def AttachToBrowser(path, timeout):
"""Invoke the browser process and connect to the socket."""
(proc, frame, wnd) = scraper.GetBrowser(path)
if not wnd: raise ValueError("Could not invoke browser.")
# Try to connect the socket. If it fails, wait and try
# again. Do this for ten seconds
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM,
socket.IPPROTO_TCP)
for attempt in xrange(10):
try:
s.connect(("localhost", PORT))
except socket.error:
time.sleep(1)
continue
break
try:
s.getpeername()
except socket.error:
raise ValueError("Could not connect to browser")
if command["--size"]:
# Resize and reposition the frame
windowing.MoveAndSizeWindow(frame, (0, 0), command["--size"], wnd)
s.settimeout(timeout)
Iterate.proc = proc
Iterate.wnd = wnd
Iterate.s = s
def Scrape(urls, outdir, size, pos, timeout=20, **kwargs):
"""Invoke a browser, send it to a series of URLs, and save its output.
Args:
urls: list of URLs to scrape
outdir: directory to place output
size: size of browser window to use
pos: position of browser window
timeout: amount of time to wait for page to load
kwargs: miscellaneous keyword args
Returns:
None if success, else an error string
"""
path = r"c:\program files\internet explorer\iexplore.exe"
if "path" in kwargs and kwargs["path"]: path = kwargs["path"]
(iewnd, ieproc, address_bar, render_pane,
tab_window) = (InvokeBrowser(path))
# Resize and reposition the frame
windowing.MoveAndSizeWindow(iewnd, pos, size, render_pane)
# Visit each URL we're given
if type(urls) in types.StringTypes: urls = [urls]
timedout = False
for url in urls:
# Double-click in the address bar, type the name, and press Enter
mouse.DoubleClickInWindow(address_bar)
keyboard.TypeString(url)
keyboard.TypeString("\n")
# Wait for the page to finish loading
load_time = windowing.WaitForThrobber(tab_window, (6, 8, 22, 24),
timeout)
timedout = load_time < 0
if timedout:
break
# Scrape the page
image = windowing.ScrapeWindow(render_pane)
# Save to disk
if "filename" in kwargs:
if callable(kwargs["filename"]):
filename = kwargs["filename"](url)
else:
filename = kwargs["filename"]
else:
filename = windowing.URLtoFilename(url, outdir, ".bmp")
image.save(filename)
windowing.EndProcess(ieproc)
if timedout:
return "timeout"
def Time(urls, size, timeout, **kwargs):
"""Measure how long it takes to load each of a series of URLs
Args:
urls: list of URLs to time
size: size of browser window to use
timeout: amount of time to wait for page to load
kwargs: miscellaneous keyword args
Returns:
A list of tuples (url, time). "time" can be "crashed" or "timeout"
"""
if "path" in kwargs and kwargs["path"]: path = kwargs["path"]
else: path = DEFAULT_PATH
proc = None
# Visit each URL we're given
if type(urls) in types.StringTypes: urls = [urls]
ret = []
for url in urls:
try:
# Invoke the browser if necessary
if not proc:
(wnd, proc, address_bar, render_pane,
tab_window) = InvokeBrowser(path)
# Resize and reposition the frame
windowing.MoveAndSizeWindow(wnd, (0, 0), size, render_pane)
# Double-click in the address bar, type the name, and press Enter
mouse.DoubleClickInWindow(address_bar)
keyboard.TypeString(url)
keyboard.TypeString("\n")
# Wait for the page to finish loading
load_time = windowing.WaitForThrobber(tab_window, (6, 8, 22, 24),
timeout)
timedout = load_time < 0
if timedout:
load_time = "timeout"
# Send an alt-F4 to make the browser close; if this times out,
# we've probably got a crash
keyboard.TypeString(r"{\4}", use_modifiers=True)
if not windowing.WaitForProcessExit(proc, timeout):
windowing.EndProcess(proc)
load_time = "crashed"
proc = None
except pywintypes.error:
load_time = "crashed"
proc = None
ret.append((url, load_time))
# Send an alt-F4 to make the browser close; if this times out,
# we've probably got a crash
if proc:
keyboard.TypeString(r"{\4}", use_modifiers=True)
if not windowing.WaitForProcessExit(proc, timeout):
windowing.EndProcess(proc)
return ret
def Scrape(urls, outdir, size, pos, timeout=20, **kwargs):
"""Invoke a browser, send it to a series of URLs, and save its output.
Args:
urls: list of URLs to scrape
outdir: directory to place output
size: size of browser window to use
pos: position of browser window
timeout: amount of time to wait for page to load
kwargs: miscellaneous keyword args
Returns:
None if success, else an error string
"""
if "path" in kwargs and kwargs["path"]: path = kwargs["path"]
else: path = DEFAULT_PATH
(wnd, proc, render_pane) = InvokeBrowser(path)
# Resize and reposition the frame
windowing.MoveAndSizeWindow(wnd, pos, size, render_pane)
time.sleep(3)
# Firefox is a bit of a pain: it doesn't use standard edit controls,
# and it doesn't display a throbber when there's no tab. Let's make
# sure there's at least one tab, then select the first one
mouse.ClickInWindow(wnd)
keyboard.TypeString("[t]", True)
mouse.ClickInWindow(wnd, (30, 115))
time.sleep(2)
timedout = False
# Visit each URL we're given
if type(urls) in types.StringTypes: urls = [urls]
for url in urls:
# Use keyboard shortcuts
keyboard.TypeString("{d}", True)
keyboard.TypeString(url)
keyboard.TypeString("\n")
# Wait for the page to finish loading
load_time = windowing.WaitForThrobber(wnd, (10, 96, 26, 112), timeout)
timedout = load_time < 0
if timedout:
break
# Scrape the page
image = windowing.ScrapeWindow(render_pane)
# Save to disk
if "filename" in kwargs:
if callable(kwargs["filename"]):
filename = kwargs["filename"](url)
else:
filename = kwargs["filename"]
else:
filename = windowing.URLtoFilename(url, outdir, ".bmp")
image.save(filename)
# Close all the tabs, cheesily
mouse.ClickInWindow(wnd)
while len(windowing.FindChildWindows(0, "MozillaUIWindowClass")):
keyboard.TypeString("[w]", True)
time.sleep(1)
if timedout:
return "timeout"
def Time(urls, size, timeout, **kwargs):
"""Measure how long it takes to load each of a series of URLs
Args:
urls: list of URLs to time
size: size of browser window to use
timeout: amount of time to wait for page to load
kwargs: miscellaneous keyword args
Returns:
A list of tuples (url, time). "time" can be "crashed" or "timeout"
"""
if "path" in kwargs and kwargs["path"]: path = kwargs["path"]
else: path = DEFAULT_PATH
proc = None
# Visit each URL we're given
if type(urls) in types.StringTypes: urls = [urls]
ret = []
for url in urls:
try:
# Invoke the browser if necessary
if not proc:
(wnd, proc, render_pane) = InvokeBrowser(path)
# Resize and reposition the frame
windowing.MoveAndSizeWindow(wnd, (0, 0), size, render_pane)
time.sleep(3)
# Firefox is a bit of a pain: it doesn't use standard edit controls,
# and it doesn't display a throbber when there's no tab. Let's make
# sure there's at least one tab, then select the first one
mouse.ClickInWindow(wnd)
keyboard.TypeString("[t]", True)
mouse.ClickInWindow(wnd, (30, 115))
time.sleep(2)
# Use keyboard shortcuts
keyboard.TypeString("{d}", True)
keyboard.TypeString(url)
keyboard.TypeString("\n")
# Wait for the page to finish loading
load_time = windowing.WaitForThrobber(wnd, (10, 96, 26, 112),
timeout)
timedout = load_time < 0
if timedout:
load_time = "timeout"
# Try to close the browser; if this fails it's probably a crash
mouse.ClickInWindow(wnd)
count = 0
while (len(
windowing.FindChildWindows(0, "MozillaUIWindowClass"))
and count < 5):
keyboard.TypeString("[w]", True)
time.sleep(1)
count = count + 1
if len(windowing.FindChildWindows(0, "MozillaUIWindowClass")):
windowing.EndProcess(proc)
load_time = "crashed"
proc = None
except pywintypes.error:
proc = None
load_time = "crashed"
ret.append((url, load_time))
if proc:
count = 0
while (len(windowing.FindChildWindows(0, "MozillaUIWindowClass"))
and count < 5):
keyboard.TypeString("[w]", True)
time.sleep(1)
count = count + 1
return ret
def Scrape(urls, outdir, size, pos, timeout, kwargs):
"""Invoke a browser, send it to a series of URLs, and save its output.
Args:
urls: list of URLs to scrape
outdir: directory to place output
size: size of browser window to use
pos: position of browser window
timeout: amount of time to wait for page to load
kwargs: miscellaneous keyword args
Returns:
None if success, else an error string
"""
if "path" in kwargs and kwargs["path"]: path = kwargs["path"]
else: path = DEFAULT_PATH
(wnd, proc, address_bar, render_pane) = InvokeBrowser(path)
# Resize and reposition the frame
windowing.MoveAndSizeWindow(wnd, pos, size, render_pane)
# Visit each URL we're given
if type(urls) in types.StringTypes: urls = [urls]
timedout = False
for url in urls:
# Double-click in the address bar, type the name, and press Enter
mouse.ClickInWindow(address_bar)
keyboard.TypeString(url, 0.1)
keyboard.TypeString("\n")
# Wait for the page to finish loading
load_time = windowing.WaitForThrobber(wnd, (20, 16, 36, 32), timeout)
timedout = load_time < 0
if timedout:
break
# Scrape the page
image = windowing.ScrapeWindow(render_pane)
# Save to disk
if "filename" in kwargs:
if callable(kwargs["filename"]):
filename = kwargs["filename"](url)
else:
filename = kwargs["filename"]
else:
filename = windowing.URLtoFilename(url, outdir, ".bmp")
image.save(filename)
if proc:
windowing.SetForegroundWindow(wnd)
# Send Alt-F4, then wait for process to end
keyboard.TypeString(r"{\4}", use_modifiers=True)
if not windowing.WaitForProcessExit(proc, timeout):
windowing.EndProcess(proc)
return "crashed"
if timedout:
return "timeout"
return None
