test: cover forensic ingestion safeguards

This commit is contained in:
Christian Krakau-Louis
2026-05-23 13:08:22 +02:00
parent 817270b1fb
commit 8a4795187e
4 changed files with 179 additions and 1 deletions
+71 -1
View File
@@ -1,8 +1,17 @@
import email
from unittest.mock import MagicMock
import pytest
from app.services.forensic_parser import ForensicParser, MAX_FORENSIC_REPORT_SIZE
from app.services import forensic_parser as forensic_parser_module
from app.services.forensic_parser import (
ForensicParser,
MAX_FORENSIC_REPORT_SIZE,
_coerce_text,
_domain_from_address,
_message_part_payload,
_payload_text,
)
SAMPLE_FORENSIC_EMAIL = b"""\
@@ -177,3 +186,64 @@ def test_parse_forensic_email_rejects_empty_and_oversized_payloads():
with pytest.raises(ValueError, match="too large"):
ForensicParser.parse_bytes(b"x" * (MAX_FORENSIC_REPORT_SIZE + 1))
def test_forensic_parser_helper_fallbacks(monkeypatch):
multipart = email.message_from_bytes(
b"Subject: container\r\nContent-Type: multipart/mixed; boundary=x\r\n\r\n--x--"
)
message = email.message_from_bytes(b"Subject: plain\r\n\r\nbody")
assert _coerce_text(b"hello") == "hello"
assert _payload_text(multipart) == ""
assert _message_part_payload(message) is None
assert _domain_from_address("not-an-address") == ""
part = MagicMock()
part.get_payload.side_effect = [None, ["nested"]]
assert _payload_text(part) == ""
part = MagicMock()
part.get_payload.side_effect = [None, "plain text"]
assert _payload_text(part) == "plain text"
monkeypatch.setattr(forensic_parser_module, "parsedate_to_datetime", lambda _value: None)
assert forensic_parser_module._parse_datetime("Fri, 22 May 2026 10:15:00 +0000") is None
def test_is_forensic_report_detects_feedback_part_without_report_container():
msg = email.message_from_bytes(
b"""\
Subject: Delivery notice
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="b"
--b
Content-Type: message/feedback-report
Feedback-Type: auth-failure
--b--
"""
)
assert ForensicParser.is_forensic_report(msg) is True
def test_is_forensic_report_ignores_non_dmarc_header_parts():
msg = email.message_from_bytes(
b"""\
Subject: DMARC forensic notice
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="b"
--b
Content-Type: text/rfc822-headers
Authentication-Results: mx; spf=pass
--b--
"""
)
assert ForensicParser.is_forensic_report(msg) is True
+66
View File
@@ -1,5 +1,8 @@
import pytest
from fastapi import HTTPException
from sqlalchemy.exc import IntegrityError
from app.api.api_v1.endpoints import forensics as forensics_endpoint
from app.models.report import ForensicReport
from app.services.forensic_parser import ForensicParser
from app.services.forensic_persistence import (
@@ -77,6 +80,22 @@ def test_upload_forensic_report_rejects_empty_file(authed_client):
assert response.status_code == 400
def test_validate_upload_rejects_missing_name_and_large_file():
missing_name = type("Upload", (), {"filename": ""})()
too_large = type("Upload", (), {"filename": "report.eml"})()
with pytest.raises(HTTPException) as missing:
forensics_endpoint._validate_upload(missing_name, b"content")
assert missing.value.status_code == 400
with pytest.raises(HTTPException) as large:
forensics_endpoint._validate_upload(
too_large,
b"x" * (forensics_endpoint.MAX_FORENSIC_REPORT_SIZE + 1),
)
assert large.value.status_code == 413
def test_upload_forensic_report_rejects_invalid_email(authed_client):
response = authed_client.post(
"/api/v1/forensics/upload",
@@ -86,6 +105,39 @@ def test_upload_forensic_report_rejects_invalid_email(authed_client):
assert response.status_code == 400
def test_upload_forensic_report_handles_save_duplicate_race(authed_client, monkeypatch):
parsed = ForensicParser.parse_bytes(SAMPLE_FORENSIC_EMAIL)
row = ForensicReport(report_id=parsed["report_id"], reported_domain=parsed["reported_domain"])
monkeypatch.setattr(forensics_endpoint, "forensic_report_exists", lambda *_args: False)
monkeypatch.setattr(
forensics_endpoint,
"save_forensic_report",
lambda *_args: (row, False),
)
response = authed_client.post(
"/api/v1/forensics/upload",
files={"file": ("report.eml", SAMPLE_FORENSIC_EMAIL, "message/rfc822")},
)
assert response.status_code == 409
def test_upload_forensic_report_unexpected_error_returns_500(authed_client, monkeypatch):
monkeypatch.setattr(
forensics_endpoint.ForensicParser,
"parse_bytes",
lambda _content: (_ for _ in ()).throw(RuntimeError("boom")),
)
response = authed_client.post(
"/api/v1/forensics/upload",
files={"file": ("report.eml", SAMPLE_FORENSIC_EMAIL, "message/rfc822")},
)
assert response.status_code == 500
def test_forensic_detail_returns_404(authed_client):
response = authed_client.get("/api/v1/forensics/999")
@@ -119,3 +171,17 @@ def test_save_forensic_report_duplicate_and_invalid_domain_paths(db_session):
row, invalid_created = save_forensic_report(db_session, invalid)
assert invalid_created is True
assert row.domain_id is None
def test_save_forensic_report_reraises_unexpected_integrity_errors(db_session, monkeypatch):
parsed = ForensicParser.parse_bytes(SAMPLE_FORENSIC_EMAIL)
parsed["report_id"] = "ruf-race-without-existing-row"
parsed["reported_domain"] = ""
def raise_integrity_error():
raise IntegrityError("insert", {}, Exception("unique"))
monkeypatch.setattr(db_session, "flush", raise_integrity_error)
with pytest.raises(IntegrityError):
save_forensic_report(db_session, parsed)
+24
View File
@@ -493,6 +493,24 @@ class TestProcessMessage:
assert stats["duplicate_forensic_reports"] == 1
assert stats["details"][1]["status"] == "duplicate"
def test_forensic_save_duplicate_result_is_skipped(self, db_session):
client = _make_client(db=db_session)
stats = {"forensic_reports_found": 0, "duplicate_forensic_reports": 0, "errors": []}
with (
patch("app.services.gmail_client.forensic_report_exists", return_value=False),
patch("app.services.gmail_client.save_forensic_report", return_value=(None, False)),
):
imported = client._process_forensic_message(
SAMPLE_FORENSIC_EMAIL,
stats,
message_id="msg-1",
)
assert imported == 0
assert stats["duplicate_forensic_reports"] == 1
assert stats["details"][0]["status"] == "duplicate"
def test_forensic_parse_error_adds_error_detail(self, db_session):
client = _make_client(db=db_session)
stats = {"errors": []}
@@ -795,11 +813,17 @@ class TestFetchReports:
class TestIngestedIdHelpers:
def test_load_empty_json_text_returns_empty_list(self):
assert GmailClient.load_ingested_ids("") == []
def test_load_non_list_json_returns_no_error(self):
# Valid JSON but not a list should gracefully not raise
result = GmailClient.load_ingested_ids('{"key": "value"}')
assert result is not None # no crash
def test_load_invalid_json_returns_empty_list(self):
assert GmailClient.load_ingested_ids("{not-json") == []
def test_dump_preserves_order(self):
ids = ["z", "a", "m"]
dumped = GmailClient.dump_ingested_ids(ids)
+18
View File
@@ -667,6 +667,24 @@ class TestProcessSingleEmail:
assert stats["duplicate_forensic_reports"] == 1
assert stats["details"][1]["status"] == "duplicate"
def test_forensic_save_duplicate_result_adds_detail(self, db_session):
client = self._make_client(db=db_session)
stats = {"forensic_reports_found": 0, "duplicate_forensic_reports": 0, "errors": []}
with (
patch("app.services.imap_client.forensic_report_exists", return_value=False),
patch("app.services.imap_client.save_forensic_report", return_value=(None, False)),
):
imported = client._process_forensic_email(
SAMPLE_FORENSIC_EMAIL,
stats=stats,
message_id="msg-1",
)
assert imported is False
assert stats["duplicate_forensic_reports"] == 1
assert stats["details"][0]["status"] == "duplicate"
def test_forensic_parse_error_adds_error_detail(self, db_session):
client = self._make_client(db=db_session)
stats = {"errors": []}