Made key_filters part of Answer

paperqa/docs.py

@@ -364,10 +364,12 @@ async def aget_evidence(
     ) -> Answer:
         if len(self.docs) == 0:
             return answer
+        if key_filter is not None:
+            answer.key_filter = key_filter
         if self._faiss_index is None:
             self._build_faiss_index()
         _k = k
-        if key_filter is not None:
+        if answer.key_filter is not None:
             _k = k * 10  # heuristic
         # want to work through indices but less k
         if marginal_relevance:
@@ -379,8 +381,17 @@ async def aget_evidence(
                 answer.question, k=_k, fetch_k=5 * _k
             )
         # ok now filter
-        if key_filter is not None:
-            docs = [doc for doc in docs if doc.metadata["dockey"] in key_filter][:k]
+        if answer.key_filter is not None:
+            # I realize that by testing for existence
+            # in strings that weird cases can
+            # happen - like FooBar can match FooBar2023
+            # but remember there are later checks
+            # The risk of explicitly parsing is that the
+            # language model may not give back in predictable
+            # format (e.g., - "FooBar and BarSoo are good papers")
+            docs = [doc for doc in docs if doc.metadata["dockey"] in answer.key_filter][
+                :k
+            ]
 
         async def process(doc):
             if doc.metadata["dockey"] in self._deleted_keys:
@@ -524,12 +535,12 @@ async def aquery(
                 answer.tokens += callbacks[0].total_tokens
                 answer.cost += callbacks[0].total_cost
                 key_filter = True if len(keys) > 0 else False
+                answer.key_filter = keys
             answer = await self.aget_evidence(
                 answer,
                 k=k,
                 max_sources=max_sources,
                 marginal_relevance=marginal_relevance,
-                key_filter=keys if key_filter else None,
                 get_callbacks=get_callbacks,
             )
         context_str, contexts = answer.context, answer.contexts

paperqa/types.py

@@ -1,6 +1,6 @@
 from dataclasses import dataclass
 from pathlib import Path
-from typing import Any, Dict, List, Union
+from typing import Any, Dict, List, Union, Optional
 
 StrPath = Union[str, Path]
 
@@ -33,6 +33,7 @@ class Answer:
     passages: Dict[str, str] = None
     tokens: int = 0
     cost: float = 0
+    key_filter: Optional[str] = None
 
     def __post_init__(self):
         """Initialize the answer."""

paperqa/version.py

@@ -1 +1 @@
-__version__ = "2.0.2"
+__version__ = "2.0.3"