Better handle non-ascii urls for crawl log hop info

setup.py

@@ -35,6 +35,7 @@ deps = [
     'idna==2.10',
     'PyYAML>=5.1',
     'cachetools',
+    'rfc3986>=2.0.0',
 ]
 try:
     import concurrent.futures

tests/Dockerfile

@@ -19,7 +19,7 @@
 # USA.
 #
 
-FROM phusion/baseimage
+FROM ubuntu:focal-20220404
 MAINTAINER Noah Levitt <nlevitt@archive.org>
 
 # see https://github.com/stuartpb/rethinkdb-dockerfiles/blob/master/trusty/2.1.3/Dockerfile
@@ -28,10 +28,11 @@ MAINTAINER Noah Levitt <nlevitt@archive.org>
 ENV LANG=C.UTF-8
 
 RUN apt-get update && apt-get --auto-remove -y dist-upgrade
+RUN apt-get install -y ca-certificates curl gnupg wget
 
 # Add the RethinkDB repository and public key
-RUN curl -s https://download.rethinkdb.com/apt/pubkey.gpg | apt-key add - \
+RUN curl -Ss https://download.rethinkdb.com/repository/raw/pubkey.gpg | apt-key add - 
-    && echo "deb http://download.rethinkdb.com/apt xenial main" > /etc/apt/sources.list.d/rethinkdb.list \
+RUN echo "deb https://download.rethinkdb.com/repository/ubuntu-focal focal main" > /etc/apt/sources.list.d/rethinkdb.list \
     && apt-get update && apt-get -y install rethinkdb
 
 RUN mkdir -vp /etc/service/rethinkdb \
@@ -57,25 +58,54 @@ RUN mkdir -vp /etc/service/tor \
     && chmod a+x /etc/service/tor/run
 
 # hadoop hdfs for trough
-RUN curl -s https://archive.cloudera.com/cdh5/ubuntu/xenial/amd64/cdh/archive.key | apt-key add - \
-    && . /etc/lsb-release \
-    && echo "deb [arch=amd64] http://archive.cloudera.com/cdh5/ubuntu/$DISTRIB_CODENAME/amd64/cdh $DISTRIB_CODENAME-cdh5 contrib" >> /etc/apt/sources.list.d/cloudera.list
 
-RUN apt-get update
+ARG DEBIAN_FRONTEND=noninteractive
-RUN apt-get install -y openjdk-8-jdk hadoop-conf-pseudo
+ENV TZ=Etc/UTC
+RUN apt-get install -y openjdk-8-jdk openssh-server 
 
-RUN su hdfs -c 'hdfs namenode -format'
+# set java home
+ENV JAVA_HOME=/usr/lib/jvm/java-8-openjdk-amd64
 
-RUN mv -v /etc/hadoop/conf/core-site.xml /etc/hadoop/conf/core-site.xml.orig \
+# setup ssh with no passphrase
-    && cat /etc/hadoop/conf/core-site.xml.orig | sed 's,localhost:8020,0.0.0.0:8020,' > /etc/hadoop/conf/core-site.xml
+RUN ssh-keygen -t rsa -f $HOME/.ssh/id_rsa -P "" \
+    && cat $HOME/.ssh/id_rsa.pub >> $HOME/.ssh/authorized_keys
 
-RUN mv -v /etc/hadoop/conf/hdfs-site.xml /etc/hadoop/conf/hdfs-site.xml.orig \
+RUN wget -O /hadoop-2.7.3.tar.gz -q https://archive.apache.org/dist/hadoop/common/hadoop-2.7.3/hadoop-2.7.3.tar.gz \
-    && cat /etc/hadoop/conf/hdfs-site.xml.orig | sed 's,^</configuration>$,  <property>\n    <name>dfs.permissions.enabled</name>\n    <value>false</value>\n  </property>\n</configuration>,' > /etc/hadoop/conf/hdfs-site.xml
+        && tar xfz hadoop-2.7.3.tar.gz \
+        && mv /hadoop-2.7.3 /usr/local/hadoop \
+        && rm /hadoop-2.7.3.tar.gz
+	
+# hadoop environment variables
+ENV HADOOP_HOME=/usr/local/hadoop
+ENV PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
 
-RUN echo '#!/bin/bash\nservice hadoop-hdfs-namenode start\nservice hadoop-hdfs-datanode start' > /etc/my_init.d/50_start_hdfs.sh \
+# hadoop-store
-    && chmod a+x /etc/my_init.d/50_start_hdfs.sh
+RUN mkdir -p $HADOOP_HOME/hdfs/namenode \
+        && mkdir -p $HADOOP_HOME/hdfs/datanode
 
-RUN apt-get install -y libsqlite3-dev
+# Temporary files: http://refspecs.linuxfoundation.org/FHS_3.0/fhs/ch03s18.html
+COPY config/ /tmp/
+RUN mv /tmp/ssh_config $HOME/.ssh/config \
+    && mv /tmp/hadoop-env.sh $HADOOP_HOME/etc/hadoop/hadoop-env.sh \
+    && mv /tmp/core-site.xml $HADOOP_HOME/etc/hadoop/core-site.xml \
+    && mv /tmp/hdfs-site.xml $HADOOP_HOME/etc/hadoop/hdfs-site.xml \
+    && mv /tmp/mapred-site.xml $HADOOP_HOME/etc/hadoop/mapred-site.xml.template \
+    && cp $HADOOP_HOME/etc/hadoop/mapred-site.xml.template $HADOOP_HOME/etc/hadoop/mapred-site.xml \
+    && mv /tmp/yarn-site.xml $HADOOP_HOME/etc/hadoop/yarn-site.xml
+
+# Add startup script
+ADD config/hadoop-services.sh $HADOOP_HOME/hadoop-services.sh
+
+# set permissions
+RUN chmod 744 -R $HADOOP_HOME
+
+# format namenode
+RUN $HADOOP_HOME/bin/hdfs namenode -format
+
+# run hadoop services
+#ENTRYPOINT $HADOOP_HOME/hadoop-services.sh; bash
+
+RUN apt-get install -y libsqlite3-dev build-essential
 
 # trough itself
 RUN virtualenv -p python3 /opt/trough-ve3 \
@@ -107,3 +137,4 @@ RUN mkdir -vp /etc/service/trough-segment-manager-server \
     && echo '#!/bin/bash\nvenv=/opt/trough-ve3\nsource $venv/bin/activate\nsleep 5\npython -c $"import doublethink ; from trough.settings import settings ; rr = doublethink.Rethinker(settings[\"RETHINKDB_HOSTS\"]) ; rr.db(\"trough_configuration\").wait().run()"\nexec uwsgi --venv=$venv --http :6111 --master --processes=2 --harakiri=7200 --http-timeout=7200 --max-requests=50000 --vacuum --die-on-term --mount /=trough.wsgi.segment_manager:server >>/tmp/trough-segment-manager-server.out 2>&1' > /etc/service/trough-segment-manager-server/run \
     && chmod a+x /etc/service/trough-segment-manager-server/run
 
+RUN apt-get install -y daemontools daemontools-run

tests/run-tests.sh

@@ -31,15 +31,18 @@ script_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 
 docker build -t internetarchive/warcprox-tests $script_dir
 
-docker run --rm --volume="$script_dir/..:/warcprox" internetarchive/warcprox-tests /sbin/my_init -- \
+docker run --rm --volume="$script_dir/..:/warcprox" internetarchive/warcprox-tests \
     bash -x -c "cd /tmp && git clone /warcprox && cd /tmp/warcprox \
         && (cd /warcprox && git diff HEAD) | patch -p1 \
         && virtualenv -p python3 /tmp/venv \
         && source /tmp/venv/bin/activate \
-        && pip --log-file /tmp/pip.log install . pytest mock requests warcio \
+        && pip --log-file /tmp/pip.log install . pytest mock requests warcio trough \
-        && py.test -v tests \
+        && py.test -v tests; \
-        && py.test -v --rethinkdb-dedup-url=rethinkdb://localhost/test1/dedup tests \
+	svscan /etc/service & \
+	sleep 10; \
+        py.test -v --rethinkdb-dedup-url=rethinkdb://localhost/test1/dedup tests \
         && py.test -v --rethinkdb-big-table-url=rethinkdb://localhost/test2/captures tests \
+	&& /usr/local/hadoop/hadoop-services.sh \
         && py.test -v --rethinkdb-trough-db-url=rethinkdb://localhost/trough_configuration tests \
         "
 

tests/test_warcprox.py

@@ -2106,6 +2106,40 @@ def test_crawl_log(warcprox_, http_daemon, archiving_proxies):
     extra_info = json.loads(fields[12].decode('utf-8'))
     assert set(extra_info.keys()) == {'exception'}
 
+    #Verify non-ascii urls are encoded properly
+    url = 'http://localhost:%s/b/¶-non-ascii' % http_daemon.server_port
+    headers = {
+        "Warcprox-Meta": json.dumps({"warc-prefix":"test_crawl_log_8",
+                                     "metadata":{'seed': 'http://example.com/¶-non-ascii', 'hop_path': 'L', 'brozzled_url': 'http://localhost:%s/b/¶-non-ascii' % http_daemon.server_port, 'hop_via_url': 'http://чунджа.kz/b/¶-non-ascii'}}),
+    }
+    response = requests.get(url, proxies=archiving_proxies, headers=headers)
+    assert response.status_code == 200
+
+    # wait for postfetch chain
+    wait(lambda: warcprox_.proxy.running_stats.urls - urls_before == 9)
+
+    file = os.path.join(
+        warcprox_.options.crawl_log_dir,
+        'test_crawl_log_8-%s-%s.log' % (hostname, port))
+
+    assert os.path.exists(file)
+    crawl_log_8 = open(file, 'rb').read()
+    assert re.match(br'\A2[^\n]+\n\Z', crawl_log_8)
+    assert crawl_log_8[24:31] == b'   200 '
+    assert crawl_log_8[31:42] == b'       154 '
+    fields = crawl_log_8.split()
+    assert len(fields) == 13
+    assert fields[3].endswith(b'/b/%C2%B6-non-ascii')
+    assert fields[4] == b'L'
+    assert fields[5].endswith(b'http://xn--80ahg0a3ax.kz/b/%C2%B6-non-ascii')
+    assert fields[6] == b'text/plain'
+    assert fields[7] == b'-'
+    assert re.match(br'^\d{17}[+]\d{3}', fields[8])
+    assert fields[9] == b'sha1:cdd841ea7c5e46fde3fba56b2e45e4df5aeec439'
+    assert fields[10].endswith('/¶-non-ascii'.encode('utf-8'))
+    assert fields[11] == b'-'
+    extra_info = json.loads(fields[12].decode('utf-8'))
+
 def test_long_warcprox_meta(
         warcprox_, http_daemon, archiving_proxies, playback_proxies):
     urls_before = warcprox_.proxy.running_stats.urls

warcprox/crawl_log.py

@@ -25,6 +25,7 @@ import json
 import os
 import warcprox
 import socket
+import rfc3986
 from urllib3.exceptions import TimeoutError, HTTPError, NewConnectionError, MaxRetryError
 
 class CrawlLogger(object):
@@ -67,8 +68,9 @@ class CrawlLogger(object):
         logging.info('warcprox_meta %s' , recorded_url.warcprox_meta)
 
         hop_path = recorded_url.warcprox_meta.get('metadata', {}).get('hop_path')
-        brozzled_url = recorded_url.warcprox_meta.get('metadata', {}).get('brozzled_url')
+        #URLs are url encoded into plain ascii urls by HTTP spec. Since we're comparing against those, our urls sent over the json blob need to be encoded similarly
-        hop_via_url = recorded_url.warcprox_meta.get('metadata', {}).get('hop_via_url')
+        brozzled_url = self.canonicalize_url(recorded_url.warcprox_meta.get('metadata', {}).get('brozzled_url'))
+        hop_via_url = self.canonicalize_url(recorded_url.warcprox_meta.get('metadata', {}).get('hop_via_url'))
 
         if hop_path is None and brozzled_url is None and hop_via_url is None:
             #No hop info headers provided
@@ -81,13 +83,11 @@ class CrawlLogger(object):
                 hop_via_url = "-"
             #Prefer referer header. Otherwise use provided via_url
             via_url = recorded_url.referer or hop_via_url if hop_path != "-" else "-"
+            logging.info('brozzled_url:%s recorded_url:%s' , brozzled_url, recorded_url.url)
             if brozzled_url != recorded_url.url.decode('ascii') and "brozzled_url" in recorded_url.warcprox_meta.get('metadata', {}).keys():
                 #Requested page is not the Brozzled url, thus we are an embed or redirect.
                 via_url = brozzled_url
-                if hop_path == "-":
+                hop_path = "B" if hop_path == "-" else "".join([hop_path,"B"])
-                    hop_path = "B"
-                else:
-                    hop_path = "".join([hop_path,"B"])
 
         fields = [
             '{:%Y-%m-%dT%H:%M:%S}.{:03d}Z'.format(now, now.microsecond//1000),
@@ -148,3 +148,15 @@ class CrawlLogger(object):
         else:
             return recorded_url.status
 
+    def canonicalize_url(self, url):
+        #URL needs to be split out to separately encode the hostname from the rest of the path.
+        #hostname will be idna encoded (punycode)
+        #The rest of the URL will be urlencoded, but browsers only encode "unsafe" and not "reserved" characters, so ignore the reserved chars.
+        try:
+            parsed_url=rfc3986.urlparse(url)
+            encoded_url=parsed_url.copy_with(host=parsed_url.host.encode('idna'))
+            return encoded_url.unsplit()
+        except (TypeError, ValueError) as e:
+            logging.warning("URL Canonicalization failure. Returning raw url: rfc3986 %s - %s", url, e)
+            return url
+