Update src/aibom_generator/utils.py

src/aibom_generator/utils.py

@@ -98,25 +98,36 @@ def validate_spdx(license_entry):
 
 def calculate_completeness_score(aibom: Dict[str, Any]) -> Dict[str, Any]:
     field_checklist = {}
+    max_scores = {
+        "required_fields": 20,
+        "metadata": 20,
+        "component_basic": 20,
+        "component_model_card": 30,
+        "external_references": 10
+    }
 
+    # Required Fields (20 points max)
     required_fields = ["bomFormat", "specVersion", "serialNumber", "version"]
     required_score = sum([5 if aibom.get(field) else 0 for field in required_fields])
     for field in required_fields:
         field_checklist[field] = "✔" if aibom.get(field) else "✘"
 
+    # Metadata (20 points max)
     metadata = aibom.get("metadata", {})
     metadata_fields = ["timestamp", "tools", "authors", "component"]
     metadata_score = sum([5 if metadata.get(field) else 0 for field in metadata_fields])
     for field in metadata_fields:
         field_checklist[f"metadata.{field}"] = "✔" if metadata.get(field) else "✘"
 
+    # Component Basic Info (20 points max)
     components = aibom.get("components", [])
-    component_scores = []
-    model_card_scores = []
-
-    for comp in components:
+    component_score = 0
+    
+    if components:
+        # Use the first component as specified in the design
+        comp = components[0]
         comp_fields = ["type", "name", "bom-ref", "purl", "description", "licenses"]
-        comp_score = sum([
+        component_score = sum([
             2 if comp.get("type") else 0,
             4 if comp.get("name") else 0,
             2 if comp.get("bom-ref") else 0,
@@ -124,24 +135,34 @@ def calculate_completeness_score(aibom: Dict[str, Any]) -> Dict[str, Any]:
             4 if comp.get("description") and len(comp["description"]) > 20 else 0,
             4 if comp.get("licenses") and validate_spdx(comp["licenses"]) else 0
         ])
-        component_scores.append(comp_score)
         for field in comp_fields:
             field_checklist[f"component.{field}"] = "✔" if comp.get(field) else "✘"
-
+            if field == "purl" and comp.get(field) and not re.match(r'^pkg:huggingface/.+', comp["purl"]):
+                field_checklist[f"component.{field}"] = "✘"
+            if field == "description" and comp.get(field) and len(comp["description"]) <= 20:
+                field_checklist[f"component.{field}"] = "✘"
+            if field == "licenses" and comp.get(field) and not validate_spdx(comp["licenses"]):
+                field_checklist[f"component.{field}"] = "✘"
+
+    # Model Card Section (30 points max)
+    model_card_score = 0
+    
+    if components:
+        # Use the first component's model card as specified in the design
+        comp = components[0]
         card = comp.get("modelCard", {})
         card_fields = ["modelParameters", "quantitativeAnalysis", "considerations"]
-        card_score = sum([
+        model_card_score = sum([
             10 if card.get("modelParameters") else 0,
             10 if card.get("quantitativeAnalysis") else 0,
             10 if card.get("considerations") and len(card["considerations"]) > 50 else 0
         ])
-        model_card_scores.append(card_score)
         for field in card_fields:
             field_checklist[f"modelCard.{field}"] = "✔" if field in card else "✘"
+            if field == "considerations" and field in card and len(card["considerations"]) <= 50:
+                field_checklist[f"modelCard.{field}"] = "✘"
 
-    avg_comp_score = (sum(component_scores) / len(component_scores)) if component_scores else 0
-    avg_model_card_score = (sum(model_card_scores) / len(model_card_scores)) if model_card_scores else 0
-
+    # External References (10 points max)
     ext_refs = aibom.get("externalReferences", [])
     ext_score = 0
     for ref in ext_refs:
@@ -155,11 +176,12 @@ def calculate_completeness_score(aibom: Dict[str, Any]) -> Dict[str, Any]:
     ext_score = min(ext_score, 10)
     field_checklist["externalReferences"] = "✔" if ext_refs else "✘"
 
+    # Calculate total score
     total_score = (
         (required_score * 0.20) +
         (metadata_score * 0.20) +
-        (avg_comp_score * 0.20) +
-        (avg_model_card_score * 0.30) +
+        (component_score * 0.20) +
+        (model_card_score * 0.30) +
         (ext_score * 0.10)
     )
 
@@ -168,10 +190,11 @@ def calculate_completeness_score(aibom: Dict[str, Any]) -> Dict[str, Any]:
         "section_scores": {
             "required_fields": required_score,
             "metadata": metadata_score,
-            "component_basic": avg_comp_score,
-            "component_model_card": avg_model_card_score,
+            "component_basic": component_score,
+            "component_model_card": model_card_score,
             "external_references": ext_score
         },
+        "max_scores": max_scores,
         "field_checklist": field_checklist
     }