class Puppet:
def __init__(self, binary: str, profile: str):
self.__has_marionette = False
self.__auto_download = False
self.__download_dir = ""
if not Path(binary).is_file():
return
if not Path(profile).is_dir():
return
# geckodriver の log ファイル出力を抑止する
NO_LOG = "-"
self.marionette = Marionette(bin=binary,
gecko_log=NO_LOG,
profile=profile)
# start_session しないと quit もできない
self.marionette.start_session()
self.__has_marionette = True
@property
def has_marionette(self):
return self.__has_marionette
@property
def auto_download(self):
return self.__auto_download
def __activate_auto_download(self):
# 一度有効にすると同セッション内では無効にできない
# firefox52 では MIME_TYPES.rdf, firefox60 では handlers.json に
# ファイルダウンロード時の動作設定が記述されている（text/plain はプログラムで開く、など）
# 自動ダウンロードするため既存の設定は削除する
MIME_TYPES_HANDLERS = ["MIME_TYPES.rdf", "handlers.json"]
for name in MIME_TYPES_HANDLERS:
p = Path(self.marionette.profile_path).joinpath(name)
if p.is_file():
p.unlink()
self.marionette.set_pref("browser.download.useDownloadDir", True)
self.marionette.set_pref("browser.helperApps.neverAsk.saveToDisk",
",".join(MIME_TYPES))
USER_DEFINED = 2
self.marionette.set_pref("browser.download.folderList", USER_DEFINED)
@property
def download_dir(self):
if self.__auto_download == False:
raise Exception("auto download has not been activated")
return self.__download_dir
@download_dir.setter
def download_dir(self, dir: str):
p = Path(dir)
if not p.is_dir():
return
full_path = str(p.resolve())
if self.__auto_download == False:
self.__activate_auto_download()
self.__auto_download = True
self.marionette.set_pref("browser.download.dir", full_path)
self.__download_dir = full_path
def set_download(self, dir: str):
self.download_dir = dir
def query_selector(self, selectors: str) -> HTMLElement:
METHOD_CSS_SELECTOR = "css selector"
return self.marionette.find_element(METHOD_CSS_SELECTOR, selectors)
def query_selectors(self, selectors: str) -> List[HTMLElement]:
METHOD_CSS_SELECTOR = "css selector"
return self.marionette.find_elements(METHOD_CSS_SELECTOR, selectors)
def wait(self, seconds: int):
actions = Actions(self.marionette)
actions.wait(seconds).perform()
def quit(self):
self.marionette.quit()
def exec(self, script: str):
# script 内での記述簡略化のため
mrnt = self.marionette
set_download = self.set_download
wait = self.wait
quit = self.quit
query_selector = self.query_selector
query_selectors = self.query_selectors
exec(script)
class Puppet:
MIME_TYPES = [
"application/epub+zip",
"application/gzip",
"application/java-archive",
"application/json",
"application/ld+json",
"application/msword",
"application/octet-stream",
"application/ogg",
"application/pdf",
"application/rtf",
"application/vnd.amazon.ebook",
"application/vnd.apple.installer+xml",
"application/vnd.mozilla.xul+xml",
"application/vnd.ms-excel",
"application/vnd.ms-fontobject",
"application/vnd.ms-powerpoint",
"application/vnd.oasis.opendocument.presentation",
"application/vnd.oasis.opendocument.spreadsheet",
"application/vnd.oasis.opendocument.text",
"application/vnd.openxmlformats-officedocument.presentationml.presentation",
"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
"application/vnd.openxmlformats-officedocument.wordprocessingml.document",
"application/vnd.visio",
"application/x-7z-compressed",
"application/x-abiword",
"application/x-bzip",
"application/x-bzip2",
"application/x-csh",
"application/x-freearc",
"application/xhtml+xml",
"application/xml",
"application/x-rar-compressed",
"application/x-sh",
"application/x-shockwave-flash",
"application/x-tar",
"application/zip",
"appliction/php",
"audio/aac",
"audio/midi audio/x-midi",
"audio/mpeg",
"audio/ogg",
"audio/wav",
"audio/webm",
"font/otf",
"font/ttf",
"font/woff",
"font/woff2",
"image/bmp",
"image/gif",
"image/jpeg",
"image/png",
"image/svg+xml",
"image/tiff",
"image/vnd.microsoft.icon",
"image/webp",
"text/calendar",
"text/css",
"text/csv",
"text/html",
"text/javascript",
"text/javascript",
"text/plain",
"text/xml",
"video/3gpp",
"video/3gpp2",
"video/mp2t",
"video/mpeg",
"video/ogg",
"video/webm",
"video/x-msvideo"
]
METHOD_CSS_SELECTOR = "css selector"
NO_LOG = "-"
DELETE_TARGET_FILES = ["mimeTypes.rdf", "handlers.json"]
USER_DEFINED = 2
GECKO_LOG = Path(__file__).parent.resolve()
def __init__(self, binary: str, profile: str):
self.__has_session = False
self.__auto_download = False
self.__download_dir = ""
if not Path(binary).is_file():
print(f"Binary {binary} Not Found")
return
if not Path(profile).is_dir():
print(f"Profile {profile} Not Found")
return
# geckodriver の log ファイル出力を抑止する
self.marionette = Marionette(
bin=binary, gecko_log=self.NO_LOG, profile=profile)
# start_session 前にファイルを消しておかないと
# 後で自動ダウンロードできない
self.__delete_download_profile()
# start_session しないと quit もできない
self.marionette.start_session()
self.__has_session = True
def __enter__(self):
return self
def __exit__(self, ex_type, ex_value, trace):
if self.has_session:
self.quit()
@property
def has_session(self):
return self.__has_session
@property
def auto_download(self):
return self.__auto_download
def __delete_download_profile(self):
# mimeTypes.rdf と handlers.json に
# ファイル読み込み時の動作設定が保存されている（text/plain はファイルを保存、など）
# 自動ダウンロードするため既存の設定は削除する
for name in self.DELETE_TARGET_FILES:
p = Path(self.marionette.profile_path).joinpath(name)
if p.is_file():
p.unlink()
def __activate_auto_download(self):
# 一度有効にすると同セッション内では無効にできない
self.marionette.set_pref("browser.download.useDownloadDir", True)
self.marionette.set_pref("browser.helperApps.neverAsk.saveToDisk",
",".join(self.MIME_TYPES))
self.marionette.set_pref(
"browser.download.folderList", self.USER_DEFINED)
self.marionette.set_pref("browser.download.lastDir", None)
self.__auto_download = True
@property
def download_dir(self):
if self.__auto_download == False:
raise Exception("auto download not activated")
return self.__download_dir
@download_dir.setter
def download_dir(self, dir: str):
p = Path(dir)
if not p.is_dir():
print(f"Download Dir {dir} Not Found")
return
full_path = str(p.resolve())
if self.__auto_download == False:
self.__activate_auto_download()
# self.__auto_download = True
self.marionette.set_pref("browser.download.dir", full_path)
self.marionette.set_pref("browser.download.downloadDir", full_path)
self.__download_dir = full_path
def set_download(self, dir: str):
self.download_dir = dir
def query_selector(self, selectors: str) -> HTMLElement:
return self.marionette.find_element(self.METHOD_CSS_SELECTOR, selectors)
def query_selectors(self, selectors: str) -> List[HTMLElement]:
return self.marionette.find_elements(self.METHOD_CSS_SELECTOR, selectors)
def wait(self, seconds: int):
actions = Actions(self.marionette)
actions.wait(seconds).perform()
def quit(self):
profile = Path(self.marionette.profile_path)
self.marionette.quit(clean=True)
# self.__forced_rmdir(profile)
# Path(self.GECKO_LOG).unlink()
self.__has_session = False
def exec(self, script: str) -> Optional[str]:
# script 内での記述簡略化のため
mrnt = self.marionette
set_download = self.set_download
wait = self.wait
quit = self.quit
query_selector = self.query_selector
query_selectors = self.query_selectors
try:
exec(script)
return None
except Exception as err:
return str(err.args[0])
@classmethod
def __forced_rmdir(self, p: Path):
if p.is_dir():
for f in p.iterdir():
if f.is_file():
f.unlink()
elif f.is_dir():
self.__forced_rmdir(f)
p.rmdir()
class FirefoxMarionetteBase(object):
"""
Wrap Marionette/Firefox into convenient interface.
- https://marionette-client.readthedocs.io/
- https://marionette-client.readthedocs.io/en/master/reference.html
- https://marionette-client.readthedocs.io/en/master/interactive.html
"""
def __init__(self):
logger.info('Starting Marionette Gecko wrapper')
# Configuration
self.firefox_bin = self.find_firefox()
self.firefox_host = 'localhost'
self.firefox_port = 2828
# TODO: Make configurable
self.firefox_verbosity = 1
#self.firefox_verbosity = 2
# Timeout configuration
self.startup_timeout = 20.0
self.socket_timeout = 32.0
self.page_timeout = 30.0
self.script_timeout = 20.0
self.shutdown_timeout = 10.0
# Instance state defaults
self.marionette = None
self.firefox_run_headless = True
self.firefox_do_shutdown = False
self.firefox_already_started = False
def enable_headless(self, run_headless=True):
self.firefox_run_headless = run_headless
def enable_shutdown(self, do_shutdown=True):
self.firefox_do_shutdown = do_shutdown
def boot_firefox(self, headless=True):
# Indicate whether to run in headless mode
self.enable_headless(headless)
# Optionally shut down Marionette/Firefox after performing work
# This will just be called if Python exits normally
atexit.register(self.shutdown)
# Check whether Firefox is already running
logger.info(
'Check for running instance of Marionette/Firefox at {}:{}'.format(
self.firefox_host, self.firefox_port))
if check_socket(self.firefox_host, self.firefox_port):
logger.info('Will reuse running Marionette/Firefox')
self.firefox_bin = None
self.firefox_already_started = True
else:
logger.info('Will launch new Marionette/Firefox instance')
# Connect to / start Marionette Gecko engine
self.marionette = Marionette(host=self.firefox_host,
port=self.firefox_port,
bin=self.firefox_bin,
socket_timeout=self.socket_timeout,
startup_timeout=self.startup_timeout,
headless=self.firefox_run_headless,
verbose=self.firefox_verbosity)
self.marionette.DEFAULT_SHUTDOWN_TIMEOUT = self.shutdown_timeout
# Start a session with Marionette Gecko engine
self.marionette.start_session()
# Configure Marionette
self.configure_marionette()
def configure_marionette(self):
# This specifies the time to wait for the page loading to complete.
self.marionette.timeout.page_load = self.page_timeout
# This specifies the time to wait for injected scripts to finish
# before interrupting them.
self.marionette.timeout.script = self.script_timeout
# Configure a HTTP proxy server
self.marionette.set_pref('network.proxy.type', 0, default_branch=True)
@classmethod
def find_firefox(cls):
candidates = where.where('firefox')
candidates += [
'/Applications/Firefox.app/Contents/MacOS/firefox-bin',
]
firefox = find_program_candidate(candidates)
logger.info('Found "firefox" program at {}'.format(firefox))
return firefox
def get_status(self):
attributes = ['session', 'session_id']
data = OrderedDict()
for attribute in attributes:
data[attribute] = getattr(self.marionette, attribute)
return data
def log_status(self):
logger.info('Marionette report: {}'.format(
json.dumps(self.get_status(), indent=4)))
def has_active_session(self):
is_initialized = self.marionette is not None and self.marionette.session_id is not None
return is_initialized
def ensure_session(self):
#self.log_status()
if not self.has_active_session():
self.boot_firefox()
logger.info(
'No session with Marionette, started new session {}'.format(
self.marionette.session_id))
def shutdown(self):
if self.firefox_do_shutdown:
logger.info('Aiming at shutdown')
if self.firefox_already_started:
logger.warning(
'Can not shutdown Firefox as it was already running before starting this program'
)
return False
logger.info('Shutting down Marionette/Firefox')
if self.marionette is not None:
self.marionette.quit()
return True
def find_tag(self, tagname):
try:
element = self.marionette.find_element("tag name", tagname)
return element
except NoSuchElementException:
pass
def wait_for_element_tag(self, tagname):
"""
Wait for element to appear.
"""
waiter = Wait(self.marionette, timeout=20.0, interval=0.1)
element = waiter.until(lambda m: self.find_tag(tagname))
return element
def render_image(self, element=None):
"""
Return screenshot from element.
"""
image = self.marionette.screenshot(element=element, format='binary')
return image
def set_window_size(self, width, height):
self.marionette.set_window_rect(width=width, height=height)
def get_window_rect(self):
return self.marionette.window_rect