1
0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-20 10:49:11 +01:00
pywb/webagg/proxyindexsource.py
Ilya Kreymer 107ba9aabc add ProxyLiveIndexSource for proxying upstream conn directly w/o a second index query
liveloader: if 'memento_url' key is set, then memento-datetime header must be present or its an error response
liveindexsource: add option to specify custom live path (eg. prefix for cacheing)
fix test cases changed due to ia (todo: mock up all external data!)
2016-03-08 10:27:13 -08:00

55 lines
1.9 KiB
Python

from pywb.cdx.cdxobject import CDXObject
from pywb.utils.wbexception import NotFoundException
from webagg.indexsource import BaseIndexSource, RemoteIndexSource
from webagg.responseloader import LiveWebLoader
from webagg.utils import ParamFormatter, res_template
from pywb.utils.timeutils import timestamp_now
#=============================================================================
class UpstreamAggIndexSource(RemoteIndexSource):
def __init__(self, base_url):
api_url = base_url + '/index?url={url}'
proxy_url = base_url + '/resource?url={url}&closest={timestamp}'
super(UpstreamAggIndexSource, self).__init__(api_url, proxy_url, 'filename')
def _set_load_url(self, cdx):
super(UpstreamAggIndexSource, self)._set_load_url(cdx)
cdx['offset'] = '0'
cdx.pop('load_url', '')
#=============================================================================
class ProxyMementoIndexSource(BaseIndexSource):
def __init__(self, proxy_url='{url}'):
self.proxy_url = proxy_url
self.loader = LiveWebLoader()
def load_index(self, params):
cdx = CDXObject()
cdx['urlkey'] = params.get('key').decode('utf-8')
closest = params.get('closest')
cdx['timestamp'] = closest if closest else timestamp_now()
cdx['url'] = params['url']
cdx['load_url'] = res_template(self.proxy_url, params)
cdx['memento_url'] = cdx['load_url']
return self._do_load(cdx, params)
def _do_load(self, cdx, params):
result = self.loader.load_resource(cdx, params)
if not result:
raise NotFoundException('Not a memento: ' + cdx['url'])
cdx['_cached_result'] = result
yield cdx
def __str__(self):
return 'proxy'
@staticmethod
def upstream_resource(base_url):
return ProxyMementoIndexSource(base_url + '/resource?url={url}&closest={closest}')