handoff fixes

backend/tests/test_email_threading_roundtrip.py

@@ -0,0 +1,76 @@
+import pytest
+from bson import ObjectId
+
+from app.services.thread_resolver import KnownMessage, resolve_thread_from_known
+
+
+def _known(
+    internet_id: str,
+    thread_id: ObjectId,
+    *,
+    references: list[str] | None = None,
+) -> KnownMessage:
+    from datetime import datetime, timezone
+
+    from app.services.message_ids import normalize_message_id
+    from app.services.subject_normalize import normalize_subject
+
+    now = datetime.now(timezone.utc)
+    return KnownMessage(
+        id=ObjectId(),
+        thread_id=thread_id,
+        internet_message_id=normalize_message_id(internet_id) or "",
+        account_id="acct-1",
+        sent_at=now,
+        received_at=now,
+        created_at=now,
+        participants=frozenset({"customer@gmail.com"}),
+        normalized_subject=normalize_subject("lets try one more time"),
+    )
+
+
+def test_gmail_reply_to_gentic_resolves_same_thread():
+    thread_id = ObjectId()
+    a1 = _known("<a1@gmail.com>", thread_id)
+    g1 = _known("<g1@gentic.email>", thread_id, references=["<a1@gmail.com>"])
+    known = {
+        a1.internet_message_id: a1,
+        g1.internet_message_id: g1,
+    }
+    resolution = resolve_thread_from_known(
+        account_id="acct-1",
+        in_reply_to="<g1@gentic.email>",
+        references=["<a1@gmail.com>", "<g1@gentic.email>"],
+        subject="Re: lets try one more time",
+        participants={"customer@gmail.com"},
+        known_by_internet_id=known,
+    )
+    assert resolution.thread_id == thread_id
+    assert resolution.thread_resolution_method == "in_reply_to"
+    assert resolution.parent_message_id == g1.id
+
+
+@pytest.mark.asyncio
+async def test_lookup_known_messages_awaits_aggregate():
+    from app.services.thread_resolver import lookup_known_messages
+
+    class FakeCursor:
+        def __aiter__(self):
+            return self
+
+        async def __anext__(self):
+            raise StopAsyncIteration
+
+    class FakeMessages:
+        def __init__(self):
+            self.aggregate_awaited = False
+
+        async def aggregate(self, pipeline):
+            self.aggregate_awaited = True
+            return FakeCursor()
+
+    messages = FakeMessages()
+    database = type("DB", (), {"messages": messages})()
+    result = await lookup_known_messages(database, "acct-1", ["<g1@gentic.email>"])
+    assert result == {}
+    assert messages.aggregate_awaited is True

backend/tests/test_external_reply_service.py

@@ -0,0 +1,57 @@
+from bson import ObjectId
+
+from app.services.external_reply_service import build_external_reply_plan
+from app.services.thread_service import threading_fields_from_parent
+
+
+def test_build_external_reply_references_gmail_parent():
+    parent = {
+        "_id": ObjectId(),
+        "internet_message_id": "<gmail-a@mail.gmail.com>",
+        "message_id": "<gmail-a@mail.gmail.com>",
+        "references": [],
+        "subject": "lets try one more time",
+    }
+    plan = build_external_reply_plan(
+        parent_external_message=parent,
+        body_text="I hear you.",
+        recipients=["customer@gmail.com"],
+    )
+    assert plan.threading.in_reply_to == "<gmail-a@mail.gmail.com>"
+    assert plan.threading.references == ["<gmail-a@mail.gmail.com>"]
+    assert plan.subject == "Re: lets try one more time"
+    assert plan.threading.replying_to_internet_message_id == "<gmail-a@mail.gmail.com>"
+
+
+def test_build_external_reply_chains_references_for_multi_round():
+    parent = {
+        "_id": ObjectId(),
+        "internet_message_id": "<a2@gmail.com>",
+        "message_id": "<a2@gmail.com>",
+        "references": ["<a1@gmail.com>", "<g1@gentic.email>"],
+        "subject": "Re: lets try one more time",
+    }
+    threading = threading_fields_from_parent(parent)
+    assert threading["in_reply_to"] == "<a2@gmail.com>"
+    assert threading["references"] == [
+        "<a1@gmail.com>",
+        "<g1@gentic.email>",
+        "<a2@gmail.com>",
+    ]
+
+
+def test_external_reply_never_uses_internal_object_id_as_in_reply_to():
+    parent = {
+        "_id": ObjectId(),
+        "internet_message_id": "<gmail-a@mail.gmail.com>",
+        "message_id": "<gmail-a@mail.gmail.com>",
+        "references": [],
+        "subject": "hello",
+    }
+    plan = build_external_reply_plan(
+        parent_external_message=parent,
+        body_text="reply",
+        recipients=["customer@gmail.com"],
+    )
+    assert str(parent["_id"]) not in (plan.threading.in_reply_to or "")
+    assert all(str(parent["_id"]) not in ref for ref in plan.threading.references)

backend/tests/test_handoff_mirror.py

@@ -0,0 +1,85 @@
+from app.services.handoff_email_templates import (
+    build_mirror_handoff_html,
+    build_mirror_handoff_text,
+    external_reply_subject,
+    format_sender_line,
+    handoff_email_subject,
+)
+from app.services.html_sanitizer import sanitize_email_html
+from app.services.mime_body_extractor import extract_mirror_content
+from app.services.thread_service import threading_fields_from_parent
+
+
+def test_mirror_template_is_minimal():
+    html = build_mirror_handoff_html(
+        sender_line="091rabbytes@gmail.com",
+        external_subject="lets try one more time",
+        original_html="<p>one last attempt. i am tired</p>",
+        agent_email="suyash@gentic.email",
+    )
+    assert "LATEST MESSAGE" not in html
+    assert "Why you" not in html
+    assert "gentic-handoff-original" in html
+    assert "one last attempt. i am tired" in html
+    assert "✦ Human Handoff" in html
+    assert "091rabbytes@gmail.com" in html
+    assert "lets try one more time" in html
+
+
+def test_mirror_plain_text_template():
+    text = build_mirror_handoff_text(
+        sender_line="091rabbytes@gmail.com",
+        external_subject="lets try one more time",
+        original_body_text="one last attempt. i am tired",
+        agent_email="suyash@gentic.email",
+    )
+    assert "one last attempt. i am tired" in text
+    assert "LATEST MESSAGE" not in text
+    assert "Reply normally" in text
+
+
+def test_format_sender_with_name():
+    assert (
+        format_sender_line({"name": "Rahul Sharma", "address": "rahul@example.com"})
+        == "Rahul Sharma <rahul@example.com>"
+    )
+
+
+def test_external_threading_headers_from_customer_message():
+    parent = {
+        "_id": "mongo-id",
+        "message_id": "<a@gmail.com>",
+        "references": [],
+    }
+    threading = threading_fields_from_parent(parent)
+    assert threading["in_reply_to"] == "<a@gmail.com>"
+    assert threading["references"] == ["<a@gmail.com>"]
+
+
+def test_external_reply_subject_uses_customer_subject():
+    assert external_reply_subject("lets try one more time") == "Re: lets try one more time"
+    assert handoff_email_subject("lets try one more time") == "Human Handoff · lets try one more time"
+
+
+def test_sanitize_removes_script():
+    dirty = '<p>Hello</p><script>alert(1)</script><img src="x" onerror="alert(1)">'
+    clean = sanitize_email_html(dirty)
+    assert "<script" not in clean.lower()
+    assert "onerror" not in clean.lower()
+    assert "Hello" in clean
+
+
+def test_extract_html_from_multipart():
+    raw = (
+        b"From: sender@example.com\r\n"
+        b"To: agent@gentic.email\r\n"
+        b"Subject: Invoice approval\r\n"
+        b'MIME-Version: 1.0\r\nContent-Type: multipart/alternative; boundary="b"\r\n\r\n'
+        b"--b\r\nContent-Type: text/plain; charset=utf-8\r\n\r\nPlain fallback\r\n"
+        b"--b\r\nContent-Type: text/html; charset=utf-8\r\n\r\n"
+        b"<h2>Invoice</h2><table><tr><td>Amount</td></tr></table>\r\n"
+        b"--b--\r\n"
+    )
+    extracted = extract_mirror_content(raw)
+    assert "Invoice" in (extracted.body_html or "")
+    assert "<table>" in (extracted.body_html or "")

backend/tests/test_reply_extractor.py

@@ -1,4 +1,4 @@
-from app.services.handoff_email_templates import external_reply_subject, handoff_email_subject
+from app.services.handoff_email_templates import external_reply_subject
 from app.services.reply_extractor import extract_new_reply_content, validate_external_reply_body
 
 SCREENSHOT_REPLY = """ahan, did you receive a response?
@@ -49,10 +49,6 @@ def test_external_reply_subject_uses_customer_subject():
     assert external_reply_subject("Human Handoff · are you ready?") == "Re: are you ready?"
 
 
-def test_handoff_subject_format():
-    assert handoff_email_subject("are you ready?") == "Human Handoff · are you ready?"
-
-
 def test_outlook_original_message_split():
     body = """Thanks for the update.