Make the code work with Python 3 + latest Selenium

html/miniwob/click-dialog-2.html

@@ -41,11 +41,11 @@
     position: {my: 'center', at: 'center', of: document.getElementById('area')},
     buttons: [
       { text: 'Cancel', click: function(e) {
-        var r = e.toElement.innerHTML === expectedButton ? 1.0 : -1.0;
+        var r = e.target.innerHTML === expectedButton ? 1.0 : -1.0;
         core.endEpisode(r, r > 0);
       } },
       { text: 'OK', click: function(e) {
-        var r = e.toElement.innerHTML === expectedButton ? 1.0 : -1.0;
+        var r = e.target.innerHTML === expectedButton ? 1.0 : -1.0;
         core.endEpisode(r, r > 0);
       } }
     ]

python/README.md

@@ -10,7 +10,7 @@
 
 - Selenium
   - Outside this repository, download
-    [ChromeDriver](https://sites.google.com/a/chromium.org/chromedriver/downloads).
+    [ChromeDriver](https://chromedriver.chromium.org/downloads).
     Unzip it and then add the directory
     containing the `chromedriver` executable to the `PATH` environment variable
     ```

python/miniwob/action.py

@@ -3,11 +3,11 @@
 import logging
 
 from selenium.webdriver.common.action_chains import ActionChains
+from selenium.webdriver.common.by import By
 
 
-class MiniWoBAction(object):
+class MiniWoBAction(metaclass=abc.ABCMeta):
     """Defines an action in its __call__ method."""
-    __metaclass__ = abc.ABCMeta
 
     @abc.abstractmethod
     def __call__(self, driver):
@@ -61,10 +61,12 @@ def __init__(self, left, top):
 
     def __call__(self, driver):
         """Clicks at coordinates (left, top)"""
-        body = driver.find_element_by_tag_name('body')
+        body = driver.find_element(By.TAG_NAME, 'body')
+        # The offset is from the center, not top-left.
+        x = - body.size['width'] / 2 + self.left
+        y = - body.size['height'] / 2 + self.top
         chain = ActionChains(driver)
-        chain.move_to_element_with_offset(
-                body, self.left, self.top).click().perform()
+        chain.move_to_element_with_offset(body, x, y).click().perform()
 
     @property
     def left(self):
@@ -114,7 +116,7 @@ def __init__(self, element, fail_hard=False):
     def __call__(self, driver):
         if self.element.tag == 'select':
             # SPECIAL CASE: <select>
-            body = driver.find_element_by_tag_name('body')
+            body = driver.find_element(By.TAG_NAME, 'body')
             chain = ActionChains(driver)
             chain.move_to_element_with_offset(
                     body, self.element.left + 5,
@@ -126,7 +128,7 @@ def __call__(self, driver):
             if self._fail_hard:
                 raise RuntimeError('{} failed: {}'.format(self, result))
             else:
-                logging.warn('%s failed: %s', self, result)
+                logging.warning('%s failed: %s', self, result)
 
     @property
     def ref(self):

python/miniwob/environment.py

@@ -56,7 +56,7 @@ def configure(self, num_instances=1, seeds=None, **kwargs):
         for instance in self.instances:
             instance.close()
         self.instances = []
-        for index in xrange(num_instances):
+        for index in range(num_instances):
             logging.info('Starting WebDriver Instance %d', index)
             instance = MiniWoBInstance(
                     index, self.subdomain, seeds[index], **kwargs)
@@ -176,13 +176,13 @@ def test_environment():
     try:
         task_name = sys.argv[1]
     except IndexError:
-        print 'Usage: python {} TASK_NAME'.format(sys.argv[0])
+        print('Usage: python {} TASK_NAME'.format(sys.argv[0]))
         exit(1)
     env = MiniWoBEnvironment(task_name)
     base_url = os.environ.get('MINIWOB_BASE_URL')
     env.configure(num_instances=1, seeds=[0], base_url=base_url)
     states = env.reset()
-    print states[0].dom.visualize()
+    print(states[0].dom.visualize())
     env.close()
 
 if __name__ == '__main__':

python/miniwob/fields.py

@@ -45,14 +45,14 @@ def __len__(self):
 
     @property
     def keys(self):
-        return self._d.keys()
+        return list(self._d.keys())
 
     @property
     def values(self):
-        return self._d.values()
+        return list(self._d.values())
 
     def __repr__(self):
-        return '\n'.join(u'{}: {}'.format(k, repr(v)) for k, v in self._d.items())
+        return '\n'.join('{}: {}'.format(k, repr(v)) for k, v in self._d.items())
     __str__ = __repr__
 
 
@@ -134,7 +134,7 @@ def extract_click_checkboxes(utterance):
     else:
         targets = re.split(', ?', targets)
     fields = dict(zip(
-        ["target {}".format(i) for i in xrange(len(targets))], targets))
+        ["target {}".format(i) for i in range(len(targets))], targets))
     fields["button"] = "submit"
     return Fields(fields)
 FIELD_EXTRACTORS['click-checkboxes'] = extract_click_checkboxes
@@ -156,7 +156,7 @@ def extract_click_checkboxes_soft(utterance):
     targets = re.match(r'Select words similar to (.*) and click Submit\.', utterance).group(1)
     targets = re.split(', ?', targets)
     fields = dict(zip(
-        ["target {}".format(i) for i in xrange(len(targets))], targets))
+        ["target {}".format(i) for i in range(len(targets))], targets))
     fields["button"] = "submit"
     return Fields(fields)
 FIELD_EXTRACTORS['click-checkboxes-soft'] = extract_click_checkboxes_soft
@@ -900,17 +900,17 @@ def extract_utterances():
     try:
         task_name = sys.argv[1]
     except:
-        print >> sys.stderr, 'Usage: {} task_name'.format(sys.argv[0])
+        print('Usage: {} task_name'.format(sys.argv[0]), file=sys.stderr)
         exit(1)
     FIELD_EXTRACTORS[task_name] = lambda utt: Fields({})
     from miniwob.environment import MiniWoBEnvironment
     env = MiniWoBEnvironment(task_name)
     base_url = os.environ.get('MINIWOB_BASE_URL')
-    env.configure(num_instances=4, seeds=range(4), base_url=base_url)
-    for i in xrange(25):
+    env.configure(num_instances=4, seeds=list(range(4)), base_url=base_url)
+    for i in range(25):
         states = env.reset()
         for state in states:
-            print 'UTT:\t{}'.format(state.utterance.replace('\n', ' '))
+            print('UTT:\t{}'.format(state.utterance.replace('\n', ' ')))
     env.close()
 
 if __name__ == '__main__':

python/miniwob/instance.py

@@ -1,9 +1,9 @@
 import json
 import logging
-from Queue import Queue
+from queue import Queue
 import time
 import traceback
-import urlparse
+import urllib.parse
 from threading import Thread
 
 import numpy as np
@@ -76,14 +76,14 @@ def __init__(self, index, subdomain, seed, headless=False,
             assert not base_url.startswith('file://'),\
                     ('For {} domain, MINIWOB_BASE_URL cannot be file://. '
                      ' See "Run a simple server" in README').format(subdomain)
-            self.url = urlparse.urljoin(base_url,
+            self.url = urllib.parse.urljoin(base_url,
                     subdomain.replace('.', '/') + '/wrapper.html')
             self.window_width = self.FLIGHT_WINDOW_WIDTH
             self.window_height = self.FLIGHT_WINDOW_HEIGHT
             self.task_width = self.FLIGHT_TASK_WIDTH
             self.task_height = self.FLIGHT_TASK_HEIGHT
         else:
-            self.url = urlparse.urljoin(base_url,
+            self.url = urllib.parse.urljoin(base_url,
                     'miniwob/{}.html'.format(subdomain))
             self.window_width = self.WINDOW_WIDTH
             self.window_height = self.WINDOW_HEIGHT
@@ -156,7 +156,7 @@ def create_driver(self):
                     .format(self.window_width, self.window_height))
             options.add_argument('window-position={},{}'
                     .format(9000, 30 + self.index * (self.window_height + 30)))
-        self.driver = webdriver.Chrome(chrome_options=options)
+        self.driver = webdriver.Chrome(options=options)
         self.driver.implicitly_wait(5)
         if self.headless:
             self.driver.get(self.url)
@@ -247,13 +247,9 @@ def begin_task(self, seed=None):
         if seed is not None:
             self.set_seed(seed)
         self.set_mode(self.mode)
-        # Wait for the sync screen, then click on it
-        #element = WebDriverWait(self.driver, 5).until(
-        #                EC.element_to_be_clickable((By.ID, self.SYNC_SCREEN_ID)))
-        #self.driver.find_element_by_id(self.SYNC_SCREEN_ID).click()
         self.driver.execute_script('core.startEpisodeReal();')
         if self.block_on_reset:
-            for _ in xrange(self.RESET_BLOCK_MAX_ATTEMPT):
+            for _ in range(self.RESET_BLOCK_MAX_ATTEMPT):
                 if self.driver.execute_script('return WOB_TASK_READY;'):
                     break
                 time.sleep(self.RESET_BLOCK_SLEEP_TIME)
@@ -274,7 +270,7 @@ def perform(self, action):
         """
         if action is not None:
             if self.get_metadata()['done']:
-                logging.warn('Cannot call %s on instance %d, which is already done',
+                logging.warning('Cannot call %s on instance %d, which is already done',
                         action, self.index)
             else:
                 action(self.driver)

python/miniwob/screenshot.py

@@ -4,7 +4,7 @@
 
 import numpy as np
 
-from cStringIO import StringIO
+from io import StringIO
 from PIL import Image, ImageDraw
 
 
@@ -79,8 +79,8 @@ def create_gif(path_prefix):
 
 if __name__ == '__main__':
     if len(sys.argv) != 2:
-        print 'Usage: {} PATH_PREFIX'.format(sys.argv[0])
-        print '  where PATH_PREFIX is something like '\
-                'data/experiments/123_unnamed/traces/test/2000-img/2000-3'
+        print('Usage: {} PATH_PREFIX'.format(sys.argv[0]))
+        print('  where PATH_PREFIX is something like '
+              'data/experiments/123_unnamed/traces/test/2000-img/2000-3')
         exit(1)
     create_gif(sys.argv[1])

python/miniwob/state.py

@@ -24,7 +24,7 @@ def __init__(self, utterance, fields, dom_info):
         """
         ################
         # Parse utterance
-        assert isinstance(utterance, basestring)
+        assert isinstance(utterance, str)
         self._phrase = Phrase(utterance)
         self._fields = fields
         ################
@@ -145,7 +145,7 @@ def __init__(self, raw_dom, parent=None, dom_elements=None):
         if self.tag == 't':
             self._ref = None  # ignore refs for text, since they are unreliable
         if 'text' in raw_dom:
-            self._text = unicode(raw_dom['text'])
+            self._text = str(raw_dom['text'])
         else:
             self._text = None
         self._value = raw_dom.get('value')
@@ -168,7 +168,7 @@ def __init__(self, raw_dom, parent=None, dom_elements=None):
         # Fix a bug where sometimes children are created even though all children are <t>
         # (which will incorrectly make this element a non-leaf and thus unclickable)
         if self._children and all(child.tag == 't' for child in self._children):
-            self._text = u' '.join(child.text for child in self._children)
+            self._text = ' '.join(child.text for child in self._children)
             self._children = []
         # Add to the collection
         if dom_elements is not None:
@@ -350,7 +350,7 @@ def visualize(self, join=True):
         lines = []
         lines.append('- {}'.format(self))
         for i, child in enumerate(self.children):
-            if isinstance(child, unicode):
+            if isinstance(child, str):
                 child = child[:20] + '...' if len(child) > 20 else child
                 lines.append('  |- "{}"'.format(child))
             else:

python/miniwob/utils.py

@@ -1,4 +1,4 @@
-import regex as re
+import re
 
 
 def strip_punctuation(uni):
@@ -10,7 +10,7 @@ def strip_punctuation(uni):
     Returns:
         unicode
     """
-    return re.sub(ur"\p{P}+", " ", uni)
+    return re.sub(r"\p{P}+", " ", uni)
 
 
 def strip_whitespace(uni):
@@ -22,7 +22,7 @@ def strip_whitespace(uni):
     Returns:
         unicode
     """
-    return re.sub(ur"\s+", u"", uni)
+    return re.sub(r"\s+", "", uni)
 
 
 def find_sublist(l, sublist):
@@ -36,7 +36,7 @@ def find_sublist(l, sublist):
     Returns
         int
     """
-    for i in xrange(len(l)):
+    for i in range(len(l)):
         # Check index 0 first for optimization
         if l[i] == sublist[0] and l[i: i + len(sublist)] == sublist:
             return i
@@ -59,7 +59,7 @@ def __init__(self, text):
         Args:
             text (str or unicode)
         """
-        self._text = unicode(text)
+        self._text = str(text)
         self._tokens = None
 
     @property

python/requirements.txt

@@ -1,3 +1,4 @@
-Pillow>=4.0
-selenium>=3.4.3
+Pillow
+selenium>=4.5.0
 pytest
+numpy