Created Answer.get_unique_docs_from_contexts

paperqa/types.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+from collections.abc import Collection
 from typing import Any, Callable
 from uuid import UUID, uuid4
 
@@ -233,6 +234,15 @@ def add_tokens(self, result: LLMResult):
             self.token_counts[result.model][0] += result.prompt_count
             self.token_counts[result.model][1] += result.completion_count
 
+    def get_unique_docs_from_contexts(
+        self, score_threshold: int = 0
+    ) -> Collection[Doc]:
+        """Parse contexts for docs with scores above the input threshold."""
+        return {
+            c.text.doc
+            for c in filter(lambda x: x.score >= score_threshold, self.contexts)
+        }
+
 
 class ChunkMetadata(BaseModel):
     """Metadata for chunking algorithm."""