예제 #1
0
    def content_stream(self):
        if not self.http_headers:
            return self.raw_stream

        encoding = self.http_headers.get_header('content-encoding')

        if encoding:
            encoding = encoding.lower()

            if encoding not in BufferedReader.get_supported_decompressors():
                encoding = None

        if self.http_headers.get_header('transfer-encoding') == 'chunked':
            return ChunkedDataReader(self.raw_stream, decomp_type=encoding)
        elif encoding:
            return BufferedReader(self.raw_stream, decomp_type=encoding)
        else:
            return self.raw_stream
예제 #2
0
    def test_unseekable_gz(self):
        """ Test iterator on unseekable 3 record uncompressed gzipped WARC input
        """
        proc = subprocess.Popen(
            ['cat', get_test_file('example-resource.warc.gz')],
            stdout=subprocess.PIPE)

        def raise_tell(x):
            raise Exception()

        # on windows, this tell() exists but doesn't work correctly, so just override (in py3)
        # this is designed to emulated stdin, which does not have a tell(), as expected
        stdout = proc.stdout
        if os.name == 'nt' and hasattr(proc.stdout, 'tell'):
            #can't override tell() in py2
            if sys.version_info < (3, 0):
                stdout = BufferedReader(stdout)
            else:
                stdout.tell = raise_tell

        with closing(ArchiveIterator(stdout)) as a:
            for record in a:
                assert record.rec_type == 'warcinfo'
                assert a.get_record_offset() == 0
                break

            record = next(a)
            assert record.rec_type == 'warcinfo'
            assert a.get_record_offset() == 361

            for record in a:
                assert record.rec_type == 'resource'
                assert a.get_record_offset() == 802
                break

            with pytest.raises(StopIteration):
                record = next(a)

        assert a.record == None
        assert a.reader == None
        assert a.read_to_end() == None

        proc.stdout.close()
        proc.wait()
예제 #3
0
 def read_and_keep(self, size):
     buff = self.content_stream.read(size)
     self._content_stream = BufferedReader(self._content_stream,
                                           starting_data=buff)
     return buff
예제 #4
0
    def render_content(self, wb_url, kwargs, environ):
        wb_url = wb_url.replace('#', '%23')
        wb_url = WbUrl(wb_url)
        is_timegate = self._check_accept_dt(wb_url, environ)

        host_prefix = self.get_host_prefix(environ)
        rel_prefix = self.get_rel_prefix(environ)
        full_prefix = host_prefix + rel_prefix

        is_proxy = ('wsgiprox.proxy_host' in environ)

        response = self.handle_custom_response(environ, wb_url,
                                               full_prefix, host_prefix,
                                               kwargs)

        if response:
            return self.format_response(response, wb_url, full_prefix, is_timegate, is_proxy)

        if is_proxy:
            environ['pywb_proxy_magic'] = environ['wsgiprox.proxy_host']
            urlrewriter = IdentityUrlRewriter(wb_url, '')
            framed_replay = False

        else:
            urlrewriter = UrlRewriter(wb_url,
                                      prefix=full_prefix,
                                      full_prefix=full_prefix,
                                      rel_prefix=rel_prefix)

            framed_replay = self.framed_replay

        url_parts = urlsplit(wb_url.url)
        if not url_parts.path:
            scheme, netloc, path, query, frag = url_parts
            path = '/'
            url = urlunsplit((scheme, netloc, path, query, frag))
            resp = WbResponse.redir_response(urlrewriter.rewrite(url),
                                             '307 Temporary Redirect')

            if self.enable_memento:
                resp.status_headers['Link'] = MementoUtils.make_link(url, 'original')

            return resp

        self.unrewrite_referrer(environ, full_prefix)

        urlkey = canonicalize(wb_url.url)

        if self.use_js_obj_proxy:
            content_rw = self.js_proxy_rw
        else:
            content_rw = self.default_rw

        inputreq = RewriteInputRequest(environ, urlkey, wb_url.url, content_rw)

        inputreq.include_method_query(wb_url.url)

        range_start, range_end, skip_record = self._check_range(inputreq, wb_url)

        setcookie_headers = None
        if self.cookie_tracker:
            cookie_key = self.get_cookie_key(kwargs)
            res = self.cookie_tracker.get_cookie_headers(wb_url.url, urlrewriter, cookie_key)
            inputreq.extra_cookie, setcookie_headers = res

        r = self._do_req(inputreq, wb_url, kwargs, skip_record)

        if r.status_code >= 400:
            error = None
            try:
                error = r.raw.read()
                r.raw.close()
            except:
                pass

            if error:
                error = error.decode('utf-8')
            else:
                error = ''

            details = dict(args=kwargs, error=error)
            raise UpstreamException(r.status_code, url=wb_url.url, details=details)

        stream = BufferedReader(r.raw, block_size=BUFF_SIZE)
        record = self.loader.parse_record_stream(stream,
                                                 ensure_http_headers=True)

        memento_dt = r.headers.get('Memento-Datetime')
        target_uri = r.headers.get('WARC-Target-URI')

        cdx = CDXObject(r.headers.get('Warcserver-Cdx').encode('utf-8'))

        #cdx['urlkey'] = urlkey
        #cdx['timestamp'] = http_date_to_timestamp(memento_dt)
        #cdx['url'] = target_uri

        set_content_loc = False

        # Check if Fuzzy Match
        if target_uri != wb_url.url and cdx.get('is_fuzzy') == '1':
            set_content_loc = True

        # if redir to exact, redir if url or ts are different
        if self.redirect_to_exact:
            if (set_content_loc or
                (wb_url.timestamp != cdx.get('timestamp') and not cdx.get('is_live'))):

                new_url = urlrewriter.get_new_url(url=target_uri,
                                                  timestamp=cdx['timestamp'],
                                                  mod=wb_url.mod)

                resp = WbResponse.redir_response(new_url, '307 Temporary Redirect')
                if self.enable_memento:
                    if is_timegate and not is_proxy:
                        self._add_memento_links(target_uri, full_prefix,
                                                memento_dt, cdx['timestamp'],
                                                resp.status_headers,
                                                is_timegate, is_proxy)

                    else:
                        resp.status_headers['Link'] = MementoUtils.make_link(target_uri, 'original')

                return resp

        self._add_custom_params(cdx, r.headers, kwargs)

        if self._add_range(record, wb_url, range_start, range_end):
            wb_url.mod = 'id_'

        is_ajax = self.is_ajax(environ)

        if is_ajax:
            head_insert_func = None
            urlrewriter.rewrite_opts['is_ajax'] = True
        else:
            top_url = self.get_top_url(full_prefix, wb_url, cdx, kwargs)
            head_insert_func = (self.head_insert_view.
                                    create_insert_func(wb_url,
                                                       full_prefix,
                                                       host_prefix,
                                                       top_url,
                                                       environ,
                                                       framed_replay,
                                                       config=self.config))

        cookie_rewriter = None
        if self.cookie_tracker:
            cookie_rewriter = self.cookie_tracker.get_rewriter(urlrewriter,
                                                               cookie_key)

        urlrewriter.rewrite_opts['ua_string'] = environ.get('HTTP_USER_AGENT')

        result = content_rw(record, urlrewriter, cookie_rewriter, head_insert_func, cdx)

        status_headers, gen, is_rw = result

        if setcookie_headers:
            status_headers.headers.extend(setcookie_headers)

        if ' ' not in status_headers.statusline:
            status_headers.statusline += ' None'

        if not is_ajax and self.enable_memento:
            self._add_memento_links(cdx['url'], full_prefix,
                                    memento_dt, cdx['timestamp'], status_headers,
                                    is_timegate, is_proxy, cdx.get('source-coll'))

            set_content_loc = True

        if set_content_loc and not self.redirect_to_exact:
            status_headers.headers.append(('Content-Location', urlrewriter.get_new_url(timestamp=cdx['timestamp'],
                                                                                       url=cdx['url'])))
        if not is_proxy:
            self.add_csp_header(wb_url, status_headers)

        response = WbResponse(status_headers, gen)

        return response
예제 #5
0
    def render_content(self, wb_url, kwargs, environ):
        wb_url = wb_url.replace('#', '%23')
        wb_url = WbUrl(wb_url)

        history_page = environ.pop('HTTP_X_WOMBAT_HISTORY_PAGE', '')
        if history_page:
            wb_url.url = history_page
            is_ajax = True
        else:
            is_ajax = self.is_ajax(environ)

        is_timegate = self._check_accept_dt(wb_url, environ)

        self.prepare_env(environ)

        host_prefix = environ['pywb.host_prefix']
        rel_prefix = self.get_rel_prefix(environ)
        full_prefix = host_prefix + rel_prefix

        pywb_static_prefix = environ['pywb.static_prefix'] + '/'
        is_proxy = ('wsgiprox.proxy_host' in environ)

        # if OPTIONS in proxy mode, just generate the proxy responss
        if is_proxy and self.is_preflight(environ):
            return WbResponse.options_response(environ)

        if self.use_js_obj_proxy:
            content_rw = self.js_proxy_rw
        else:
            content_rw = self.default_rw

        # no redirects if in proxy
        redirect_to_exact = self.redirect_to_exact and not is_proxy

        # Check Prefer
        pref_mod, pref_applied = self._get_prefer_mod(wb_url, environ,
                                                      content_rw, is_proxy)

        response = None
        keep_frame_response = False

        # prefer overrides custom response?
        if pref_mod is not None:
            # fast-redirect to preferred
            if redirect_to_exact and not is_timegate and pref_mod != wb_url.mod:
                new_url = full_prefix + wb_url.to_str(mod=pref_mod)
                headers = [('Preference-Applied', pref_applied),
                           ('Vary', 'Prefer')]

                return WbResponse.redir_response(new_url,
                                                 '307 Temporary Redirect',
                                                 headers=headers)
            else:
                wb_url.mod = pref_mod
        else:
            if kwargs.get('output'):
                response = self.handle_timemap(wb_url, kwargs, full_prefix)

            elif wb_url.is_query():
                response = self.handle_query(environ, wb_url, kwargs,
                                             full_prefix)

            else:
                response = self.handle_custom_response(environ, wb_url,
                                                       full_prefix,
                                                       host_prefix, kwargs)

                keep_frame_response = (not kwargs.get('no_timegate_check')
                                       and is_timegate
                                       and not is_proxy) or redirect_to_exact

        if response and not keep_frame_response:
            return self.format_response(response, wb_url, full_prefix,
                                        is_timegate, is_proxy)

        if is_proxy:
            environ['pywb_proxy_magic'] = environ['wsgiprox.proxy_host']
            urlrewriter = IdentityUrlRewriter(wb_url, '')
            framed_replay = False

        else:
            urlrewriter = UrlRewriter(wb_url,
                                      prefix=full_prefix,
                                      full_prefix=full_prefix,
                                      rel_prefix=rel_prefix,
                                      pywb_static_prefix=pywb_static_prefix)

            framed_replay = self.framed_replay

        url_parts = urlsplit(wb_url.url)
        if not url_parts.path:
            return self.send_redirect('/', url_parts, urlrewriter)

        self.unrewrite_referrer(environ, full_prefix)

        urlkey = canonicalize(wb_url.url)

        inputreq = RewriteInputRequest(environ, urlkey, wb_url.url, content_rw)

        inputreq.include_method_query(wb_url.url)

        range_start, range_end, skip_record = self._check_range(
            inputreq, wb_url)

        setcookie_headers = None
        cookie_key = None
        if self.cookie_tracker:
            cookie_key = self.get_cookie_key(kwargs)
            if cookie_key:
                res = self.cookie_tracker.get_cookie_headers(
                    wb_url.url, urlrewriter, cookie_key,
                    environ.get('HTTP_COOKIE', ''))
                inputreq.extra_cookie, setcookie_headers = res

        r = self._do_req(inputreq, wb_url, kwargs, skip_record)

        if r.status_code >= 400:
            error = None
            try:
                error = r.raw.read()
            except Exception:
                pass
            finally:
                no_except_close(r.raw)

            if error:
                error = error.decode('utf-8')
            else:
                error = ''

            details = dict(args=kwargs, error=error)
            if r.status_code == 404:
                raise NotFoundException(url=wb_url.url, msg=details)

            else:
                raise UpstreamException(r.status_code,
                                        url=wb_url.url,
                                        details=details)

        cdx = CDXObject(r.headers.get('Warcserver-Cdx').encode('utf-8'))

        cdx_url_parts = urlsplit(cdx['url'])

        if cdx_url_parts.path.endswith(
                '/') and not url_parts.path.endswith('/'):
            # add trailing slash
            new_path = url_parts.path + '/'

            no_except_close(r.raw)

            return self.send_redirect(new_path, url_parts, urlrewriter)

        # only redirect to exact if not live, otherwise set to false
        redirect_to_exact = redirect_to_exact and not cdx.get('is_live')

        # return top-frame timegate response, with timestamp from cdx
        if response and keep_frame_response and (not redirect_to_exact
                                                 or not is_timegate):
            no_except_close(r.raw)
            return self.format_response(response, wb_url, full_prefix,
                                        is_timegate, is_proxy,
                                        cdx['timestamp'])

        stream = BufferedReader(r.raw, block_size=BUFF_SIZE)
        record = self.loader.parse_record_stream(stream,
                                                 ensure_http_headers=True)

        memento_dt = r.headers.get('Memento-Datetime')
        target_uri = r.headers.get('WARC-Target-URI')

        # cdx['urlkey'] = urlkey
        # cdx['timestamp'] = http_date_to_timestamp(memento_dt)
        # cdx['url'] = target_uri

        set_content_loc = False

        # Check if Fuzzy Match
        if target_uri != wb_url.url and cdx.get('is_fuzzy') == '1':
            set_content_loc = True

        # if redirect to exact timestamp (only set if not live)
        if redirect_to_exact:
            if set_content_loc or is_timegate or wb_url.timestamp != cdx.get(
                    'timestamp'):
                new_url = urlrewriter.get_new_url(url=target_uri,
                                                  timestamp=cdx['timestamp'],
                                                  mod=wb_url.mod)

                resp = WbResponse.redir_response(new_url,
                                                 '307 Temporary Redirect')
                if self.enable_memento:
                    if is_timegate and not is_proxy:
                        self._add_memento_links(target_uri,
                                                full_prefix,
                                                memento_dt,
                                                cdx['timestamp'],
                                                resp.status_headers,
                                                is_timegate,
                                                is_proxy,
                                                pref_applied=pref_applied,
                                                mod=pref_mod,
                                                is_memento=False)

                    else:
                        resp.status_headers['Link'] = MementoUtils.make_link(
                            target_uri, 'original')

                return resp

        self._add_custom_params(cdx, r.headers, kwargs, record)

        if self._add_range(record, wb_url, range_start, range_end):
            wb_url.mod = 'id_'

        if is_ajax:
            head_insert_func = None
            urlrewriter.rewrite_opts['is_ajax'] = True
        else:
            top_url = self.get_top_url(full_prefix, wb_url, cdx, kwargs)
            head_insert_func = (self.head_insert_view.create_insert_func(
                wb_url,
                full_prefix,
                host_prefix,
                top_url,
                environ,
                framed_replay,
                coll=kwargs.get('coll', ''),
                replay_mod=self.replay_mod,
                metadata=kwargs.get('metadata', {}),
                config=self.config))

        cookie_rewriter = None
        if self.cookie_tracker and cookie_key:
            # skip add cookie if service worker is not 200
            # it seems cookie headers from service workers are not applied, so don't update in cache
            if wb_url.mod == 'sw_':
                cookie_key = None

            cookie_rewriter = self.cookie_tracker.get_rewriter(
                urlrewriter, cookie_key)

        urlrewriter.rewrite_opts['ua_string'] = environ.get('HTTP_USER_AGENT')

        result = content_rw(record, urlrewriter, cookie_rewriter,
                            head_insert_func, cdx, environ)

        status_headers, gen, is_rw = result

        if history_page:
            title = DefaultRewriter._extract_title(gen)
            if not title:
                title = unquote(environ.get('HTTP_X_WOMBAT_HISTORY_TITLE', ''))

            if not title:
                title = history_page

            self._add_history_page(cdx, kwargs, title)
            return WbResponse.json_response({'title': title})

        if setcookie_headers:
            status_headers.headers.extend(setcookie_headers)

        if ' ' not in status_headers.statusline:
            status_headers.statusline += ' None'

        if not is_ajax and self.enable_memento:
            self._add_memento_links(cdx['url'],
                                    full_prefix,
                                    memento_dt,
                                    cdx['timestamp'],
                                    status_headers,
                                    is_timegate,
                                    is_proxy,
                                    cdx.get('source-coll'),
                                    mod=pref_mod,
                                    pref_applied=pref_applied)

            set_content_loc = True

        if set_content_loc and not redirect_to_exact and not is_proxy:
            status_headers.headers.append(
                ('Content-Location',
                 urlrewriter.get_new_url(timestamp=cdx['timestamp'],
                                         url=cdx['url'])))

        if not is_proxy:
            self.add_csp_header(wb_url, status_headers)

        response = WbResponse(status_headers, gen)

        if is_proxy and environ.get('HTTP_ORIGIN'):
            response.add_access_control_headers(environ)

        if r.status_code == 200 and kwargs.get(
                'cache') == 'always' and environ.get('HTTP_REFERER'):
            response.status_headers[
                'Cache-Control'] = 'public, max-age=31536000, immutable'

        return response
예제 #6
0
    def render_content(self, wb_url, kwargs, environ):
        wb_url = wb_url.replace('#', '%23')
        wb_url = WbUrl(wb_url)

        proto = environ.get('HTTP_X_FORWARDED_PROTO', self.force_scheme)

        if proto:
            environ['wsgi.url_scheme'] = proto

        history_page = environ.pop('HTTP_X_WOMBAT_HISTORY_PAGE', '')
        if history_page:
            wb_url.url = history_page
            is_ajax = True
        else:
            is_ajax = self.is_ajax(environ)

        is_timegate = self._check_accept_dt(wb_url, environ)

        host_prefix = self.get_host_prefix(environ)
        rel_prefix = self.get_rel_prefix(environ)
        full_prefix = host_prefix + rel_prefix
        environ['pywb.host_prefix'] = host_prefix
        pywb_static_prefix = host_prefix + environ.get(
            'pywb.app_prefix', '') + environ.get('pywb.static_prefix',
                                                 '/static/')
        is_proxy = ('wsgiprox.proxy_host' in environ)

        response = self.handle_custom_response(environ, wb_url, full_prefix,
                                               host_prefix, kwargs)

        if response:
            return self.format_response(response, wb_url, full_prefix,
                                        is_timegate, is_proxy)

        if is_proxy:
            environ['pywb_proxy_magic'] = environ['wsgiprox.proxy_host']
            urlrewriter = IdentityUrlRewriter(wb_url, '')
            framed_replay = False

        else:
            urlrewriter = UrlRewriter(wb_url,
                                      prefix=full_prefix,
                                      full_prefix=full_prefix,
                                      rel_prefix=rel_prefix,
                                      pywb_static_prefix=pywb_static_prefix)

            framed_replay = self.framed_replay

        url_parts = urlsplit(wb_url.url)
        if not url_parts.path:
            return self.send_redirect('/', url_parts, urlrewriter)

        self.unrewrite_referrer(environ, full_prefix)

        urlkey = canonicalize(wb_url.url)

        if self.use_js_obj_proxy:
            content_rw = self.js_proxy_rw
        else:
            content_rw = self.default_rw

        inputreq = RewriteInputRequest(environ, urlkey, wb_url.url, content_rw)

        inputreq.include_method_query(wb_url.url)

        range_start, range_end, skip_record = self._check_range(
            inputreq, wb_url)

        setcookie_headers = None
        cookie_key = None
        if self.cookie_tracker:
            cookie_key = self.get_cookie_key(kwargs)
            if cookie_key:
                res = self.cookie_tracker.get_cookie_headers(
                    wb_url.url, urlrewriter, cookie_key,
                    environ.get('HTTP_COOKIE', ''))
                inputreq.extra_cookie, setcookie_headers = res

        r = self._do_req(inputreq, wb_url, kwargs, skip_record)

        if r.status_code >= 400:
            error = None
            try:
                error = r.raw.read()
            except Exception:
                pass
            finally:
                no_except_close(r.raw)

            if error:
                error = error.decode('utf-8')
            else:
                error = ''

            details = dict(args=kwargs, error=error)
            raise UpstreamException(r.status_code,
                                    url=wb_url.url,
                                    details=details)

        cdx = CDXObject(r.headers.get('Warcserver-Cdx').encode('utf-8'))

        cdx_url_parts = urlsplit(cdx['url'])

        if cdx_url_parts.path.endswith(
                '/') and not url_parts.path.endswith('/'):
            # add trailing slash
            new_path = url_parts.path + '/'

            no_except_close(r.raw)

            return self.send_redirect(new_path, url_parts, urlrewriter)

        stream = BufferedReader(r.raw, block_size=BUFF_SIZE)
        record = self.loader.parse_record_stream(stream,
                                                 ensure_http_headers=True)

        memento_dt = r.headers.get('Memento-Datetime')
        target_uri = r.headers.get('WARC-Target-URI')

        # cdx['urlkey'] = urlkey
        # cdx['timestamp'] = http_date_to_timestamp(memento_dt)
        # cdx['url'] = target_uri

        set_content_loc = False

        # Check if Fuzzy Match
        if target_uri != wb_url.url and cdx.get('is_fuzzy') == '1':
            set_content_loc = True

        # if redir to exact, redir if url or ts are different
        if self.redirect_to_exact:
            if (set_content_loc or (wb_url.timestamp != cdx.get('timestamp')
                                    and not cdx.get('is_live'))):

                new_url = urlrewriter.get_new_url(url=target_uri,
                                                  timestamp=cdx['timestamp'],
                                                  mod=wb_url.mod)

                resp = WbResponse.redir_response(new_url,
                                                 '307 Temporary Redirect')
                if self.enable_memento:
                    if is_timegate and not is_proxy:
                        self._add_memento_links(target_uri, full_prefix,
                                                memento_dt, cdx['timestamp'],
                                                resp.status_headers,
                                                is_timegate, is_proxy)

                    else:
                        resp.status_headers['Link'] = MementoUtils.make_link(
                            target_uri, 'original')

                return resp

        self._add_custom_params(cdx, r.headers, kwargs, record)

        if self._add_range(record, wb_url, range_start, range_end):
            wb_url.mod = 'id_'

        if is_ajax:
            head_insert_func = None
            urlrewriter.rewrite_opts['is_ajax'] = True
        else:
            top_url = self.get_top_url(full_prefix, wb_url, cdx, kwargs)
            head_insert_func = (self.head_insert_view.create_insert_func(
                wb_url,
                full_prefix,
                host_prefix,
                top_url,
                environ,
                framed_replay,
                coll=kwargs.get('coll', ''),
                replay_mod=self.replay_mod,
                config=self.config))

        cookie_rewriter = None
        if self.cookie_tracker and cookie_key:
            # skip add cookie if service worker is not 200
            # it seems cookie headers from service workers are not applied, so don't update in cache
            if wb_url.mod == 'sw_':
                cookie_key = None

            cookie_rewriter = self.cookie_tracker.get_rewriter(
                urlrewriter, cookie_key)

        urlrewriter.rewrite_opts['ua_string'] = environ.get('HTTP_USER_AGENT')

        result = content_rw(record, urlrewriter, cookie_rewriter,
                            head_insert_func, cdx, environ)

        status_headers, gen, is_rw = result

        if history_page:
            title = DefaultRewriter._extract_title(gen)
            if not title:
                title = unquote(environ.get('HTTP_X_WOMBAT_HISTORY_TITLE', ''))

            if not title:
                title = history_page

            self._add_history_page(cdx, kwargs, title)
            return WbResponse.json_response({'title': title})

        if setcookie_headers:
            status_headers.headers.extend(setcookie_headers)

        if ' ' not in status_headers.statusline:
            status_headers.statusline += ' None'

        if not is_ajax and self.enable_memento:
            self._add_memento_links(cdx['url'], full_prefix, memento_dt,
                                    cdx['timestamp'], status_headers,
                                    is_timegate, is_proxy,
                                    cdx.get('source-coll'))

            set_content_loc = True

        if set_content_loc and not self.redirect_to_exact:
            status_headers.headers.append(
                ('Content-Location',
                 urlrewriter.get_new_url(timestamp=cdx['timestamp'],
                                         url=cdx['url'])))
        if not is_proxy:
            self.add_csp_header(wb_url, status_headers)

        response = WbResponse(status_headers, gen)

        return response