fix: str issues in squad_to_dpr (#4826)

* fix #4754 * fix #4753 * run black formatting --------- Co-authored-by: Julian Risch <julian.risch@deepset.ai>
deepset-ai · May 7, 2023 · 2ff8b0d · 2ff8b0d
1 parent 705a2c0
commit 2ff8b0d
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 2 deletions.
diff --git a/haystack/document_stores/utils.py b/haystack/document_stores/utils.py
@@ -197,7 +197,9 @@ def _extract_docs_and_labels_from_dict(
                         )
                         labels.append(label)
                     else:
-                        ans_position = cur_full_doc.content[answer["answer_start"] : answer["answer_start"] + len(ans)]
+                        ans_position = cur_full_doc.content[
+                            answer["answer_start"] : answer["answer_start"] + len(str(ans))
+                        ]
                         if ans != ans_position:
                             # do not use answer
                             problematic_ids.append(qa.get("id", "missing"))

diff --git a/haystack/utils/squad_to_dpr.py b/haystack/utils/squad_to_dpr.py
@@ -227,7 +227,7 @@ def get_hard_negative_contexts(retriever: BaseRetriever, question: str, answers:
     for retrieved_doc in retrieved_docs:
         retrieved_doc_id = retrieved_doc.meta.get("name", "")
         retrieved_doc_text = retrieved_doc.content
-        if any(answer.lower() in retrieved_doc_text.lower() for answer in answers):
+        if any(str(answer).lower() in retrieved_doc_text.lower() for answer in answers):
             continue
         list_hard_neg_ctxs.append({"title": retrieved_doc_id, "text": retrieved_doc_text, "passage_id": ""})