mirror of
https://github.com/ArchiveBox/ArchiveBox.git
synced 2026-01-03 17:35:45 +10:00
cfa5edb160
Real integration tests using Chrome sessions with example.com: - accessibility: Tests page outline and accessibility tree extraction - parse_dom_outlinks: Tests link extraction and categorization - consolelog: Tests console output capture
124 lines
4.3 KiB
Python
124 lines
4.3 KiB
Python
"""
|
|
Tests for the consolelog plugin.
|
|
|
|
Tests the real consolelog hook with an actual URL to verify
|
|
console output capture.
|
|
"""
|
|
|
|
import json
|
|
import shutil
|
|
import subprocess
|
|
import sys
|
|
import tempfile
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
from django.test import TestCase
|
|
|
|
# Import chrome test helpers
|
|
sys.path.insert(0, str(Path(__file__).parent.parent.parent / 'chrome' / 'tests'))
|
|
from chrome_test_helpers import (
|
|
chrome_session,
|
|
get_test_env,
|
|
get_plugin_dir,
|
|
get_hook_script,
|
|
)
|
|
|
|
|
|
def chrome_available() -> bool:
|
|
"""Check if Chrome/Chromium is available."""
|
|
for name in ['chromium', 'chromium-browser', 'google-chrome', 'chrome']:
|
|
if shutil.which(name):
|
|
return True
|
|
return False
|
|
|
|
|
|
# Get the path to the consolelog hook
|
|
PLUGIN_DIR = get_plugin_dir(__file__)
|
|
CONSOLELOG_HOOK = get_hook_script(PLUGIN_DIR, 'on_Snapshot__*_consolelog.*')
|
|
|
|
|
|
class TestConsolelogPlugin(TestCase):
|
|
"""Test the consolelog plugin."""
|
|
|
|
def test_consolelog_hook_exists(self):
|
|
"""Consolelog hook script should exist."""
|
|
self.assertIsNotNone(CONSOLELOG_HOOK, "Consolelog hook not found in plugin directory")
|
|
self.assertTrue(CONSOLELOG_HOOK.exists(), f"Hook not found: {CONSOLELOG_HOOK}")
|
|
|
|
|
|
@pytest.mark.skipif(not chrome_available(), reason="Chrome not installed")
|
|
class TestConsolelogWithChrome(TestCase):
|
|
"""Integration tests for consolelog plugin with Chrome."""
|
|
|
|
def setUp(self):
|
|
"""Set up test environment."""
|
|
self.temp_dir = Path(tempfile.mkdtemp())
|
|
|
|
def tearDown(self):
|
|
"""Clean up."""
|
|
shutil.rmtree(self.temp_dir, ignore_errors=True)
|
|
|
|
def test_consolelog_captures_output(self):
|
|
"""Consolelog hook should capture console output from page."""
|
|
test_url = 'https://example.com'
|
|
snapshot_id = 'test-consolelog-snapshot'
|
|
|
|
try:
|
|
with chrome_session(
|
|
self.temp_dir,
|
|
crawl_id='test-consolelog-crawl',
|
|
snapshot_id=snapshot_id,
|
|
test_url=test_url,
|
|
navigate=True,
|
|
timeout=30,
|
|
) as (chrome_process, chrome_pid, snapshot_chrome_dir):
|
|
# Get environment and run the consolelog hook
|
|
env = get_test_env()
|
|
env['CHROME_HEADLESS'] = 'true'
|
|
|
|
# Run consolelog hook with the active Chrome session
|
|
result = subprocess.run(
|
|
['node', str(CONSOLELOG_HOOK), f'--url={test_url}', f'--snapshot-id={snapshot_id}'],
|
|
cwd=str(snapshot_chrome_dir),
|
|
capture_output=True,
|
|
text=True,
|
|
timeout=120, # Longer timeout as it waits for navigation
|
|
env=env
|
|
)
|
|
|
|
# Check for output file
|
|
console_output = snapshot_chrome_dir / 'console.jsonl'
|
|
|
|
# Verify hook ran (may succeed or timeout waiting for navigation)
|
|
# The hook is designed to wait for page_loaded.txt from chrome_navigate
|
|
# In test mode, that file may not exist, so hook may timeout
|
|
# But it should still create the console.jsonl file
|
|
|
|
# At minimum, verify no crash
|
|
self.assertNotIn('Traceback', result.stderr)
|
|
|
|
# If output file exists, verify it's valid JSONL
|
|
if console_output.exists():
|
|
with open(console_output) as f:
|
|
content = f.read().strip()
|
|
if content:
|
|
for line in content.split('\n'):
|
|
if line.strip():
|
|
try:
|
|
record = json.loads(line)
|
|
# Verify structure
|
|
self.assertIn('timestamp', record)
|
|
self.assertIn('type', record)
|
|
except json.JSONDecodeError:
|
|
pass # Some lines may be incomplete
|
|
|
|
except RuntimeError as e:
|
|
if 'Chrome' in str(e) or 'CDP' in str(e):
|
|
self.skipTest(f"Chrome session setup failed: {e}")
|
|
raise
|
|
|
|
|
|
if __name__ == '__main__':
|
|
pytest.main([__file__, '-v'])
|