def LoadImagesFromBrowser():
global bingSearchCache, bgCurrentQuery
print("Loading browser results..")
ScrollBingPage(1)
time.sleep(scroll_wait)
elist = None
try:
bingSearchCache[bgCurrentQuery] = None
elist = GetBingResultElements()
results = {}
for e in elist:
try:
image = ProcessBingImageElement(e)
index = e["data-idx"]
results[str(index)] = image
except Exception as e:
view_traceback()
# Save Images to in-memory cache
# TODO : Convert to simple storage using sqlalchemy and sqllite3
bingSearchCache[bgCurrentQuery] = results
rescount = str(len(results))
print(rescount + " no. of images loaded onto cache for : " +
bgCurrentQuery)
return rescount
except Exception as e:
view_traceback()
return "-1"
def ScrollBingPage(nPages):
# Scrolls a Bing Image search by a defined
# number of pages.
print("Scrolling target page.")
global bgMoreResultsContainer, bgMoreResultsButton
showMoreContainer = None
showMoreButton = None
try:
# Need to scroll once to ensure presence of indicating element
# quirk observed in Bing page.
browser.execute_script(
"window.scrollTo(0, document.body.scrollHeight);")
time.sleep(scroll_wait)
if nPages > 2:
nPages = nPages - 2
for i in range(nPages):
showMoreContainer = browser.find_element_by_css_selector(
bgMoreResultsContainer)
cls = showMoreContainer.get_attribute("class")
if "b_hide" in cls or cls == "":
# Button is hidden. Scrolling can be made directly
browser.execute_script(
"window.scrollTo(0, document.body.scrollHeight);")
else:
showMoreButton = browser.find_element_by_css_selector(
bgMoreResultsButton)
showMoreButton.click()
return True
else:
return True
except Exception as e:
view_traceback()
return False
def GetBingSearchImages(n, offset):
global bingSearchCache, bgCurrentQuery
imageCount = None
try:
imageCount = len(bingSearchCache[bgCurrentQuery])
if imageCount == 0 or imageCount is None:
LoadBingSearchPage(bgCurrentQuery)
except Exception as e:
imageCount = 0
return "404"
# Cleaning the input
if n <= 0:
n = 1
if offset <= 0:
offset = imageCount
if offset >= imageCount:
# If asking for more than available images, return a default value.
offset = imageCount - 1
if n + offset > imageCount:
ScrollBingPage(2)
# Update the results in the browser.
LoadBingSearchResults(bgCurrentQuery)
imageCount = len(bingSearchCache[bgCurrentQuery])
#n = imageCount-(offset)
try:
if bingSearchCache:
results = bingSearchCache[bgCurrentQuery]
idxrange = range(n, n + offset + 1)
apiresults = []
# Note sometimes, keys may be missing
for key in idxrange:
try:
apiresults.append(results[str(key)])
except KeyError as ke:
pass
jsonresponse = "{\n"
for apiresult in apiresults:
id = apiresult.id
image = apiresult
jsonresponse += "\"" + id + "\"" + ": "
jsonresponse += JSONEncoder().encode({
"base64": image.base64,
"url": image.url,
"width": image.width,
"height": image.height,
})
jsonresponse += ","
jsonresponse = jsonresponse[:-1]
jsonresponse += "}\n"
return jsonresponse
else:
return "500"
except Exception as e:
view_traceback()
return "501"
def InitializeBrowser():
global browser,logpath,pjspath,wd
profile_name = "SELENIUM"
geckopath = os.path.join(wd,"geckodriver.exe")
try:
browser = webdriver.Firefox(log_path=logpath, executable_path=geckopath,firefox_profile=ffprofilepath)
browser.implicitly_wait(3)
#browser = webdriver.PhantomJS(logpath=logpath)
#browser.implicitly_wait(10)
except Exception as e:
print("Error starting PhantomJS")
view_traceback()
def LoadBingSearchResults(query):
global maxResults, bgResultsPerRow
#ScrollBingPage(2)
elements = GetBingResultElements()
results = {}
for e in elements:
try:
image = ProcessBingImageElement(e)
index = e["data-idx"]
results[str(index)] = image
except Exception as e:
print("LoadBingSearchResults Error : " + e)
view_traceback()
# Save Images to in-memory cache
# TODO : Convert to simple storage using sqlalchemy and sqllite3
bingSearchCache[query] = results
rescount = str(len(results))
return rescount