From 1aad1db9f69692c8a206f2964ec9f01ec2d21f72 Mon Sep 17 00:00:00 2001
From: Afonso Coutinho <afonso@omelhorsite.pt>
Date: Mon, 20 Jul 2026 08:39:16 +0100
Subject: [PATCH] fix: services research source extraction crashes on a
 non-dict finding (#1868)

---
 services/research/research_handler.py      |  2 ++
 tests/test_svc_research_sources_nondict.py | 15 +++++++++++++++
 2 files changed, 17 insertions(+)
 create mode 100644 tests/test_svc_research_sources_nondict.py

diff --git a/services/research/research_handler.py b/services/research/research_handler.py
index 2521f61e1..2ef74a8ef 100644
--- a/services/research/research_handler.py
+++ b/services/research/research_handler.py
@@ -186,6 +186,8 @@ class ResearchHandler:
         seen = set()
         sources = []
         for f in findings:
+            if not isinstance(f, dict):
+                continue
             url = f.get("url", "")
             title = f.get("title", "") or url
             summary = f.get("summary", "") or f.get("evidence", "")
diff --git a/tests/test_svc_research_sources_nondict.py b/tests/test_svc_research_sources_nondict.py
new file mode 100644
index 000000000..d7ea63ab2
--- /dev/null
+++ b/tests/test_svc_research_sources_nondict.py
@@ -0,0 +1,15 @@
+from services.research.research_handler import ResearchHandler
+
+
+def test_extract_sources_skips_non_dict_findings():
+    # findings come from the DeepResearcher result list / cached JSON; a
+    # malformed entry (None or a bare string) made f.get crash and drop every
+    # real source.
+    findings = [
+        {"url": "https://a.com", "title": "A", "summary": "real analysis of the topic"},
+        "junk-row",
+        None,
+        {"url": "https://b.com", "summary": "more genuine detail here"},
+    ]
+    out = ResearchHandler._extract_sources(findings)
+    assert [s["url"] for s in out] == ["https://a.com", "https://b.com"]