backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-24 06:59:52 +01:00
Code Issues Projects Releases Activity

frontendapp: encapsulate, don't extend rewriterapp

Browse Source 
rewriterapp: add 'Content-Location' if fuzzy match, or if using memento
tests: fix test to check for Content-Location for fuzzy match instead of redirect
This commit is contained in:
Ilya Kreymer 2017-04-21 15:37:21 -07:00
parent 4b055c9394
commit e060ea7b56
3 changed files with 27 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
pywb/urlrewrite/frontendapp.py
View File

@ -27,13 +27,13 @@ class NewWbRequest(object):
# ============================================================================ # ============================================================================
class FrontEndApp(RewriterApp): class FrontEndApp(object):
def __init__(self, config_file='./config.yaml', custom_config=None): def __init__(self, config_file='./config.yaml', custom_config=None):
self.debug = True self.debug = True
self.webagg = AutoConfigApp(config_file=config_file, self.webagg = AutoConfigApp(config_file=config_file,
custom_config=custom_config) custom_config=custom_config)
super(FrontEndApp, self).__init__(True, config=self.webagg.config) self.rewriterapp = RewriterApp(True, config=self.webagg.config)
self.webagg_server = GeventServer(self.webagg, port=0) self.webagg_server = GeventServer(self.webagg, port=0)
@ -46,7 +46,7 @@ class FrontEndApp(RewriterApp):
self.url_map.add(Rule('/collinfo.json', endpoint=self.serve_listing)) self.url_map.add(Rule('/collinfo.json', endpoint=self.serve_listing))
self.url_map.add(Rule('/', endpoint=self.serve_home)) self.url_map.add(Rule('/', endpoint=self.serve_home))
self.paths = self.get_upstream_paths(self.webagg_server.port) self.rewriterapp.paths = self.get_upstream_paths(self.webagg_server.port)
def get_upstream_paths(self, port): def get_upstream_paths(self, port):
return {'replay-dyn': 'http://localhost:%s/_/resource/postreq?param.coll={coll}' % port, return {'replay-dyn': 'http://localhost:%s/_/resource/postreq?param.coll={coll}' % port,
@ -54,7 +54,7 @@ class FrontEndApp(RewriterApp):
} }
def serve_home(self, environ): def serve_home(self, environ):
home_view = BaseInsertView(self.jinja_env, 'new_index.html') home_view = BaseInsertView(self.rewriterapp.jinja_env, 'new_index.html')
routes = self.webagg.list_fixed_routes() + self.webagg.list_dynamic_routes() routes = self.webagg.list_fixed_routes() + self.webagg.list_dynamic_routes()
content = home_view.render_to_string(environ, routes=routes) content = home_view.render_to_string(environ, routes=routes)
@ -64,14 +64,14 @@ class FrontEndApp(RewriterApp):
try: try:
return self.static_handler(NewWbRequest(environ, filepath, '')) return self.static_handler(NewWbRequest(environ, filepath, ''))
except: except:
raise NotFound(response=self._error_response(environ, 'Static File Not Found: {0}'.format(filepath))) self.raise_not_found(environ, 'Static File Not Found: {0}'.format(filepath))
def serve_coll_page(self, environ, coll): def serve_coll_page(self, environ, coll):
if not self.is_valid_coll(coll): if not self.is_valid_coll(coll):
raise NotFound(response=self._error_response(environ, 'No handler for "/{0}"'.format(coll))) self.raise_not_found(environ, 'No handler for "/{0}"'.format(coll))
wbrequest = NewWbRequest(environ, '', '/') wbrequest = NewWbRequest(environ, '', '/')
view = BaseInsertView(self.jinja_env, 'search.html') view = BaseInsertView(self.rewriterapp.jinja_env, 'search.html')
content = view.render_to_string(environ, wbrequest=wbrequest) content = view.render_to_string(environ, wbrequest=wbrequest)
return WbResponse.text_response(content, content_type='text/html; charset="utf-8"') return WbResponse.text_response(content, content_type='text/html; charset="utf-8"')
@ -87,12 +87,15 @@ class FrontEndApp(RewriterApp):
return (coll in self.webagg.list_fixed_routes() or return (coll in self.webagg.list_fixed_routes() or
coll in self.webagg.list_dynamic_routes()) coll in self.webagg.list_dynamic_routes())
def raise_not_found(self, environ, msg):
raise NotFound(response=self.rewriterapp._error_response(environ, msg))
def serve_content(self, environ, coll='', url=''): def serve_content(self, environ, coll='', url=''):
if not self.is_valid_coll(coll): if not self.is_valid_coll(coll):
raise NotFound(response=self._error_response(environ, 'No handler for "/{0}"'.format(coll))) self.raise_not_found(environ, 'No handler for "/{0}"'.format(coll))
pop_path_info(environ) pop_path_info(environ)
wb_url = self.get_wburl(environ) wb_url = self.rewriterapp.get_wburl(environ)
kwargs = {'coll': coll} kwargs = {'coll': coll}
@ -102,9 +105,9 @@ class FrontEndApp(RewriterApp):
kwargs['type'] = 'replay-dyn' kwargs['type'] = 'replay-dyn'
try: try:
response = self.render_content(wb_url, kwargs, environ) response = self.rewriterapp.render_content(wb_url, kwargs, environ)
except UpstreamException as ue: except UpstreamException as ue:
response = self.handle_error(environ, ue) response = self.rewriterapp.handle_error(environ, ue)
raise HTTPException(response=response) raise HTTPException(response=response)
return response return response
@ -156,7 +159,7 @@ class FrontEndApp(RewriterApp):
if self.debug: if self.debug:
traceback.print_exc() traceback.print_exc()
response = self._error_response(environ, 'Internal Error: ' + str(e), '500 Server Error') response = self.rewriterapp._error_response(environ, 'Internal Error: ' + str(e), '500 Server Error')
return response(environ, start_response) return response(environ, start_response)
@classmethod @classmethod

11
pywb/urlrewrite/rewriterapp.py
View File

@ -216,8 +216,12 @@ class RewriterApp(object):
cdx['timestamp'] = http_date_to_timestamp(memento_dt) cdx['timestamp'] = http_date_to_timestamp(memento_dt)
cdx['url'] = target_uri cdx['url'] = target_uri
# Disable Fuzzy Match Redir set_content_loc = False
#if target_uri != wb_url.url and r.headers.get('WebAgg-Fuzzy-Match') == '1':
# Check if Fuzzy Match
if target_uri != wb_url.url and r.headers.get('WebAgg-Fuzzy-Match') == '1':
set_content_loc = True
# return WbResponse.redir_response(urlrewriter.rewrite(target_uri), # return WbResponse.redir_response(urlrewriter.rewrite(target_uri),
# '307 Temporary Redirect') # '307 Temporary Redirect')
@ -272,6 +276,9 @@ class RewriterApp(object):
if not is_ajax and self.enable_memento: if not is_ajax and self.enable_memento:
self._add_memento_links(urlrewriter, full_prefix, memento_dt, status_headers) self._add_memento_links(urlrewriter, full_prefix, memento_dt, status_headers)
set_content_loc = True
if set_content_loc:
status_headers.headers.append(('Content-Location', urlrewriter.get_new_url(timestamp=cdx['timestamp'], status_headers.headers.append(('Content-Location', urlrewriter.get_new_url(timestamp=cdx['timestamp'],
url=cdx['url']))) url=cdx['url'])))
#gen = buffer_iter(status_headers, gen) #gen = buffer_iter(status_headers, gen)

5
tests/test_integration.py
View File

@ -106,8 +106,9 @@ class TestWbIntegration(BaseConfigTest):
def test_replay_fuzzy_1(self): def test_replay_fuzzy_1(self):
resp = self.testapp.get('/pywb/20140127171238mp_/http://www.iana.org/?_=123') resp = self.testapp.get('/pywb/20140127171238mp_/http://www.iana.org/?_=123')
assert resp.status_int == 307 assert resp.status_int == 200
assert resp.headers['Location'].endswith('/pywb/20140127171238mp_/http://www.iana.org/') assert resp.headers['Content-Location'].endswith('/pywb/20140126200624mp_/http://www.iana.org/')
#assert resp.headers['Location'].endswith('/pywb/20140127171238mp_/http://www.iana.org/')
def test_replay_no_fuzzy_match(self): def test_replay_no_fuzzy_match(self):
resp = self.testapp.get('/pywb/20140127171238mp_/http://www.iana.org/?foo=bar', status=404) resp = self.testapp.get('/pywb/20140127171238mp_/http://www.iana.org/?foo=bar', status=404)