handoff fixes
This commit is contained in:
parent
3f4bcc26d9
commit
6c52cd19cb
8 changed files with 561 additions and 59 deletions
198
backend/app/services/handoff_email_templates.py
Normal file
198
backend/app/services/handoff_email_templates.py
Normal file
|
|
@ -0,0 +1,198 @@
|
|||
from html import escape
|
||||
|
||||
DEBUG_REASON_MARKERS = (
|
||||
"step passes",
|
||||
"classifier",
|
||||
"rule evaluation",
|
||||
"no handoff instructions were set",
|
||||
)
|
||||
|
||||
|
||||
def _clean_external_subject(subject: str) -> str:
|
||||
value = (subject or "(No subject)").strip()
|
||||
for prefix in ("Re: ", "RE: ", "re: "):
|
||||
if value.startswith(prefix):
|
||||
value = value[len(prefix) :].strip()
|
||||
for prefix in ("Human Handoff · ", "Human Handoff: "):
|
||||
if value.lower().startswith(prefix.lower()):
|
||||
value = value[len(prefix) :].strip()
|
||||
return value or "(No subject)"
|
||||
|
||||
|
||||
def handoff_email_subject(external_subject: str) -> str:
|
||||
return f"Human Handoff · {_clean_external_subject(external_subject)}"
|
||||
|
||||
|
||||
def external_reply_subject(external_subject: str) -> str:
|
||||
cleaned = _clean_external_subject(external_subject)
|
||||
return f"Re: {cleaned}"
|
||||
|
||||
|
||||
def format_reason_for_human(reason: str | None, *, is_update: bool = False) -> str | None:
|
||||
if is_update:
|
||||
return None
|
||||
if not reason or not reason.strip():
|
||||
return None
|
||||
lower = reason.strip().lower()
|
||||
if any(marker in lower for marker in DEBUG_REASON_MARKERS):
|
||||
return "This conversation matched your Human Handoff settings."
|
||||
if lower in {
|
||||
"matched handoff rules.",
|
||||
"new message in the linked conversation.",
|
||||
}:
|
||||
return "This conversation requires human attention."
|
||||
return reason.strip()
|
||||
|
||||
|
||||
def build_initial_handoff_text(
|
||||
*,
|
||||
external_sender: str,
|
||||
external_subject: str,
|
||||
external_body: str,
|
||||
agent_email: str,
|
||||
handoff_reason: str | None,
|
||||
) -> str:
|
||||
reason = format_reason_for_human(handoff_reason)
|
||||
lines = [
|
||||
"GENTIC — HUMAN HANDOFF",
|
||||
"",
|
||||
"This conversation needs your attention.",
|
||||
"",
|
||||
f"From: {external_sender}",
|
||||
f"Subject: {_clean_external_subject(external_subject)}",
|
||||
"",
|
||||
"LATEST MESSAGE",
|
||||
"",
|
||||
external_body.strip(),
|
||||
"",
|
||||
]
|
||||
if reason:
|
||||
lines.extend(["WHY YOU'RE RECEIVING THIS", "", reason, ""])
|
||||
lines.extend(
|
||||
[
|
||||
"REPLY NORMALLY",
|
||||
"",
|
||||
f"Reply to this email and Gentic will send only your new reply to the "
|
||||
f"original sender through {agent_email}.",
|
||||
"",
|
||||
"This private handoff conversation and quoted history will not be shared "
|
||||
"with the sender.",
|
||||
"",
|
||||
"—",
|
||||
"Gentic Human Handoff",
|
||||
"Private Agent ↔ Human conversation",
|
||||
]
|
||||
)
|
||||
return "\n".join(lines)
|
||||
|
||||
|
||||
def build_initial_handoff_html(
|
||||
*,
|
||||
external_sender: str,
|
||||
external_subject: str,
|
||||
external_body: str,
|
||||
agent_email: str,
|
||||
handoff_reason: str | None,
|
||||
) -> str:
|
||||
reason = format_reason_for_human(handoff_reason)
|
||||
reason_block = ""
|
||||
if reason:
|
||||
reason_block = f"""
|
||||
<tr><td style="padding:18px 24px 0;">
|
||||
<p style="margin:0 0 8px;font-size:11px;font-weight:600;letter-spacing:.05em;text-transform:uppercase;color:#8A8695;">Why you're receiving this</p>
|
||||
<p style="margin:0;font-size:13px;line-height:1.55;color:#4A4652;">{escape(reason)}</p>
|
||||
</td></tr>"""
|
||||
|
||||
return f"""<!DOCTYPE html>
|
||||
<html><body style="margin:0;padding:0;background:#F5F4F2;">
|
||||
<table role="presentation" width="100%" cellpadding="0" cellspacing="0" style="background:#F5F4F2;padding:24px 12px;">
|
||||
<tr><td align="center">
|
||||
<table role="presentation" width="600" cellpadding="0" cellspacing="0" style="width:100%;max-width:600px;background:#FFFFFF;border:1px solid #E8E5EA;border-radius:12px;overflow:hidden;">
|
||||
<tr><td style="padding:24px 24px 8px;">
|
||||
<p style="margin:0 0 4px;font-size:12px;font-weight:600;color:#7357F6;">✦ Gentic</p>
|
||||
<h1 style="margin:0 0 8px;font-size:20px;font-weight:700;color:#171424;">Human Handoff</h1>
|
||||
<p style="margin:0;font-size:13px;line-height:1.5;color:#77758A;">This conversation needs your attention.</p>
|
||||
</td></tr>
|
||||
<tr><td style="padding:8px 24px 0;">
|
||||
<table role="presentation" width="100%" cellpadding="0" cellspacing="0" style="background:#FAF9FB;border:1px solid #ECE9EF;border-radius:8px;">
|
||||
<tr><td style="padding:14px;">
|
||||
<p style="margin:0 0 10px;font-size:10px;font-weight:600;letter-spacing:.05em;text-transform:uppercase;color:#8A8695;">From</p>
|
||||
<p style="margin:0 0 14px;font-size:13px;font-weight:500;color:#292536;">{escape(external_sender)}</p>
|
||||
<p style="margin:0 0 10px;font-size:10px;font-weight:600;letter-spacing:.05em;text-transform:uppercase;color:#8A8695;">Subject</p>
|
||||
<p style="margin:0;font-size:13px;font-weight:500;color:#292536;">{escape(_clean_external_subject(external_subject))}</p>
|
||||
</td></tr>
|
||||
</table>
|
||||
</td></tr>
|
||||
<tr><td style="padding:18px 24px 0;">
|
||||
<p style="margin:0 0 8px;font-size:11px;font-weight:600;letter-spacing:.05em;text-transform:uppercase;color:#8A8695;">Latest message</p>
|
||||
<div style="border-left:3px solid #7357F6;padding:4px 0 4px 14px;font-size:14px;line-height:1.55;color:#292536;white-space:pre-wrap;">{escape(external_body.strip())}</div>
|
||||
</td></tr>
|
||||
{reason_block}
|
||||
<tr><td style="padding:18px 24px 24px;">
|
||||
<table role="presentation" width="100%" cellpadding="0" cellspacing="0" style="background:#F7F5FF;border:1px solid #E5DFFF;border-radius:8px;">
|
||||
<tr><td style="padding:14px;">
|
||||
<p style="margin:0 0 8px;font-size:13px;font-weight:650;color:#5633CE;">↩ Reply normally</p>
|
||||
<p style="margin:0;font-size:12px;line-height:1.5;color:#625D70;">
|
||||
Reply to this email and Gentic will send your response to the original sender as
|
||||
<strong>{escape(agent_email)}</strong>. Only the new text you write will be sent.
|
||||
This handoff email and quoted history stay private.
|
||||
</p>
|
||||
</td></tr>
|
||||
</table>
|
||||
</td></tr>
|
||||
<tr><td style="padding:0 24px 24px;text-align:center;">
|
||||
<p style="margin:0;font-size:10px;color:#A19DAA;">Gentic Human Handoff<br>Private Agent ↔ Human conversation</p>
|
||||
</td></tr>
|
||||
</table>
|
||||
</td></tr>
|
||||
</table>
|
||||
</body></html>"""
|
||||
|
||||
|
||||
def build_update_handoff_text(
|
||||
*,
|
||||
external_sender: str,
|
||||
external_body: str,
|
||||
agent_email: str,
|
||||
) -> str:
|
||||
return "\n".join(
|
||||
[
|
||||
"Customer replied",
|
||||
"",
|
||||
external_sender,
|
||||
"",
|
||||
external_body.strip(),
|
||||
"",
|
||||
f"Reply normally to respond through {agent_email}.",
|
||||
]
|
||||
)
|
||||
|
||||
|
||||
def build_update_handoff_html(
|
||||
*,
|
||||
external_sender: str,
|
||||
external_body: str,
|
||||
agent_email: str,
|
||||
) -> str:
|
||||
return f"""<!DOCTYPE html>
|
||||
<html><body style="margin:0;padding:0;background:#F5F4F2;">
|
||||
<table role="presentation" width="100%" cellpadding="0" cellspacing="0" style="background:#F5F4F2;padding:24px 12px;">
|
||||
<tr><td align="center">
|
||||
<table role="presentation" width="600" cellpadding="0" cellspacing="0" style="width:100%;max-width:600px;background:#FFFFFF;border:1px solid #E8E5EA;border-radius:12px;">
|
||||
<tr><td style="padding:24px;">
|
||||
<p style="margin:0 0 8px;font-size:12px;font-weight:600;color:#7357F6;">✦ Gentic</p>
|
||||
<h1 style="margin:0 0 16px;font-size:18px;font-weight:700;color:#171424;">Customer replied</h1>
|
||||
<p style="margin:0 0 12px;font-size:13px;font-weight:500;color:#292536;">{escape(external_sender)}</p>
|
||||
<div style="border-left:3px solid #7357F6;padding:4px 0 4px 14px;font-size:14px;line-height:1.55;color:#292536;white-space:pre-wrap;">{escape(external_body.strip())}</div>
|
||||
<p style="margin:18px 0 0;font-size:12px;line-height:1.5;color:#625D70;">
|
||||
Reply normally to respond through <strong>{escape(agent_email)}</strong>.
|
||||
</p>
|
||||
</td></tr>
|
||||
</table>
|
||||
</td></tr>
|
||||
</table>
|
||||
</body></html>"""
|
||||
|
||||
|
||||
def build_relay_confirmation_text(customer_body: str) -> str:
|
||||
return f"Sent to customer:\n\n{customer_body.strip()}"
|
||||
|
|
@ -4,9 +4,19 @@ from datetime import datetime, timezone
|
|||
|
||||
from bson import ObjectId
|
||||
|
||||
from app.services.handoff_email_templates import (
|
||||
build_initial_handoff_html,
|
||||
build_initial_handoff_text,
|
||||
build_relay_confirmation_text,
|
||||
build_update_handoff_html,
|
||||
build_update_handoff_text,
|
||||
external_reply_subject,
|
||||
handoff_email_subject,
|
||||
)
|
||||
from app.services.handoff_evaluator import evaluate_handoff
|
||||
from app.services.mail_sender import deliver_email
|
||||
from app.services.message_ids import build_references_header, normalize_message_id, parse_message_id_list
|
||||
from app.services.message_ids import normalize_message_id, parse_message_id_list
|
||||
from app.services.reply_extractor import extract_new_reply_content, validate_external_reply_body
|
||||
from app.services.thread_service import (
|
||||
attach_message_to_thread,
|
||||
message_event_at,
|
||||
|
|
@ -42,36 +52,8 @@ def _snippet(text: str, limit: int = 240) -> str:
|
|||
return re.sub(r"\s+", " ", (text or "")).strip()[:limit]
|
||||
|
||||
|
||||
def _handoff_subject(external_subject: str) -> str:
|
||||
subject = (external_subject or "(No subject)").strip()
|
||||
if subject.lower().startswith("human handoff:"):
|
||||
return subject
|
||||
return f"Human Handoff: {subject}"
|
||||
|
||||
|
||||
def _build_handoff_body(
|
||||
*,
|
||||
external_sender: str,
|
||||
external_subject: str,
|
||||
external_body: str,
|
||||
handoff_reason: str,
|
||||
agent_email: str,
|
||||
) -> str:
|
||||
return (
|
||||
"Human Handoff\n\n"
|
||||
"A conversation needs your attention.\n\n"
|
||||
f"From\n{external_sender}\n\n"
|
||||
f"Subject\n{external_subject}\n\n"
|
||||
"Latest message\n\n"
|
||||
f"\"{external_body.strip()}\"\n\n"
|
||||
"Why this was handed off\n\n"
|
||||
f"{handoff_reason}\n\n"
|
||||
"────────────────────────────────────────\n\n"
|
||||
f"Reply to this email to respond to the original sender through {agent_email}.\n\n"
|
||||
"Your reply will be sent to the customer.\n\n"
|
||||
"────────────────────────────────────────\n\n"
|
||||
"Gentic Human Handoff"
|
||||
)
|
||||
def _handoff_reply_subject(external_subject: str) -> str:
|
||||
return f"Re: {handoff_email_subject(external_subject)}"
|
||||
|
||||
|
||||
async def create_empty_handoff_thread(
|
||||
|
|
@ -114,6 +96,47 @@ async def _record_message_link(
|
|||
)
|
||||
|
||||
|
||||
async def _latest_external_customer_message(
|
||||
database,
|
||||
account_id: str,
|
||||
external_thread_id: ObjectId,
|
||||
customer_address: str,
|
||||
) -> dict | None:
|
||||
customer = customer_address.lower()
|
||||
pipeline = [
|
||||
{
|
||||
"$match": {
|
||||
"account_id": account_id,
|
||||
"thread_id": external_thread_id,
|
||||
"message_scope": MESSAGE_SCOPE_EXTERNAL,
|
||||
"deletion_state": {"$in": [None, ""]},
|
||||
}
|
||||
},
|
||||
{"$sort": {"event_at": -1, "_id": -1}},
|
||||
{
|
||||
"$lookup": {
|
||||
"from": "messages",
|
||||
"localField": "message_id",
|
||||
"foreignField": "_id",
|
||||
"as": "message",
|
||||
}
|
||||
},
|
||||
{"$unwind": "$message"},
|
||||
{
|
||||
"$match": {
|
||||
"$or": [
|
||||
{"message.direction": "inbound"},
|
||||
{"message.sender.address": customer},
|
||||
]
|
||||
}
|
||||
},
|
||||
{"$limit": 1},
|
||||
]
|
||||
cursor = await database.mailbox_entries.aggregate(pipeline)
|
||||
rows = [row async for row in cursor]
|
||||
return rows[0]["message"] if rows else None
|
||||
|
||||
|
||||
async def _latest_external_message(database, account_id: str, external_thread_id: ObjectId) -> dict | None:
|
||||
pipeline = [
|
||||
{
|
||||
|
|
@ -250,8 +273,12 @@ async def store_handoff_message(
|
|||
link_relationship: str | None = None,
|
||||
external_message_id: ObjectId | None = None,
|
||||
event_at: datetime | None = None,
|
||||
raw_body: str | None = None,
|
||||
new_reply_content: str | None = None,
|
||||
body_html: str | None = None,
|
||||
) -> dict:
|
||||
now = event_at or _now()
|
||||
display_body = new_reply_content if new_reply_content is not None else body_text
|
||||
document = {
|
||||
"direction": direction,
|
||||
"message_scope": MESSAGE_SCOPE_HANDOFF,
|
||||
|
|
@ -259,9 +286,11 @@ async def store_handoff_message(
|
|||
"sender": sender,
|
||||
"to": recipients,
|
||||
"subject": subject,
|
||||
"snippet": _snippet(body_text),
|
||||
"body_text": body_text,
|
||||
"body_html": None,
|
||||
"snippet": _snippet(display_body),
|
||||
"body_text": display_body,
|
||||
"raw_body": raw_body,
|
||||
"new_reply_content": new_reply_content,
|
||||
"body_html": body_html,
|
||||
"attachments": [],
|
||||
"created_at": now,
|
||||
"sent_at": now if direction == "outbound" else None,
|
||||
|
|
@ -331,17 +360,21 @@ async def send_handoff_to_human(
|
|||
external_subject = external_message.get("subject", "")
|
||||
external_body = external_message.get("body_text") or external_message.get("snippet") or ""
|
||||
|
||||
stable_subject = handoff_email_subject(external_subject)
|
||||
if existing_handoff:
|
||||
handoff_id = existing_handoff["_id"]
|
||||
handoff_thread_id = existing_handoff["handoff_thread_id"]
|
||||
parent = await _latest_handoff_message(database, account_id, handoff_thread_id)
|
||||
threading = threading_fields_from_parent(parent) if parent else {}
|
||||
subject = _handoff_subject(external_subject)
|
||||
body = _build_handoff_body(
|
||||
subject = _handoff_reply_subject(external_subject)
|
||||
body_text = build_update_handoff_text(
|
||||
external_sender=external_sender_text,
|
||||
external_body=external_body,
|
||||
agent_email=agent_email,
|
||||
)
|
||||
body_html = build_update_handoff_html(
|
||||
external_sender=external_sender_text,
|
||||
external_subject=external_subject,
|
||||
external_body=external_body,
|
||||
handoff_reason="New message in the linked conversation.",
|
||||
agent_email=agent_email,
|
||||
)
|
||||
relationship = LINK_EXTERNAL_UPDATE_RELAYED
|
||||
|
|
@ -349,11 +382,18 @@ async def send_handoff_to_human(
|
|||
handoff_thread_id = await create_empty_handoff_thread(
|
||||
database,
|
||||
account_id=account_id,
|
||||
subject=_handoff_subject(external_subject),
|
||||
subject=stable_subject,
|
||||
)
|
||||
threading = {}
|
||||
subject = _handoff_subject(external_subject)
|
||||
body = _build_handoff_body(
|
||||
subject = stable_subject
|
||||
body_text = build_initial_handoff_text(
|
||||
external_sender=external_sender_text,
|
||||
external_subject=external_subject,
|
||||
external_body=external_body,
|
||||
handoff_reason=handoff_reason,
|
||||
agent_email=agent_email,
|
||||
)
|
||||
body_html = build_initial_handoff_html(
|
||||
external_sender=external_sender_text,
|
||||
external_subject=external_subject,
|
||||
external_body=external_body,
|
||||
|
|
@ -386,8 +426,8 @@ async def send_handoff_to_human(
|
|||
cc=[],
|
||||
bcc=[],
|
||||
subject=subject,
|
||||
body_text=body,
|
||||
body_html=None,
|
||||
body_text=body_text,
|
||||
body_html=body_html,
|
||||
domain=domain,
|
||||
in_reply_to=threading.get("in_reply_to"),
|
||||
references=threading.get("references"),
|
||||
|
|
@ -402,7 +442,8 @@ async def send_handoff_to_human(
|
|||
sender={"name": "Gentic Agent", "address": agent_email},
|
||||
recipients=[{"name": human_email.split("@")[0], "address": human_email}],
|
||||
subject=subject,
|
||||
body_text=body,
|
||||
body_text=body_text,
|
||||
body_html=body_html,
|
||||
internet_message_id=internet_message_id,
|
||||
in_reply_to=threading.get("in_reply_to"),
|
||||
references=threading.get("references"),
|
||||
|
|
@ -458,6 +499,13 @@ async def process_human_reply(
|
|||
rule = await _handoff_rule_for_human(database, account_id, human_email)
|
||||
allow_replies = rule.get("allow_replies", True) if rule else True
|
||||
|
||||
extracted = extract_new_reply_content(
|
||||
body_text=inbound_body,
|
||||
body_html=inbound_parsed.get("body_html"),
|
||||
)
|
||||
new_reply_content = extracted["new_reply_content"].strip()
|
||||
raw_body = extracted["raw_body"]
|
||||
|
||||
human_message = await store_handoff_message(
|
||||
database,
|
||||
account_id=account_id,
|
||||
|
|
@ -467,7 +515,10 @@ async def process_human_reply(
|
|||
sender={"name": human_email.split("@")[0], "address": human_email},
|
||||
recipients=[{"name": "Gentic Agent", "address": agent_email}],
|
||||
subject=inbound_parsed.get("subject", ""),
|
||||
body_text=inbound_body,
|
||||
body_text=raw_body,
|
||||
raw_body=raw_body,
|
||||
new_reply_content=new_reply_content,
|
||||
body_html=inbound_parsed.get("body_html"),
|
||||
internet_message_id=inbound_parsed.get("message_id"),
|
||||
in_reply_to=inbound_parsed.get("in_reply_to"),
|
||||
references=inbound_parsed.get("references", []),
|
||||
|
|
@ -490,21 +541,56 @@ async def process_human_reply(
|
|||
)
|
||||
return
|
||||
|
||||
external_parent = await _latest_external_message(database, account_id, external_thread_id)
|
||||
if not new_reply_content:
|
||||
logger.warning(
|
||||
"handoff_reply_empty_extract handoff_id=%s account=%s",
|
||||
handoff["_id"],
|
||||
account_id,
|
||||
)
|
||||
await database.human_handoffs.update_one(
|
||||
{"_id": handoff["_id"]},
|
||||
{
|
||||
"$set": {
|
||||
"status": HANDOFF_STATUS_HUMAN_REPLIED,
|
||||
"current_owner": OWNER_AGENT,
|
||||
"updated_at": now,
|
||||
"last_human_message_at": now,
|
||||
}
|
||||
},
|
||||
)
|
||||
return
|
||||
|
||||
if not validate_external_reply_body(new_reply_content):
|
||||
logger.error(
|
||||
"handoff_reply_leakage_guard handoff_id=%s account=%s",
|
||||
handoff["_id"],
|
||||
account_id,
|
||||
)
|
||||
return
|
||||
|
||||
external_recipient = handoff.get("external_sender") or ""
|
||||
external_parent = await _latest_external_customer_message(
|
||||
database,
|
||||
account_id,
|
||||
external_thread_id,
|
||||
external_recipient,
|
||||
)
|
||||
if external_parent is None:
|
||||
external_parent = await _latest_external_message(database, account_id, external_thread_id)
|
||||
if external_parent is None:
|
||||
logger.warning("handoff_reply_missing_external_parent handoff_id=%s", handoff["_id"])
|
||||
return
|
||||
|
||||
threading = threading_fields_from_parent(external_parent)
|
||||
external_subject = handoff.get("external_subject") or external_parent.get("subject", "")
|
||||
if not re.match(r"^re:\s*", external_subject, flags=re.IGNORECASE):
|
||||
external_subject = f"Re: {external_subject}"
|
||||
external_subject = external_reply_subject(
|
||||
handoff.get("external_subject") or external_parent.get("subject", "")
|
||||
)
|
||||
|
||||
external_recipient = handoff.get("external_sender") or external_parent.get("sender", {}).get("address")
|
||||
external_recipient = external_recipient or external_parent.get("sender", {}).get("address")
|
||||
if not external_recipient:
|
||||
return
|
||||
|
||||
customer_body = inbound_body.strip()
|
||||
customer_body = new_reply_content
|
||||
internet_message_id = await deliver_email(
|
||||
sender=agent_email,
|
||||
recipients=[external_recipient],
|
||||
|
|
@ -572,8 +658,8 @@ async def process_human_reply(
|
|||
}
|
||||
)
|
||||
|
||||
relay_body = f"Sent to customer:\n\n\"{customer_body}\""
|
||||
relay_subject = f"Re: {_handoff_subject(handoff.get('external_subject', ''))}"
|
||||
relay_body = build_relay_confirmation_text(customer_body)
|
||||
relay_subject = _handoff_reply_subject(handoff.get("external_subject", ""))
|
||||
relay_threading = threading_fields_from_parent(human_message)
|
||||
relay_internet_id = await deliver_email(
|
||||
sender=agent_email,
|
||||
|
|
@ -656,7 +742,7 @@ async def maybe_trigger_handoff(
|
|||
existing = await find_active_handoff_for_external_thread(
|
||||
database, account_id, external_thread_id
|
||||
)
|
||||
reason = evaluation.get("instruction_reason") or rule.get("instructions") or "Matched handoff rules."
|
||||
reason = (rule.get("instructions") or "").strip() or "This conversation requires human attention."
|
||||
await send_handoff_to_human(
|
||||
database,
|
||||
account_id=account_id,
|
||||
|
|
|
|||
119
backend/app/services/reply_extractor.py
Normal file
119
backend/app/services/reply_extractor.py
Normal file
|
|
@ -0,0 +1,119 @@
|
|||
import re
|
||||
from html import unescape
|
||||
|
||||
# Patterns that mark the start of quoted previous-message content.
|
||||
TEXT_QUOTE_MARKERS = [
|
||||
re.compile(r"\n-{2,}\s*Original Message\s*-{2,}", re.IGNORECASE),
|
||||
re.compile(
|
||||
r"\nOn .+? wrote:\s*\n",
|
||||
re.IGNORECASE,
|
||||
),
|
||||
re.compile(
|
||||
r"\nOn .+?, .+? at .+? .+? wrote:\s*\n",
|
||||
re.IGNORECASE,
|
||||
),
|
||||
re.compile(
|
||||
r"\nFrom:\s*.+\nSent:\s*.+\nTo:\s*.+\nSubject:\s*.+",
|
||||
re.IGNORECASE,
|
||||
),
|
||||
re.compile(r"\n_{5,}\n"),
|
||||
]
|
||||
|
||||
HTML_QUOTE_PATTERNS = [
|
||||
re.compile(r"<blockquote\b[^>]*>.*?</blockquote>", re.IGNORECASE | re.DOTALL),
|
||||
re.compile(r'<div[^>]+class="[^"]*gmail_quote[^"]*"[^>]*>.*?</div>', re.IGNORECASE | re.DOTALL),
|
||||
re.compile(r'<div[^>]+class="[^"]*gmail_attr[^"]*"[^>]*>.*?</div>', re.IGNORECASE | re.DOTALL),
|
||||
]
|
||||
|
||||
PRIVATE_LEAK_MARKERS = [
|
||||
"human handoff",
|
||||
"a conversation needs your attention",
|
||||
"why this was handed off",
|
||||
"why you're receiving this",
|
||||
"gentic human handoff",
|
||||
"no handoff instructions were set",
|
||||
"step passes",
|
||||
"classifier returned",
|
||||
"rule evaluation succeeded",
|
||||
"reply normally to respond through",
|
||||
"latest message",
|
||||
"private agent ↔ human conversation",
|
||||
]
|
||||
|
||||
|
||||
def _normalize_newlines(text: str) -> str:
|
||||
return text.replace("\r\n", "\n").replace("\r", "\n")
|
||||
|
||||
|
||||
def _html_to_text(html: str) -> str:
|
||||
cleaned = html
|
||||
for pattern in HTML_QUOTE_PATTERNS:
|
||||
cleaned = pattern.sub("", cleaned)
|
||||
cleaned = re.sub(r"<br\s*/?>", "\n", cleaned, flags=re.IGNORECASE)
|
||||
cleaned = re.sub(r"</p\s*>", "\n", cleaned, flags=re.IGNORECASE)
|
||||
cleaned = re.sub(r"<[^>]+>", "", cleaned)
|
||||
return unescape(cleaned)
|
||||
|
||||
|
||||
def _strip_trailing_quoted_lines(text: str) -> str:
|
||||
lines = text.split("\n")
|
||||
kept: list[str] = []
|
||||
for line in lines:
|
||||
if re.match(r"^\s*>", line):
|
||||
break
|
||||
kept.append(line)
|
||||
return "\n".join(kept).strip()
|
||||
|
||||
|
||||
def _split_at_quote_markers(text: str) -> tuple[str, str]:
|
||||
normalized = _normalize_newlines(text)
|
||||
earliest_index = None
|
||||
for pattern in TEXT_QUOTE_MARKERS:
|
||||
match = pattern.search(normalized)
|
||||
if match and (earliest_index is None or match.start() < earliest_index):
|
||||
earliest_index = match.start()
|
||||
if earliest_index is not None:
|
||||
new_content = normalized[:earliest_index].strip()
|
||||
quoted = normalized[earliest_index:].strip()
|
||||
return new_content, quoted
|
||||
return _strip_trailing_quoted_lines(normalized), ""
|
||||
|
||||
|
||||
def extract_new_reply_content(
|
||||
*,
|
||||
body_text: str | None,
|
||||
body_html: str | None = None,
|
||||
) -> dict[str, str]:
|
||||
raw_text = _normalize_newlines((body_text or "").strip())
|
||||
if body_html and not raw_text:
|
||||
raw_text = _html_to_text(body_html).strip()
|
||||
|
||||
if body_html:
|
||||
html_stripped = body_html
|
||||
for pattern in HTML_QUOTE_PATTERNS:
|
||||
html_stripped = pattern.sub("", html_stripped)
|
||||
html_text = _html_to_text(html_stripped).strip()
|
||||
if html_text:
|
||||
new_from_html, quoted_from_html = _split_at_quote_markers(html_text)
|
||||
if new_from_html:
|
||||
return {
|
||||
"raw_body": raw_text or html_text,
|
||||
"new_reply_content": new_from_html,
|
||||
"quoted_content": quoted_from_html,
|
||||
}
|
||||
|
||||
new_content, quoted_content = _split_at_quote_markers(raw_text)
|
||||
return {
|
||||
"raw_body": raw_text,
|
||||
"new_reply_content": new_content,
|
||||
"quoted_content": quoted_content,
|
||||
}
|
||||
|
||||
|
||||
def validate_external_reply_body(body: str) -> bool:
|
||||
normalized = (body or "").strip()
|
||||
if not normalized:
|
||||
return False
|
||||
lower = normalized.lower()
|
||||
hits = sum(1 for marker in PRIVATE_LEAK_MARKERS if marker in lower)
|
||||
return hits == 0
|
||||
Loading…
Add table
Add a link
Reference in a new issue