fix: escape apostrophes for parsing and rendering [JAR-9386] (#704)

Author: JoshParkSJ

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "uipath-langchain"
-version = "0.8.25"
+version = "0.8.26"
 description = "Python SDK that enables developers to build and deploy LangGraph agents to the UiPath Cloud Platform"
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.11"

src/uipath_langchain/runtime/_citations.py

@@ -22,8 +22,8 @@
 
 logger = logging.getLogger(__name__)
 
-_TAG_RE = re.compile(r'<uip:cite\s+((?:[a-z_]+="[^"]*"\s*)+)/\s*>')
-_ATTR_RE = re.compile(r'([a-z_]+)="([^"]*)"')
+_TAG_RE = re.compile(r'<uip:cite\s+((?:[a-z_]+="(?:[^"\\]|\\.)*"\s*)+)/\s*>')
+_ATTR_RE = re.compile(r'([a-z_]+)="((?:[^"\\]|\\.)*)"')
 
 
 @dataclass(frozen=True)  # frozen to make hashable / de-dupe sources
@@ -45,7 +45,9 @@ def _parse_citations(text: str) -> list[tuple[str, _ParsedCitation | None]]:
         raw_attributes = match.group(1)
 
         # title="foo" url="https://..." -> [("title","foo"), ("url","https://...")]
-        attributes = dict(_ATTR_RE.findall(raw_attributes))
+        attributes = {
+            k: v.replace('\\"', '"') for k, v in _ATTR_RE.findall(raw_attributes)
+        }
 
         title = attributes.get("title", "")
         url = attributes.get("url")

tests/runtime/test_citations.py

@@ -556,6 +556,34 @@ def test_uip_prefix_followed_by_citation_single_chunk(self):
         assert cited[0].data == "<uip "
         assert cited[0].citation.end.sources[0].url == "https://example.com"
 
+    def test_escaped_quotes_in_title(self):
+        """Citation with backslash-escaped quotes in title is parsed correctly."""
+        proc = CitationStreamProcessor()
+        text = r'Some text.<uip:cite title="The Peculiar Journey of \"Orange\"" url="https://example.com" />'
+        events = proc.add_chunk(text)
+        events.extend(proc.finalize())
+        cited = [e for e in events if e.citation is not None]
+        assert len(cited) == 1
+        assert cited[0].data == "Some text."
+        source = cited[0].citation.end.sources[0]
+        assert isinstance(source, UiPathConversationCitationSourceUrl)
+        assert source.url == "https://example.com"
+        assert source.title == 'The Peculiar Journey of "Orange"'
+
+    def test_escaped_quotes_in_title_streamed(self):
+        """Escaped quotes in title still work when streamed across chunks."""
+        proc = CitationStreamProcessor()
+        events = proc.add_chunk(r'Text.<uip:cite title="Say \"hi')
+        # "Text." is emitted immediately; the partial tag is buffered
+        assert len(events) == 1
+        assert events[0].data == "Text."
+        assert events[0].citation is None
+        events = proc.add_chunk(r' there\"" url="https://x.com" />')
+        events.extend(proc.finalize())
+        cited = [e for e in events if e.citation is not None]
+        assert len(cited) == 1
+        assert cited[0].citation.end.sources[0].title == 'Say "hi there"'
+
 
 class TestExtractCitationsFromText:
     """Test cases for extract_citations_from_text function."""
@@ -689,6 +717,35 @@ def test_reference_without_page_number_skipped(self):
         assert cleaned == "UiPath reported earnings"
         assert citations == []
 
+    def test_escaped_quotes_in_title(self):
+        """Citation with escaped quotes in title is parsed and unescaped."""
+        text = (
+            r'A fact<uip:cite title="The \"Real\" Story" url="https://example.com" />'
+        )
+        cleaned, citations = extract_citations_from_text(text)
+        assert cleaned == "A fact"
+        assert len(citations) == 1
+        source = citations[0].sources[0]
+        assert isinstance(source, UiPathConversationCitationSourceUrl)
+        assert source.title == 'The "Real" Story'
+        assert source.url == "https://example.com"
+
+    def test_escaped_quotes_in_title_debug_dump_repro(self):
+        """Reproduce the exact tag from the debug dump that was failing."""
+        text = (
+            r'some text.<uip:cite title="The Peculiar Journey of \"Orange\"" '
+            r'url="https://www.vocabulary.com/articles/wordroutes/the-peculiar-journey-of-orange/" />'
+        )
+        cleaned, citations = extract_citations_from_text(text)
+        assert cleaned == "some text."
+        assert len(citations) == 1
+        source = citations[0].sources[0]
+        assert source.title == 'The Peculiar Journey of "Orange"'
+        assert (
+            source.url
+            == "https://www.vocabulary.com/articles/wordroutes/the-peculiar-journey-of-orange/"
+        )
+
     def test_different_sources_get_different_numbers(self):
         """Different sources get incrementing numbers."""
         text = (