Selenium

An automated browser engine -- useful for things like scraping and browser based installations. Has been used in projects like PlaySureVeillance and In The Company Of Bots.

• http://docs.seleniumhq.org/

Using with python

To use selenium with python you just need the selenium python library (installable with pip) and the "webdriver" for your specific browser (for instance: firefox or chrome (maybe chromium, though haven't tested that).

from __future__ import print_function
from selenium import webdriver
from selenium.webdriver.support.ui import Select
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
from time import sleep
from selenium.common.exceptions import NoSuchElementException, NoSuchWindowException
from urllib2 import urlopen
import sys, json, os
from argparse import ArgumentParser


BUFSIZE  = 1024 * 1000
def wget (url, tofile):
    f = urlopen(url)
    count = 0
    with open(tofile, "wb") as fout:
        while True:
            data = f.read(BUFSIZE)
            if data == "":
                break
            count += len(data)
            fout.write(data)
    return count

def image_path (x):
    x = x.replace(" ", "_").replace("/", "-").lower()
    x = x+".jpg"
    return x

def log (*msg):
    print (*msg, file=sys.stderr)


ap = ArgumentParser("MRAH scaper")
ap.add_argument("--starturl", default="http://carmentis.be")
ap.add_argument("--browser", choices=("firefox", "chrome", "opera"), default="chrome", help="browser driver: firefox (default), chrome, opera")
ap.add_argument("--format", choices=("json", "csv"), default="csv", help="output format: json (default), csv")
ap.add_argument("--imagepath", default="images")
ap.add_argument("--skipimages", action="store_true", default=False)
ap.add_argument("--limit", type=int, default=None)
ap.add_argument("--sleeptime", type=float, default=None, help="sleeptime")

args = ap.parse_args()
if not args.skipimages:
    try:
        os.makedirs(args.imagepath)
    except OSError:
        pass

sleeptime = args.sleeptime
log("Opening browser...")
driver = None
if args.browser == "opera":
    b = webdriver.Opera()
elif args.browser == "chrome":
    b = webdriver.Chrome()
else:
    b = webdriver.Firefox()
    if sleeptime == None:
        sleeptime = 0.5

b.get(args.starturl)

log("Perform a search and select detail mode, then press enter to start scraping items... (Ctrl-c to cancel)")
raw_input()

props = """
collectionName
inventoryNb
objectName
objectTitle
objectCulture
geography
dating
material
technique
dimensions
legalRightOwner
""".strip().splitlines()

if args.format == "csv":
    from csv import DictWriter
    fieldnames = props[:]
    fieldnames.append("url")
    if not args.skipimages:
        fieldnames.extend(("imageurl", "image"))
    csvout = DictWriter(sys.stdout, fieldnames=fieldnames)
    csvout.writeheader()

count = 0

while True:
    b.implicitly_wait(0)
    count += 1
    item = {}
    for p in props:
        try:
            li = b.find_element_by_css_selector("li."+p)
            name = li.find_element_by_css_selector(".tspPrefix")
            span = li.find_element_by_css_selector(".tspValue")
            # item[name.text] = span.text
            item[p] = span.text
        except NoSuchElementException as e:
            pass

    # permalink / bookmark
    tries = 0
    while tries < 5:
        try:
            permalink = b.find_element_by_css_selector("li.bookmark")\
                .find_element_by_css_selector("input")\
                .get_attribute("value")
            item['url'] = permalink
            break
        except NoSuchElementException:
            tries += 1
            sleep(0.1)
    # print ("PERMALINK: {0}".format(permalink))

    if not args.skipimages:
        imglink = b.find_element_by_css_selector("dt.detailImg a")
        imglink.click()


        img_src = None
        tries = 0
        while tries < 5:
            try:
                b.switch_to_window('HighResImage')
                img = b.find_element_by_css_selector("img")
                img_src = img.get_attribute("src")
                item['imageurl'] = img_src
                b.close()
                b.switch_to_window("")
                break
            except NoSuchWindowException as e:
                log("NoSuchWindowException", tries)
                sleep(1)
                tries += 1
        

        if img_src:
            # log("IMAGE: {0}".format(img_src))
            ifilename = image_path(item['inventoryNb'])
            ipath = os.path.join(args.imagepath, ifilename)
            if wget(img_src, ipath):
                item['image'] = ifilename

    if args.format == "json":
        print (json.dumps(item))
    elif args.format == "csv":
        csvout.writerow({k:v.encode('utf8') for k,v in item.items()})

    log(u"{0} {1}/{2}".format(count, item.get('objectName', u''), item.get('objectTitle', u'')).encode("utf-8"))

    next = None
    try:
        next = b.find_element_by_css_selector('#pageSetEntries-nextSet a')
    except NoSuchElementException as e:
        pass
    if next == None:
        log("END OF LIST")
        break
        
    b.implicitly_wait(10)
    next.click()
    if sleeptime:
        sleep(sleeptime)

log("output {0} items".format(count))
b.close()