feat(medcat): CPDT-223 Add failure modes to get stats

medcat/medcat/components/linking/context_based_linker.py

@@ -169,6 +169,7 @@ def _process_entity_inference(
         # Check does it have a detected concepts
         cuis = entity.link_candidates
         if not cuis:
+            logger.debug("No candidates")
             return
         # Check does it have a detected name
         name = entity.detected_name

medcat/medcat/stats/common.py

@@ -0,0 +1,37 @@
+from typing import TypedDict
+
+from medcat.tokenizing.tokens import MutableEntity
+
+
+NER_DETECTED_ENTITY_TAG = "DETECTED_ENTITY"
+
+
+class GoldAnnotation(TypedDict):
+    """Validated gold annotation payload after CUI filtering."""
+
+    start: int
+    end: int
+    cuis: list[str]
+    cui: str
+    text: str
+    raw: object
+    document_id: int
+    document_name: str
+    # context is 60 chars before + after the entity
+    context: str
+
+
+class PredictedAnnotation(TypedDict):
+    """Predicted entity payload used for scoring and metrics."""
+
+    start: int
+    end: int
+    cui: str
+    text: str
+    confidence: float
+    raw: MutableEntity
+    no_tokens: int
+    document_id: int
+    document_name: str
+    # context is 60 chars before + after the entity
+    context: str