Suyash091/agentic.email
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

human hand off

Browse source
This commit is contained in:
debu-2 2026-10-06 07:12:55 +05:30
parent d0fee5c8f7
commit 7af776196a
34 changed files with 2941 additions and 195 deletions
Download patch file Download diff file Expand all files Collapse all files

24
backend/app/services/reply_extractor.py
View file

@ -16,6 +16,7 @@ TEXT_QUOTE_MARKERS = [
PRIVATE_LEAK_MARKERS = [
"human handoff",
" · handoff · ",
"a conversation needs your attention",
"why this was handed off",
"why you're receiving this",
@ -37,11 +38,19 @@ def _normalize_newlines(text: str) -> str:
class _ReplyHTMLTextParser(HTMLParser):
BLOCK_TAGS = {"div", "p", "pre", "li", "ul", "ol", "table", "tr"}
IGNORED_TAGS = {"script", "style", "iframe", "object"}
QUOTE_CLASSES = {
"gmail_attr",
"gmail_quote",
"protonmail_quote",
"yahoo_quoted",
"moz-cite-prefix",
}
def __init__(self) -> None:
super().__init__(convert_charrefs=True)
self.parts: list[str] = []
self.quote_depth = 0
self.quote_tags: list[str] = []
self.ignored_depth = 0
self.at_line_start = True
@ -70,11 +79,19 @@ class _ReplyHTMLTextParser(HTMLParser):
return
if self.ignored_depth:
return
classes = {
value
for key, value in attrs
if key.lower() == "class"
for value in (value or "").split()
}
is_quote = tag == "blockquote" or bool(classes.intersection(self.QUOTE_CLASSES))
if tag == "br":
self._newline()
elif tag == "blockquote":
elif is_quote:
self._newline()
self.quote_depth += 1
self.quote_tags.append(tag)
elif tag in self.BLOCK_TAGS:
self._newline()
@ -85,9 +102,10 @@ class _ReplyHTMLTextParser(HTMLParser):
return
if self.ignored_depth:
return
if tag == "blockquote":
self._newline()
if self.quote_tags and self.quote_tags[-1] == tag:
self.quote_tags.pop()
self.quote_depth = max(0, self.quote_depth - 1)
self._newline()
elif tag in self.BLOCK_TAGS:
self._newline()