

src/fairscape_cli/models/rocrate.py

@@ -22,6 +22,7 @@
 from fairscape_cli.config import NAAN, DEFAULT_CONTEXT
 from fairscape_cli.models.guid_utils import GenerateDatetimeSquid, clean_guid
 from fairscape_models.rocrate import ROCrateV1_2, ROCrateMetadataElem, ROCrateMetadataFileElem
+from fairscape_cli.utils.serialization import prune_none, model_dump_pruned
 
 def GenerateROCrate(
    path: pathlib.Path,
@@ -80,7 +81,7 @@ def GenerateROCrate(
        ]}
    )
 
-   rocrate_dict = rocrate.model_dump(by_alias=True, exclude_none=True)
+   rocrate_dict = model_dump_pruned(rocrate, by_alias=True)
 
    if 'ro-crate-metadata.json' in str(path):
        roCrateMetadataPath = path
@@ -94,7 +95,7 @@ def GenerateROCrate(
    with roCrateMetadataPath.open(mode="w") as metadataFile:
        json.dump(rocrate_dict, metadataFile, indent=2)
 
-   return root_dataset.model_dump(by_alias=True, exclude_none=True)
+   return model_dump_pruned(root_dataset, by_alias=True)
 class ROCrate(ROCrateMetadataElem):
     model_config = ConfigDict(populate_by_name=True)
 
@@ -213,7 +214,7 @@ def create_subcrate(
 
             f.seek(0)
             f.truncate()
-            json.dump(rocrate.model_dump(by_alias=True), f, indent=2)
+            json.dump(model_dump_pruned(rocrate, by_alias=True), f, indent=2)
 
         return subcrate['@id']
 
@@ -269,7 +270,7 @@ def initCrate(self):
 
         # Write to file
         with ro_crate_metadata_path.open(mode="w") as metadata_file:
-            json.dump(rocrate_metadata, metadata_file, indent=2)
+            json.dump(prune_none(rocrate_metadata), metadata_file, indent=2)
 
     def registerObject(self, model: Union[Dataset, Software, Computation]):
         """Add metadata to the graph of an ROCrate"""
@@ -279,7 +280,7 @@ def registerObject(self, model: Union[Dataset, Software, Computation]):
             rocrate_metadata = json.load(rocrate_metadata_file)
 
             # Add to the @graph
-            model_data = model.model_dump(by_alias=True, exclude_none=True)
+            model_data = model_dump_pruned(model, by_alias=True)
             rocrate_metadata['@graph'].append(model_data)
 
             # Add reference to root dataset's hasPart
@@ -294,7 +295,7 @@ def registerObject(self, model: Union[Dataset, Software, Computation]):
             # Write back to file
             rocrate_metadata_file.seek(0)
             rocrate_metadata_file.truncate()
-            json.dump(rocrate_metadata, rocrate_metadata_file, indent=2)
+            json.dump(prune_none(rocrate_metadata), rocrate_metadata_file, indent=2)
 
     def registerDataset(self, dataset: Dataset):
         self.registerObject(dataset)
@@ -336,7 +337,7 @@ def AppendCrate(
             root_dataset['hasPart'] = []
 
         for element in elements:
-            element_data = element.model_dump(by_alias=True, exclude_none=True)
+            element_data = model_dump_pruned(element, by_alias=True)
             rocrate_metadata['@graph'].append(element_data)
             root_dataset['hasPart'].append({"@id": element_data["@id"]})
 
@@ -346,7 +347,7 @@ def AppendCrate(
         # Write back to file
         rocrate_metadata_file.seek(0)
         rocrate_metadata_file.truncate()
-        json.dump(rocrate_metadata, rocrate_metadata_file, indent=2)
+        json.dump(prune_none(rocrate_metadata), rocrate_metadata_file, indent=2)
 
 
 def CopyToROCrate(source_filepath: str, destination_filepath: str):
@@ -385,7 +386,7 @@ def UpdateCrate(
         rocrate_metadata = json.load(rocrate_metadata_file)
 
         # Find and replace the element with matching @id
-        element_data = element.model_dump(by_alias=True, exclude_none=True)
+        element_data = model_dump_pruned(element, by_alias=True)
         for i, existing in enumerate(rocrate_metadata['@graph']):
             if existing.get('@id') == element_data['@id']:
                 rocrate_metadata['@graph'][i] = element_data
@@ -397,7 +398,7 @@ def UpdateCrate(
         # Write back to file
         rocrate_metadata_file.seek(0)
         rocrate_metadata_file.truncate()
-        json.dump(rocrate_metadata, rocrate_metadata_file, indent=2)
+        json.dump(prune_none(rocrate_metadata), rocrate_metadata_file, indent=2)
 
 def LinkSubcrates(parent_crate_path: pathlib.Path) -> List[str]:
     parent_metadata_file = parent_crate_path / 'ro-crate-metadata.json'
@@ -508,7 +509,7 @@ def find_and_process_subcrates(directory: pathlib.Path, base_path: pathlib.Path)
                     if modified:
                         subcrate_metadata['@graph'][subcrate_root_index] = subcrate_root
                         with subcrate_metadata_file.open('w') as f:
-                            json.dump(subcrate_metadata, f, indent=2)
+                            json.dump(prune_none(subcrate_metadata), f, indent=2)
 
                     reference_dict = dict(subcrate_root)
                     relative_path = (subcrate_metadata_file.relative_to(base_path)).as_posix()
@@ -535,7 +536,7 @@ def find_and_process_subcrates(directory: pathlib.Path, base_path: pathlib.Path)
                 parent_root_dataset['hasPart'].append({'@id': sub_id})
 
         with parent_metadata_file.open('w') as f:
-            json.dump(parent_metadata, f, indent=2)
+            json.dump(prune_none(parent_metadata), f, indent=2)
     else:
         print("No valid sub-crates found to link.")
 
@@ -881,11 +882,11 @@ def UpdateEntitiesInGraph(
             return False, f"RO-Crate became invalid after update operations. Details: {e}"
 
         with metadata_filepath.open(mode="w") as metadataFile:
-            json.dump(validated_crate.model_dump(by_alias=True), metadataFile, indent=2, ensure_ascii=False)
+            json.dump(model_dump_pruned(validated_crate, by_alias=True), metadataFile, indent=2, ensure_ascii=False)
 
         return True, f"Successfully processed entities. Matched: {matched_count}, Modified: {modified_count}."
 
     except Exception as e:
         import traceback
         print(f"DEBUG: Unexpected error in UpdateEntitiesInGraph: {traceback.format_exc()}")
-        return False, f"An unexpected error occurred: {type(e).__name__} - {e}"
+        return False, f"An unexpected error occurred: {type(e).__name__} - {e}"

src/fairscape_cli/models/schema/tabular.py

@@ -33,6 +33,7 @@
     DEFAULT_SCHEMA_TYPE,
     NAAN,
 )
+from fairscape_cli.utils.serialization import model_dump_pruned
 
 class FileType(str, Enum):
     CSV = "csv"
@@ -514,7 +515,7 @@ def from_dict(cls, data: dict) -> 'HDF5ValidationSchema':
 
 def write_schema(schema: TabularValidationSchema, output_file: str):
     """Write a schema to a file"""
-    schema_dict = schema.to_dict()
+    schema_dict = model_dump_pruned(schema, by_alias=True)
 
     with open(output_file, 'w') as f:
         json.dump(schema_dict, f, indent=2)
@@ -579,4 +580,3 @@ def ReadSchemaLocal(schemaFile: str) -> TabularValidationSchema:
     # load the model into 
     tabularSchema = TabularValidationSchema.model_validate(schemaJson)
     return tabularSchema
-