Skip to content

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

docs: fix two docstrings that overclaim their code's behavior (#5919)

LearningCircuitf10bda3a8f80
diff --git a/src/local_deep_research/advanced_search_system/strategies/langgraph_agent_strategy.py b/src/local_deep_research/advanced_search_system/strategies/langgraph_agent_strategy.py
index 52f60bd967..3810cba126 100644
--- a/src/local_deep_research/advanced_search_system/strategies/langgraph_agent_strategy.py
+++ b/src/local_deep_research/advanced_search_system/strategies/langgraph_agent_strategy.py
@@ -170,8 +170,11 @@ def _snippet_dedup_key(result: dict) -> str:
the same passage to any reader — hash alike instead of becoming two
citation indices;
* case is folded;
- * a trailing ellipsis or sentence punctuation is dropped, so
- ``"…text…"`` and ``"…text"`` agree.
+ * trailing spaces, ``.``, and ``…`` are stripped, so ``"…text…"`` and
+ ``"…text"`` agree. Trailing ``!`` and ``?`` are deliberately left
+ alone: stripping them would collapse two passages that differ only
+ in final punctuation, and this key exists to stop discarding
+ evidence, not to merge more of it.
Returns ``""`` — a sentinel, not a hash — for anything with no
content: a missing/empty snippet, whitespace or punctuation only, or
diff --git a/src/local_deep_research/web/services/report_assembly_service.py b/src/local_deep_research/web/services/report_assembly_service.py
index 9ac422cbac..343e893c0d 100644
--- a/src/local_deep_research/web/services/report_assembly_service.py
+++ b/src/local_deep_research/web/services/report_assembly_service.py
@@ -216,8 +216,8 @@ def get_research_source_links(
same URL legitimately appears several times with different snippets
(see #5894), and without this a "top 3 sources" card could render one
URL three times. Because dedup happens in Python the row cap cannot
- be pushed into SQL, so the query is walked lazily and abandoned as
- soon as ``limit`` distinct sources are in hand.
+ be pushed into SQL: the query fetches every matching row, but
+ formatting stops as soon as ``limit`` distinct sources are in hand.
Args:
research_id: The ResearchHistory id.