def test_backend(backend):
# Graph
graph = graphs.Manager()
graph.add_site_list(SITE_LIST)
# Frontier
settings = Settings()
settings.BACKEND = backend
settings.LOGGING_MANAGER_ENABLED = True
settings.LOGGING_BACKEND_ENABLED = True
settings.LOGGING_DEBUGGING_ENABLED = False
frontier = FrontierManager.from_settings(settings)
print '-'*80
print frontier.backend.name
print '-'*80
# Tester
tester = FrontierTester(frontier, graph)
tester.run()
# Show crawling sequence
for page in tester.sequence:
print page.url
def test_logic(backend):
# Graph
graph = graphs.Manager('sqlite:///data/graph.db')
# Frontier
settings = Settings()
settings.BACKEND = backend
settings.LOGGING_MANAGER_ENABLED = True
settings.LOGGING_BACKEND_ENABLED = True
settings.LOGGING_DEBUGGING_ENABLED = False
settings.TEST_MODE = True
frontier = FrontierManager.from_settings(settings)
# Tester
tester = FrontierTester(frontier, graph)
tester.run(add_all_pages=True)
# Show crawling sequence
print '-'*80
print frontier.backend.name
print '-'*80
for page in tester.sequence:
print page.url
def get_frontier(self):
"""
Returns frontierManager object
"""
return FrontierManager.from_settings(self.get_settings())
"""
from frontera import FrontierManager, FrontierTester, Settings, graphs
SETTINGS = Settings()
SETTINGS.BACKEND = 'frontera.contrib.backends.memory_heapq.FIFO'
SETTINGS.LOGGING_MANAGER_ENABLED = True
SETTINGS.LOGGING_BACKEND_ENABLED = True
SETTINGS.LOGGING_DEBUGGING_ENABLED = False
if __name__ == '__main__':
# Graph
graph = graphs.Manager('sqlite:///recordings/scrapinghub.com.db')
# Frontier
frontier = FrontierManager.from_settings(SETTINGS)
# Tester
tester = FrontierTester(frontier, graph)
# Run test
tester.run()
# Show frontier pages
print '-'*80
print ' Frontier pages'
print '-'*80
for page in frontier.backend.pages.values():
print page.url, page.depth, page.state
# Show crawling sequence
"""
Frontier tester usage example
"""
from frontera import FrontierManager, FrontierTester, Settings, graphs
if __name__ == '__main__':
# Graph
graph = graphs.Manager('sqlite:///data/graph.db')
# Frontier
settings = Settings()
settings.TEST_MODE = True
settings.LOGGING_MANAGER_ENABLED = True
settings.LOGGING_BACKEND_ENABLED = True
settings.LOGGING_DEBUGGING_ENABLED = False
frontier = FrontierManager.from_settings(settings)
# Tester
tester = FrontierTester(frontier, graph)
# Run test
tester.run()
# Show crawling sequence
for page in tester.sequence:
print page.url
"""
Frontier from parameters example
"""
from frontera import FrontierManager, graphs, Request, Response
if __name__ == '__main__':
# Create graph
graph = graphs.Manager('sqlite:///data/graph.db')
# Create frontier
frontier = FrontierManager(
request_model='frontera.core.models.Request',
response_model='frontera.core.models.Response',
backend='frontera.contrib.backends.memory.FIFO',
logger='frontera.logger.FrontierLogger',
event_log_manager='frontera.logger.events.EventLogManager',
middlewares=[
'frontera.contrib.middlewares.domain.DomainMiddleware',
'frontera.contrib.middlewares.fingerprint.UrlFingerprintMiddleware',
'frontera.contrib.middlewares.fingerprint.DomainFingerprintMiddleware',
],
test_mode=True)
# Add seeds
frontier.add_seeds([Request(seed.url) for seed in graph.seeds])
# Get next requests
next_requests = frontier.get_next_requests()
# Crawl pages
for request in next_requests: