| 1 |
NAME |
| 2 |
AnyEvent::WebDriver - control browsers using the W3C WebDriver protocol |
| 3 |
|
| 4 |
SYNOPSIS |
| 5 |
# start geckodriver or any other w3c-compatible webdriver via the shell |
| 6 |
$ geckdriver -b myfirefox/firefox --log trace --port 4444 |
| 7 |
|
| 8 |
# then use it |
| 9 |
use AnyEvent::WebDriver; |
| 10 |
|
| 11 |
# create a new webdriver object |
| 12 |
my $wd = new AnyEvent::WebDriver; |
| 13 |
|
| 14 |
# create a new session with default capabilities. |
| 15 |
$wd->new_session ({}); |
| 16 |
|
| 17 |
$wd->navigate_to ("https://duckduckgo.com/html"); |
| 18 |
my $searchbox = $wd->find_element ("css selector" => 'input[type="text"]'); |
| 19 |
|
| 20 |
$wd->element_send_keys ($searchbox => "free software"); |
| 21 |
$wd->element_click ($wd->find_element ("css selector" => 'input[type="submit"]')); |
| 22 |
|
| 23 |
sleep 10; |
| 24 |
|
| 25 |
DESCRIPTION |
| 26 |
This module aims to implement the W3C WebDriver specification which is |
| 27 |
the standardised equivalent to the Selenium WebDriver API., which in |
| 28 |
turn aims at remotely controlling web browsers such as Firefox or |
| 29 |
Chromium. |
| 30 |
|
| 31 |
At the time of this writing, it was only available as a draft document, |
| 32 |
so changes will be expected. Also, only geckodriver did implement it, or |
| 33 |
at least, most of it. |
| 34 |
|
| 35 |
To make most of this module, or, in fact, to make any reasonable use of |
| 36 |
this module, you would need to refer tot he W3C WebDriver document, |
| 37 |
which can be found here <https://w3c.github.io/webdriver/>: |
| 38 |
|
| 39 |
https://w3c.github.io/webdriver/ |
| 40 |
|
| 41 |
CREATING WEBDRIVER OBJECTS |
| 42 |
new AnyEvent::WebDriver key => value... |
| 43 |
Create a new WebDriver object. Example for a remote webdriver |
| 44 |
connection (the only type supported at the moment): |
| 45 |
|
| 46 |
my $wd = new AnyEvent::WebDriver host => "localhost", port => 4444; |
| 47 |
|
| 48 |
Supported keys are: |
| 49 |
|
| 50 |
endpoint => $string |
| 51 |
For remote connections, the endpoint to connect to (defaults to |
| 52 |
"http://localhost:4444"). |
| 53 |
|
| 54 |
proxy => $proxyspec |
| 55 |
The proxy to use (same as the "proxy" argument used by |
| 56 |
AnyEvent::HTTP). The default is "undef", which disables proxies. |
| 57 |
To use the system-provided proxy (e.g. "http_proxy" environment |
| 58 |
variable), specify a value of "default". |
| 59 |
|
| 60 |
autodelete => $boolean |
| 61 |
If true (the default), then automatically execute |
| 62 |
"delete_session" when the WebDriver object is destroyed with an |
| 63 |
active session. IF set to a false value, then the session will |
| 64 |
continue to exist. |
| 65 |
|
| 66 |
SIMPLIFIED API |
| 67 |
This section documents the simplified API, which is really just a very |
| 68 |
thin wrapper around the WebDriver protocol commands. They all block |
| 69 |
(using AnyEvent condvars) the caller until the result is available, so |
| 70 |
must not be called from an event loop callback - see "EVENT BASED API" |
| 71 |
for an alternative. |
| 72 |
|
| 73 |
The method names are preetty much taken directly from the W3C WebDriver |
| 74 |
specification, e.g. the request documented in the "Get All Cookies" |
| 75 |
section is implemented via the "get_all_cookies" method. |
| 76 |
|
| 77 |
The order is the same as in the WebDriver draft at the tiome of this |
| 78 |
writing, and only minimal massaging is done to request parameters and |
| 79 |
results. |
| 80 |
|
| 81 |
SESSIONS |
| 82 |
$wd->new_session ({ key => value... }) |
| 83 |
Try to connect to a webdriver and initialize session with a "new |
| 84 |
session" command, passing the given key-value pairs as value (e.g. |
| 85 |
"capabilities"). |
| 86 |
|
| 87 |
No session-dependent methods must be called before this function |
| 88 |
returns successfully. |
| 89 |
|
| 90 |
On success, "$wd->{sid}" is set to the session id, and |
| 91 |
"$wd->{capabilities}" is set to the returned capabilities. |
| 92 |
|
| 93 |
my $wd = new AnyEvent::Selenium host => "localhost", port => 4545; |
| 94 |
|
| 95 |
$wd->new_session ({ |
| 96 |
capabilities => { |
| 97 |
pageLoadStrategy => "normal", |
| 98 |
}. |
| 99 |
}); |
| 100 |
|
| 101 |
$wd->delete_session |
| 102 |
Deletes the session - the WebDriver object must not be used after |
| 103 |
this call. |
| 104 |
|
| 105 |
$timeouts = $wd->get_timeouts |
| 106 |
Get the current timeouts, e.g.: |
| 107 |
|
| 108 |
my $timeouts = $wd->get_timeouts; |
| 109 |
|
| 110 |
# { implicit => 0, pageLoad => 300000, script => 30000 } |
| 111 |
|
| 112 |
$wd->set_timeouts ($timeouts) |
| 113 |
Sets one or more timeouts, e.g.: |
| 114 |
|
| 115 |
$wd->set_timeouts ({ script => 60000 }); |
| 116 |
|
| 117 |
NAVIGATION |
| 118 |
$wd->navigate_to ($url) |
| 119 |
Navigates to the specified URL. |
| 120 |
|
| 121 |
$url = $wd->get_current_url |
| 122 |
Queries the czurrent page URL as set by "navigate_to". |
| 123 |
|
| 124 |
$wd->back |
| 125 |
The equivalent of pressing "back" in the browser. |
| 126 |
|
| 127 |
$wd->forward |
| 128 |
The equivalent of pressing "forward" in the browser. |
| 129 |
|
| 130 |
$wd->refresh |
| 131 |
The equivalent of pressing "refresh" in the browser. |
| 132 |
|
| 133 |
$title = $wd->get_title |
| 134 |
Returns the current document title. |
| 135 |
|
| 136 |
COMMAND CONTEXTS |
| 137 |
$handle = $wd->get_window_handle |
| 138 |
Returns the current window handle. |
| 139 |
|
| 140 |
$wd->close_window |
| 141 |
Closes the current browsing context. |
| 142 |
|
| 143 |
$wd->switch_to_window ($handle) |
| 144 |
Changes the current browsing context to the given window. |
| 145 |
|
| 146 |
$handles = $wd->get_window_handles |
| 147 |
Return the current window handles as an array-ref of handle IDs. |
| 148 |
|
| 149 |
$handles = $wd->switch_to_frame ($frame) |
| 150 |
Switch to the given frame. |
| 151 |
|
| 152 |
$handles = $wd->switch_to_parent_frame |
| 153 |
Switch to the parent frame. |
| 154 |
|
| 155 |
$rect = $wd->get_window_rect |
| 156 |
Return the current window rect, e.g.: |
| 157 |
|
| 158 |
$rect = $wd->get_window_rect |
| 159 |
# { height => 1040, width => 540, x => 0, y => 0 } |
| 160 |
|
| 161 |
$wd->set_window_rect ($rect) |
| 162 |
Sets the window rect. |
| 163 |
|
| 164 |
$wd->maximize_window |
| 165 |
$wd->minimize_window |
| 166 |
$wd->fullscreen_window |
| 167 |
Changes the window size by eithe3r maximising, minimising or making |
| 168 |
it fullscreen. In my experience, this might timeout if no window |
| 169 |
manager is running. |
| 170 |
|
| 171 |
ELEMENT RETRIEVAL |
| 172 |
$element_id = $wd->find_element ($location_strategy, $selector) |
| 173 |
Finds the first element specified by the given selector and returns |
| 174 |
its web element id (the strong, not the object from the protocol). |
| 175 |
Raises an error when no element was found. |
| 176 |
|
| 177 |
$element = $wd->find_element ("css selector" => "body a"); |
| 178 |
$element = $wd->find_element ("link text" => "Click Here For Porn"); |
| 179 |
$element = $wd->find_element ("partial link text" => "orn"); |
| 180 |
$element = $wd->find_element ("tag name" => "input"); |
| 181 |
$element = $wd->find_element ("xpath" => '//input[@type="text"]'); |
| 182 |
# "decddca8-5986-4e1d-8c93-efe952505a5f" |
| 183 |
|
| 184 |
$element_ids = $wd->find_elements ($location_strategy, $selector) |
| 185 |
As above, but returns an arrayref of all found element IDs. |
| 186 |
|
| 187 |
$element_id = $wd->find_element_from_element ($element_id, |
| 188 |
$location_strategy, $selector) |
| 189 |
Like "find_element", but looks only inside the specified $element. |
| 190 |
|
| 191 |
$element_ids = $wd->find_elements_from_element ($element_id, |
| 192 |
$location_strategy, $selector) |
| 193 |
Like "find_elements", but looks only inside the specified $element. |
| 194 |
|
| 195 |
my $head = $wd->find_element ("tag name" => "head"); |
| 196 |
my $links = $wd->find_elements_from_element ($head, "tag name", "link"); |
| 197 |
|
| 198 |
$element_id = $wd->get_active_element |
| 199 |
Returns the active element. |
| 200 |
|
| 201 |
ELEMENT STATE |
| 202 |
$bool = $wd->is_element_selected |
| 203 |
Returns whether the given input or option element is selected or |
| 204 |
not. |
| 205 |
|
| 206 |
$string = $wd->get_element_attribute ($element_id, $name) |
| 207 |
Returns the value of the given attribute. |
| 208 |
|
| 209 |
$string = $wd->get_element_property ($element_id, $name) |
| 210 |
Returns the value of the given property. |
| 211 |
|
| 212 |
$string = $wd->get_element_css_value ($element_id, $name) |
| 213 |
Returns the value of the given css value. |
| 214 |
|
| 215 |
$string = $wd->get_element_text ($element_id) |
| 216 |
Returns the (rendered) text content of the given element. |
| 217 |
|
| 218 |
$string = $wd->get_element_tag_name ($element_id) |
| 219 |
Returns the tag of the given element. |
| 220 |
|
| 221 |
$rect = $wd->get_element_rect ($element_id) |
| 222 |
Returns the element rect of the given element. |
| 223 |
|
| 224 |
$bool = $wd->is_element_enabled |
| 225 |
Returns whether the element is enabled or not. |
| 226 |
|
| 227 |
ELEMENT INTERACTION |
| 228 |
$wd->element_click ($element_id) |
| 229 |
Clicks the given element. |
| 230 |
|
| 231 |
$wd->element_clear ($element_id) |
| 232 |
Clear the contents of the given element. |
| 233 |
|
| 234 |
$wd->element_send_keys ($element_id, $text) |
| 235 |
Sends the given text as key events to the given element. |
| 236 |
|
| 237 |
DOCUMENT HANDLING |
| 238 |
$source = $wd->get_page_source |
| 239 |
Returns the (HTML/XML) page source of the current document. |
| 240 |
|
| 241 |
$results = $wd->execute_script ($javascript, $args) |
| 242 |
Synchronously execute the given script with given arguments and |
| 243 |
return its results ($args can be "undef" if no arguments are |
| 244 |
wanted/needed). |
| 245 |
|
| 246 |
$ten = $wd->execute_script ("return arguments[0]+arguments[1]", [3, 7]); |
| 247 |
|
| 248 |
$results = $wd->execute_async_script ($javascript, $args) |
| 249 |
Similar to "execute_script", but doesn't wait for script to return, |
| 250 |
but instead waits for the script to call its last argument, which is |
| 251 |
added to $args automatically. |
| 252 |
|
| 253 |
$twenty = $wd->execute_async_script ("arguments[0](20)", undef); |
| 254 |
|
| 255 |
COOKIES |
| 256 |
$cookies = $wd->get_all_cookies |
| 257 |
Returns all cookies, as an arrayref of hashrefs. |
| 258 |
|
| 259 |
# google surely sets a lot of cookies without my consent |
| 260 |
$wd->navigate_to ("http://google.com"); |
| 261 |
use Data::Dump; |
| 262 |
ddx $wd->get_all_cookies; |
| 263 |
|
| 264 |
$cookie = $wd->get_named_cookie ($name) |
| 265 |
Returns a single cookie as a hashref. |
| 266 |
|
| 267 |
$wd->add_cookie ($cookie) |
| 268 |
Adds the given cookie hashref. |
| 269 |
|
| 270 |
$wd->delete_cookie ($name) |
| 271 |
Delete the named cookie. |
| 272 |
|
| 273 |
$wd->delete_all_cookies |
| 274 |
Delete all cookies. |
| 275 |
|
| 276 |
ACTIONS |
| 277 |
$wd->perform_actions ($actions) |
| 278 |
Perform the given actions (an arrayref of action specifications |
| 279 |
simulating user activity). For further details, read the spec. |
| 280 |
|
| 281 |
An example to get you started: |
| 282 |
|
| 283 |
$wd->navigate_to ("https://duckduckgo.com/html"); |
| 284 |
$wd->set_timeouts ({ implicit => 10000 }); |
| 285 |
my $input = $wd->find_element ("css selector", 'input[type="text"]'); |
| 286 |
$wd->perform_actions ([ |
| 287 |
{ |
| 288 |
id => "myfatfinger", |
| 289 |
type => "pointer", |
| 290 |
pointerType => "touch", |
| 291 |
actions => [ |
| 292 |
{ type => "pointerMove", duration => 100, origin => $wd->element_object ($input), x => 40, y => 5 }, |
| 293 |
{ type => "pointerDown", button => 1 }, |
| 294 |
{ type => "pause", duration => 40 }, |
| 295 |
{ type => "pointerUp", button => 1 }, |
| 296 |
], |
| 297 |
}, |
| 298 |
{ |
| 299 |
id => "mykeyboard", |
| 300 |
type => "key", |
| 301 |
actions => [ |
| 302 |
{ type => "pause" }, |
| 303 |
{ type => "pause" }, |
| 304 |
{ type => "pause" }, |
| 305 |
{ type => "pause" }, |
| 306 |
{ type => "keyDown", value => "a" }, |
| 307 |
{ type => "pause", duration => 100 }, |
| 308 |
{ type => "keyUp", value => "a" }, |
| 309 |
{ type => "pause", duration => 100 }, |
| 310 |
{ type => "keyDown", value => "b" }, |
| 311 |
{ type => "pause", duration => 100 }, |
| 312 |
{ type => "keyUp", value => "b" }, |
| 313 |
{ type => "pause", duration => 2000 }, |
| 314 |
{ type => "keyDown", value => "\x{E007}" }, # enter |
| 315 |
{ type => "pause", duration => 100 }, |
| 316 |
{ type => "keyUp", value => "\x{E007}" }, # enter |
| 317 |
{ type => "pause", duration => 5000 }, |
| 318 |
], |
| 319 |
}, |
| 320 |
]); |
| 321 |
|
| 322 |
$wd->release_actions |
| 323 |
Release all keys and pointer buttons currently depressed. |
| 324 |
|
| 325 |
USER PROMPTS |
| 326 |
$wd->dismiss_alert |
| 327 |
Dismiss a simple dialog, if present. |
| 328 |
|
| 329 |
$wd->accept_alert |
| 330 |
Accept a simple dialog, if present. |
| 331 |
|
| 332 |
$text = $wd->get_alert_text |
| 333 |
Returns the text of any simple dialog. |
| 334 |
|
| 335 |
$text = $wd->send_alert_text |
| 336 |
Fills in the user prompt with the given text. |
| 337 |
|
| 338 |
SCREEN CAPTURE |
| 339 |
$wd->take_screenshot |
| 340 |
Create a screenshot, returning it as a PNG image in a data url. |
| 341 |
|
| 342 |
$wd->take_element_screenshot ($element_id) |
| 343 |
Accept a simple dialog, if present. |
| 344 |
|
| 345 |
HELPER METHODS |
| 346 |
$object = $wd->element_object ($element_id) |
| 347 |
Encoding element ids in data structures is done by represetning them |
| 348 |
as an object with a special key and the element id as value. This |
| 349 |
helper method does this for you. |
| 350 |
|
| 351 |
EVENT BASED API |
| 352 |
This module wouldn't be a good AnyEvent citizen if it didn't have a true |
| 353 |
event-based API. |
| 354 |
|
| 355 |
In fact, the simplified API, as documented above, is emulated via the |
| 356 |
event-based API and an "AUTOLOAD" function that automatically provides |
| 357 |
blocking wrappers around the callback-based API. |
| 358 |
|
| 359 |
Every method documented in the "SIMPLIFIED API" section has an |
| 360 |
equivalent event-based method that is formed by appending a underscore |
| 361 |
("_") to the method name, and appending a callback to the argument list |
| 362 |
(mnemonic: the underscore indicates the "the action is not yet finished" |
| 363 |
after the call returns). |
| 364 |
|
| 365 |
For example, instead of a blocking calls to "new_session", "navigate_to" |
| 366 |
and "back", you can make a callback-based ones: |
| 367 |
|
| 368 |
my $cv = AE::cv; |
| 369 |
|
| 370 |
$wd->new_session ({}, sub { |
| 371 |
my ($status, $value) = @_, |
| 372 |
|
| 373 |
die "error $value->{error}" if $status ne "200"; |
| 374 |
|
| 375 |
$wd->navigate_to_ ("http://www.nethype.de", sub { |
| 376 |
|
| 377 |
$wd->back_ (sub { |
| 378 |
print "all done\n"; |
| 379 |
$cv->send; |
| 380 |
}); |
| 381 |
|
| 382 |
}); |
| 383 |
}); |
| 384 |
|
| 385 |
$cv->recv; |
| 386 |
|
| 387 |
While the blocking methods "croak" on errors, the callback-based ones |
| 388 |
all pass two values to the callback, $status and $res, where $status is |
| 389 |
the HTTP status code (200 for successful requests, typically 4xx ot 5xx |
| 390 |
for errors), and $res is the value of the "value" key in the JSON |
| 391 |
response object. |
| 392 |
|
| 393 |
Other than that, the underscore variants and the blocking variants are |
| 394 |
identical. |
| 395 |
|
| 396 |
LOW LEVEL API |
| 397 |
All the simplfiied API methods are very thin wrappers around WebDriver |
| 398 |
commands of the same name. Theyx are all implemented in terms of the |
| 399 |
low-level methods ("req", "get", "post" and "delete"), which exists in |
| 400 |
blocking and callback-based variants ("req_", "get_", "post_" and |
| 401 |
"delete_"). |
| 402 |
|
| 403 |
Examples are after the function descriptions. |
| 404 |
|
| 405 |
$wd->req_ ($method, $uri, $body, $cb->($status, $value)) |
| 406 |
$value = $wd->req ($method, $uri, $body) |
| 407 |
Appends the $uri to the "endpoint/session/{sessionid}/" URL and |
| 408 |
makes a HTTP $method request ("GET", "POST" etc.). "POST" requests |
| 409 |
can provide a UTF-8-encoded JSON text as HTTP request body, or the |
| 410 |
empty string to indicate no body is used. |
| 411 |
|
| 412 |
For the callback version, the callback gets passed the HTTP status |
| 413 |
code (200 for every successful request), and the value of the |
| 414 |
"value" key in the JSON response object as second argument. |
| 415 |
|
| 416 |
$wd->get_ ($uri, $cb->($status, $value)) |
| 417 |
$value = $wd->get ($uri) |
| 418 |
Simply a call to "req_" with $method set to "GET" and an empty body. |
| 419 |
|
| 420 |
$wd->post_ ($uri, $data, $cb->($status, $value)) |
| 421 |
$value = $wd->post ($uri, $data) |
| 422 |
Simply a call to "req_" with $method set to "POST" - if $body is |
| 423 |
"undef", then an empty object is send, otherwise, $data must be a |
| 424 |
valid request object, which gets encoded into JSON for you. |
| 425 |
|
| 426 |
$wd->delete_ ($uri, $cb->($status, $value)) |
| 427 |
$value = $wd->delete ($uri) |
| 428 |
Simply a call to "req_" with $method set to "DELETE" and an empty |
| 429 |
body. |
| 430 |
|
| 431 |
Example: implement "get_all_cookies", which is a simple "GET" request |
| 432 |
without any parameters: |
| 433 |
|
| 434 |
$cookies = $wd->get ("cookie"); |
| 435 |
|
| 436 |
Example: implement "execute_script", which needs some parameters: |
| 437 |
|
| 438 |
$results = $wd->post ("execute/sync" => { script => "$javascript", args => [] }); |
| 439 |
|
| 440 |
Example: call "find_elements" to find all "IMG" elements, stripping the |
| 441 |
returned element objects to only return the element ID strings: |
| 442 |
|
| 443 |
my $elems = $wd->post (elements => { using => "css selector", value => "img" }); |
| 444 |
|
| 445 |
# yes, the W3C found an interetsing way around the typelessness of JSON |
| 446 |
$_ = $_->{"element-6066-11e4-a52e-4f735466cecf"} |
| 447 |
for @$elems; |
| 448 |
|
| 449 |
HISTORY |
| 450 |
This module was developed simply because I couldn't get the existing |
| 451 |
"Selenium::Remote::Driver" module to work, ever, despite multiple |
| 452 |
attempts over the years and trying to report multiple bugs, which have |
| 453 |
been completely ignored. It's also not event-based, so, yeah... |
| 454 |
|
| 455 |
AUTHOR |
| 456 |
Marc Lehmann <schmorp@schmorp.de> |
| 457 |
http://anyevent.schmorp.de |
| 458 |
|
