114 lines
3.6 KiB
Python
114 lines
3.6 KiB
Python
|
# Licensed to the Software Freedom Conservancy (SFC) under one
|
||
|
# or more contributor license agreements. See the NOTICE file
|
||
|
# distributed with this work for additional information
|
||
|
# regarding copyright ownership. The SFC licenses this file
|
||
|
# to you under the Apache License, Version 2.0 (the
|
||
|
# "License"); you may not use this file except in compliance
|
||
|
# with the License. You may obtain a copy of the License at
|
||
|
#
|
||
|
# http://www.apache.org/licenses/LICENSE-2.0
|
||
|
#
|
||
|
# Unless required by applicable law or agreed to in writing,
|
||
|
# software distributed under the License is distributed on an
|
||
|
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
||
|
# KIND, either express or implied. See the License for the
|
||
|
# specific language governing permissions and limitations
|
||
|
# under the License.
|
||
|
|
||
|
import json
|
||
|
import logging
|
||
|
import os
|
||
|
import tempfile
|
||
|
import zipfile
|
||
|
|
||
|
from selenium.common.exceptions import NoSuchElementException
|
||
|
|
||
|
LOGGER = logging.getLogger(__name__)
|
||
|
|
||
|
|
||
|
def format_json(json_struct):
|
||
|
return json.dumps(json_struct, indent=4)
|
||
|
|
||
|
|
||
|
def dump_json(json_struct):
|
||
|
return json.dumps(json_struct)
|
||
|
|
||
|
|
||
|
def load_json(s):
|
||
|
return json.loads(s)
|
||
|
|
||
|
|
||
|
def handle_find_element_exception(e):
|
||
|
if ("Unable to find" in e.response["value"]["message"] or "Unable to locate" in e.response["value"]["message"]):
|
||
|
raise NoSuchElementException("Unable to locate element:")
|
||
|
else:
|
||
|
raise e
|
||
|
|
||
|
|
||
|
def return_value_if_exists(resp):
|
||
|
if resp and "value" in resp:
|
||
|
return resp["value"]
|
||
|
|
||
|
|
||
|
def get_root_parent(elem):
|
||
|
parent = elem.parent
|
||
|
while True:
|
||
|
try:
|
||
|
parent.parent
|
||
|
parent = parent.parent
|
||
|
except AttributeError:
|
||
|
return parent
|
||
|
|
||
|
|
||
|
def unzip_to_temp_dir(zip_file_name):
|
||
|
"""Unzip zipfile to a temporary directory.
|
||
|
|
||
|
The directory of the unzipped files is returned if success,
|
||
|
otherwise None is returned. """
|
||
|
if not zip_file_name or not os.path.exists(zip_file_name):
|
||
|
return None
|
||
|
|
||
|
zf = zipfile.ZipFile(zip_file_name)
|
||
|
|
||
|
if zf.testzip() is not None:
|
||
|
return None
|
||
|
|
||
|
# Unzip the files into a temporary directory
|
||
|
LOGGER.info("Extracting zipped file: %s" % zip_file_name)
|
||
|
tempdir = tempfile.mkdtemp()
|
||
|
|
||
|
try:
|
||
|
# Create directories that don't exist
|
||
|
for zip_name in zf.namelist():
|
||
|
# We have no knowledge on the os where the zipped file was
|
||
|
# created, so we restrict to zip files with paths without
|
||
|
# charactor "\" and "/".
|
||
|
name = (zip_name.replace("\\", os.path.sep).
|
||
|
replace("/", os.path.sep))
|
||
|
dest = os.path.join(tempdir, name)
|
||
|
if (name.endswith(os.path.sep) and not os.path.exists(dest)):
|
||
|
os.mkdir(dest)
|
||
|
LOGGER.debug("Directory %s created." % dest)
|
||
|
|
||
|
# Copy files
|
||
|
for zip_name in zf.namelist():
|
||
|
# We have no knowledge on the os where the zipped file was
|
||
|
# created, so we restrict to zip files with paths without
|
||
|
# charactor "\" and "/".
|
||
|
name = (zip_name.replace("\\", os.path.sep).
|
||
|
replace("/", os.path.sep))
|
||
|
dest = os.path.join(tempdir, name)
|
||
|
if not (name.endswith(os.path.sep)):
|
||
|
LOGGER.debug("Copying file %s......" % dest)
|
||
|
outfile = open(dest, 'wb')
|
||
|
outfile.write(zf.read(zip_name))
|
||
|
outfile.close()
|
||
|
LOGGER.debug("File %s copied." % dest)
|
||
|
|
||
|
LOGGER.info("Unzipped file can be found at %s" % tempdir)
|
||
|
return tempdir
|
||
|
|
||
|
except IOError as err:
|
||
|
LOGGER.error("Error in extracting webdriver.xpi: %s" % err)
|
||
|
return None
|