def scrape_data(user_id, scan_list, section, elements_path, save_status, file_names):
"""Given some parameters, this function can scrap friends/photos/videos/about/posts(statuses) of a profile"""
page = []
if save_status == 4:
page.append(user_id)
page += [user_id + s for s in section]
for i, _ in enumerate(scan_list):
# try:
driver.get(page[i])
if save_status != 3:
utils.scroll(total_scrolls, driver, selectors, scroll_time, dbid)
data = bs(driver.page_source, 'lxml').find_all('div', attrs={"class":"du4w35lb k4urcfbm l9j0dhe7 sjgh65i0"})
if len(data) == 0:
driver.refresh()
time.sleep(0.5)
driver.find_element_by_xpath("//a[contains(text(),'Timeline')]").click()
time.sleep(0.5)
driver.find_element_by_xpath("//a[contains(text(),'Timeline')]").click()
time.sleep(3)
data = bs(driver.page_source, 'lxml').find_all('div', attrs={"class":"du4w35lb k4urcfbm l9j0dhe7 sjgh65i0"})
save_to_file(file_names[i], data, save_status, i)
def getTrends(driver):
driver.get("https://www.youtube.com/feed/trending")
utils.scroll(driver, numScrolls=20)
soup = BeautifulSoup(driver.page_source, "html.parser")
vids = []
for k in soup.find_all('div', id='grid-container'):
vids += k.find_all('ytd-video-renderer')
videos = utils.getVideoFromSearch(vids)
return videos
def downloadImagesByTag(tag, maxImages, isClosing):
driver = webdriver.Firefox()
driver.get("https://www.instagram.com/")
utils.login(driver, username, password)
searchByTag(driver, tag, 1)
utils.scroll(driver, 1)
#target all the link elements on the page
numberOfImagesDownloaded = download.downloadImages(driver, maxImages, tag)
print(f'Downloaded {numberOfImagesDownloaded} images')
if isClosing:
utils.end(driver, 1)
def scrape_data(user_id, scan_list, section, elements_path, save_status,
file_names):
"""Given some parameters, this function can scrap friends/photos/videos/about/posts(statuses) of a profile"""
page = []
if save_status == 4:
page.append(user_id)
page += [user_id + s for s in section]
for i, _ in enumerate(scan_list):
try:
driver.get(page[i])
if ((save_status == 0) or (save_status == 1)
or (save_status
== 2)): # Only run this for friends, photos and videos
# the bar which contains all the sections
sections_bar = driver.find_element_by_xpath(
selectors.get("sections_bar"))
if sections_bar.text.find(scan_list[i]) == -1:
continue
if save_status != 3:
utils.scroll(total_scrolls, driver, selectors, scroll_time,
dbid)
data = bs(driver.page_source, 'lxml').find_all(
'div', attrs={"class": "du4w35lb k4urcfbm l9j0dhe7 sjgh65i0"})
if len(data) == 0:
driver.refresh()
time.sleep(0.5)
driver.find_element_by_xpath(
"//a[contains(text(),'Timeline')]").click()
time.sleep(0.5)
driver.find_element_by_xpath(
"//a[contains(text(),'Timeline')]").click()
time.sleep(3)
data = bs(driver.page_source, 'lxml').find_all(
'div',
attrs={"class": "du4w35lb k4urcfbm l9j0dhe7 sjgh65i0"})
save_to_file(file_names[i], data, save_status, i)
except Exception:
print(
"Exception (scrape_data)",
str(i),
"Status =",
str(save_status),
sys.exc_info()[0],
)
def scrape_data(url, scan_list, section, elements_path, save_status,
file_names):
"""Given some parameters, this function can scrap friends/photos/videos/about/posts(statuses) of a profile"""
page = []
if save_status == 4 or save_status == 5:
page.append(url)
page += [url + s for s in section]
for i, _ in enumerate(scan_list):
try:
driver.get(page[i])
if ((save_status == 0) or (save_status == 1)
or (save_status
== 2)): # Only run this for friends, photos and videos
# the bar which contains all the sections
sections_bar = driver.find_element_by_xpath(
selectors.get("sections_bar"))
if sections_bar.text.find(scan_list[i]) == -1:
continue
if save_status != 3:
utils.scroll(total_scrolls, driver, selectors, scroll_time)
pass
data = driver.find_elements_by_xpath(elements_path[i])
save_to_file(file_names[i], data, save_status, i)
except Exception:
print(
"Exception (scrape_data)",
str(i),
"Status =",
str(save_status),
sys.exc_info()[0],
)
def scrape_data(url, scan_list, section, elements_path, save_status,
file_names):
"""Given some parameters, this function can scrap friends/photos/videos/about/posts(statuses) of a profile"""
page = []
pos = -1
if url[-1] == '/':
pos = -2
user_name = url.split("/")[pos]
print("user_name: ", user_name)
if save_status == 4 or save_status == 5:
page.append(url)
page += [url + s for s in section]
for i, _ in enumerate(scan_list):
try:
driver.get(page[i])
if save_status != 3:
utils.scroll(total_scrolls, driver, selectors, scroll_time)
pass
data = driver.find_elements_by_xpath(elements_path[i])
save_to_file(file_names[i], data, save_status, i, user_name)
except Exception:
print(
"Exception (scrape_data)",
str(i),
"Status =",
str(save_status),
sys.exc_info()[0],
)
def get_all_page_data(url, is_community=False):
name = url.split("/")[-1] if len(
url.split("/")[-1]) > 0 else url.split("/")[-2]
if is_community:
name = os.path.join(name, "community")
url = url + "/community"
data_path = os.path.join(".", "data")
if not os.path.exists(data_path):
os.mkdir(data_path)
page_data_path = os.path.join(data_path, name)
if not os.path.exists(page_data_path):
os.mkdir(page_data_path)
should_scrape_headless = is_community == False
driver = initialize_driver(args.chrome,
args.windows,
is_headless=should_scrape_headless)
driver.get(url)
page_name = get_text(driver, './/a[@class="_64-f"]')
print(f"Scrolling {url} until {cutoff_date}")
scroll(driver, pd.to_datetime(cutoff_date))
posts = driver.find_elements_by_xpath(
'//div[contains(@class, "userContentWrapper")]')
post_links = [get_post_links(post) for post in tqdm(posts)]
post_links = list(set(post_links))
with open(os.path.join(page_data_path, 'post_links.json'), 'w') as f:
json.dump(post_links, f)
driver.quit()
print(f"Now scraping {len(post_links)} posts from {name}")
for i, post_link in enumerate(post_links):
if not is_string_url(post_link):
continue
print(f"Scraping {post_link}")
driver = initialize_driver(args.chrome, args.windows)
driver.get(post_link)
if "/videos/" in post_link:
post_type = "videos"
elif "/photos/" in post_link:
post_type = "photos"
elif "/posts/" in post_link:
post_type = "posts"
elif "/notes/" in post_link:
post_type = "notes"
else:
post_type = "other"
if post_type == "notes":
post_element = driver.find_element_by_xpath(
'.//div[contains(@class, "fb_content")]')
else:
post_element = driver.find_element_by_xpath(
'.//div[contains(@class, "userContentWrapper")]')
post_data = get_post_data(driver, post_element, post_type)
post_data["page_name"] = page_name
with open(os.path.join(page_data_path, f'page_post_{i}.json'),
'w') as f:
json.dump(post_data, f)
driver.quit()
if not is_community:
get_all_page_data(url, is_community=True)
def start_crawler(url, email, password, depth):
#Check if page has loaded
def page_loaded(driver):
return driver.find_element_by_tag_name("body") != None
# Log and sign into twitter - definde or redefine driver
def log_in():
# Opening the web browser and the twitter page
driver = webdriver.Chrome(
executable_path=
r'F:\Dokumentation\Programme\ChromeDriver\chromedriver.exe')
driver.get(url)
wait = ui.WebDriverWait(driver, 10)
wait.until(page_loaded)
#Signing in
actions = ActionChains(driver)
actions.send_keys(email)
actions.send_keys(Keys.TAB)
actions.send_keys(password)
actions.send_keys(Keys.ENTER)
actions.perform()
time.sleep(5)
return driver
# Opening the web browser and the twitter page
driver = log_in()
# Get first node
html_doc = driver.page_source
soup = BeautifulSoup(html_doc, 'html.parser')
# Create Anchor
id = len(models.GephiNode.objects.all()) + 1
print("id: ", id)
label = soup.find(
"a", {
"class": "ProfileHeaderCard-nameLink u-textInheritColor js-nav"
}).get_text()
print("label: ", label)
try:
fan_count = soup.findAll("span",
{"class": "ProfileNav-value"})[2].get_text()
except:
fan_count = 0
print("fan_count: ", fan_count)
handle = soup.find("b", {"class": "u-linkComplex-target"}).get_text()
print("handle: ", handle)
Anchor = models.GephiNode(id=id,
label=label,
fan_count=fan_count,
handle=handle)
Anchor.create()
print("Anchor: ", Anchor.id, Anchor.label)
# Go to following following
driver.get(url + "/following")
wait = ui.WebDriverWait(driver, 10)
wait.until(page_loaded)
try:
lnks_cnt = int(
soup.findAll("span",
{"class": "ProfileNav-value"})[1].get_text().replace(
".", ""))
except:
lnks_cnt = 1500
print(lnks_cnt)
links = driver.find_elements_by_xpath(
"//a[@class='ProfileCard-bg js-nav']")
print(len(links))
test = 1
while len(links) < lnks_cnt - 2 and test < 1500:
utils.scroll(driver)
links = driver.find_elements_by_xpath(
"//a[@class='ProfileCard-bg js-nav']")
print("following: ", lnks_cnt)
print("links: ", len(links))
test += 1
utils.scroll(driver)
links = driver.find_elements_by_xpath(
"//a[@class='ProfileCard-bg js-nav']")
newUrlList = list()
for lnk in links:
newUrlList.append(lnk.get_attribute("href"))
print("links unshuffeld: ", newUrlList)
shuffle(newUrlList)
print("links shuffeld: ", newUrlList)
# Check everyone the Anchor follows
print("link count: ", len(links))
print("edge count: ",
len(models.GephiEdge.objects.filter(source=Anchor.id)))
if len(newUrlList) > len(
models.GephiEdge.objects.filter(source=Anchor.id)):
#First Followed
for link in newUrlList:
if link == "https://twitter.com/account/suspended":
newUrlList = [
x for x in newUrlList
if x != "https://twitter.com/account/suspended"
]
else:
try:
driver.get(link)
except:
driver = log_in()
driver.get(link)
# Get HTML
wait = ui.WebDriverWait(driver, 10)
wait.until(page_loaded)
html_doc = driver.page_source
soup = BeautifulSoup(html_doc, 'html.parser')
#Create Node
id = len(models.GephiNode.objects.all()) + 1
print("id: ", id)
label = soup.find(
"a", {
"class":
"ProfileHeaderCard-nameLink u-textInheritColor js-nav"
}).get_text()
print("label: ", label)
try:
fan_count = soup.findAll(
"span", {"class": "ProfileNav-value"})[2].get_text()
except:
fan_count = 0
print("fan_count: ", fan_count)
handle = soup.find("b", {
"class": "u-linkComplex-target"
}).get_text()
print("handle: ", handle)
tempNode = models.GephiNode(id=id,
label=label,
fan_count=fan_count,
handle=handle)
tempNode.create()
#Create Edge
print("source: ", Anchor.id)
print("target: ", tempNode.id)
print("type: ", "Directed")
edge_id = len(models.GephiEdge.objects.all())
print("id: ", edge_id)
firstEdge = models.GephiEdge(source=Anchor.id,
target=tempNode.id,
type="Directed",
id=edge_id,
weight=1)
firstEdge.create()
# Break if list is complete in DB
print("link count: ", len(links))
print("edge count: ",
len(models.GephiEdge.objects.filter(source=Anchor.id)))
if len(newUrlList) == len(
models.GephiEdge.objects.filter(source=Anchor.id)):
break
print("Nodes: ", len(models.GephiNode.objects.all()))
print("Edges: ", len(models.GephiEdge.objects.all()))
print(
"Difference: ",
len(models.GephiEdge.objects.all()) -
len(models.GephiNode.objects.all()))
driver.close()
# Rekcursion
print("depth: ", depth)
if depth > 0:
for uri in newUrlList:
try:
start_crawler(uri, email, password, depth - 1)
except:
try:
start_crawler(uri, email, password, depth - 1)
except:
pass
def open_newsfeed(self):
self.driver.get(self.url)
scroll(self.driver, 5)
def twi_hashtag_originator(hashtag, minYear=2006):
# cache the hashtag for future use (screenshot)
cache = "_".join(hashtag.split(" ")).replace("#", "")
# checking wether argument is valid
if hashtag is None or hashtag == "":
print("Wrong argument: ", hashtag)
return None
# separate the words and remove #
h = [j for j in hashtag.split(" ") if j != ""]
check = lambda x: x[1:] if x.startswith("#") else x
h = [check(j).lower() for j in h]
# forming the URL part
hashtag = f"(%23{h[0]}"
for rest in h[1:]:
hashtag += f" AND %23{rest}"
hashtag += ")"
# preliminary code
soup = BeautifulSoup("", "html.parser")
createStorage("twitter")
screen = False
# close driver as soon as work done
with setDriver(headless=False) as driver:
wait = WebDriverWait(driver, 3)
year = minYear
# checking if hashtag is actually ever been used
url = f"https://twitter.com/search?q={hashtag}"
driver.get(url)
articles = returnAllTweets(wait)
if len(articles) == 0:
print("Hashtag has not been used in any tweets that are public")
return None
# check starting from minYear
articles = []
while (len(articles) == 0):
url = f"https://twitter.com/search?q={hashtag}%20until%3A{str(year)}-01-01&src=typed_query"
driver.get(url)
articles = returnAllTweets(wait)
year += 1
scroll(driver, fastScroll=False)
ti = returnAllTweets(wait)
if len(ti) == 0:
print("Something went wrong.")
return None
# remove tweets not containing the hashtag (This occurs if the tweet is a parent tweet of the hashtag containing tweet)
t = []
for post in ti:
inner = post.get_attribute("innerHTML").lower()
good = True
for each in h:
if f">#{each}</a>" not in inner:
good = False
break
if good:
t.append(post)
# sort all posts by retrieving the timestamp
timed = []
for post in t:
try:
s = post.find_element_by_xpath('.//time')
s = s.get_attribute("datetime")
if s:
try:
timed.append({
"timestamp":
datetime.datetime.strptime(
s, "%Y-%m-%dT%H:%M:%S.%fZ"),
"elem":
post
})
except:
pass
except NoSuchElementException:
pass
t = sorted(timed, key=lambda x: x["timestamp"])
# element of earliest post selected and screenshot taken
if len(t) == 0:
print("Something is seriously wrong")
return None
t = t[0]["elem"]
try:
with open(f"data/twitter/{cache}.png", "wb") as filex:
filex.write(t.screenshot_as_png)
screen = True
except:
print("Screenshot was unsuccessfull")
t = t.get_attribute("innerHTML")
soup = BeautifulSoup(t, "html.parser")
result = {
"poster": {
"full_name": "",
"username": "",
"url": "",
"id": "",
"profile_image_url": ""
},
"timestamp": "",
"post_link": "",
"post_text": "",
"embed": "",
"screenshot": ""
}
a = soup.find("a")
if a:
result["poster"]["url"] = completeTwitterLink(a.get('href', ""))
result["poster"]["username"] = result["poster"]["url"].replace(
"https://www.twitter.com/", "")
k = a.findNext("a", {"href": a.get("href", "")})
if k:
result["poster"]["full_name"] = k.text.partition("@")[0]
k = a.find("img")
if k:
result["poster"]["profile_image_url"] = k.get("src", "")
a = soup.find(lambda tag: tag.name == 'a' and tag.find("time"))
if a:
link = completeTwitterLink(a.get("href", ""))
result["post_link"] = link
result[
"embed"] = f'<blockquote class="twitter-tweet"><a href="{link}"></a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>'
k = a.find("time")
if k:
result["timestamp"] = k.get("datetime", "")
soup = BeautifulSoup(str(soup).partition(str(a))[2], "html.parser")
result["post_text"] = " ".join([s.text for s in soup.find_all("span")])
if screen:
result["screenshot"] = os.getcwd() + f"/data/twitter/{cache}"
return result
def getPosts(self):
# navigating to page
newUrl = self.url + "/community"
if self.driver.current_url != newUrl:
self.driver.get(newUrl)
utils.scroll(self.driver)
sleep(1)
content = str(self.driver.page_source)
# open comments
comm = re.findall(r'View all [0-9]+ comments', content)
comm += re.findall(r'View comment', content)
for co in comm:
elements = self.driver.find_elements_by_xpath(
"//*[text()[contains(.,'" + co + "')]]")
for element in elements:
self.driver.execute_script("arguments[0].click();", element)
try:
self.wait.until(
EC.visibility_of_all_elements_located(
(By.TAG_NAME, 'ytd-comment-thread-renderer')))
except TimeoutException:
pass
sleep(1)
soup = BeautifulSoup(self.driver.page_source, "html.parser")
blocks = soup.find_all('ytd-backstage-post-thread-renderer')
posts = []
# extracting post one by one
for block in blocks:
post = self.initiatePostSkeleton()
# extracting author link and picture
a = block.find('div', id='author-thumbnail')
if a:
a = a.find('a')
if a:
post["post_author"]["link"] = utils.completeYoutubeLink(
a.get('href', ''))
img = a.find('img')
if img:
post["post_author"]["media_directory"] = img.get(
'src', "")
# extracting author name
a = block.find('a', id='author-text')
if a:
post["post_author"]["name"] = utils.beautifyText(a.text)
# extracting post link and pubish time
a = block.find('yt-formatted-string', id='published-time-text')
if a:
a = a.find('a')
if a:
post["timestamp"] = a.text
post["post_link"] = utils.completeYoutubeLink(
a.get('href', ''))
# extracting text
a = block.find('yt-formatted-string', id='content-text')
if a:
post["post_type"]["is_text"] = True
post["post_text"] = a.text.replace('\ufeff', '')
# extracting vote data
v = block.find('ytd-backstage-poll-renderer', id='poll-attachment')
if v:
a = v.find('yt-formatted-string', id='vote-info')
if a:
count = a.text.split()[0]
if count.isnumeric():
post["votes"]["total_votes"] = int(count)
post["post_type"]["is_vote"] = True
post["votes"]["vote_text"] = post["post_text"]
a = v.find_all('yt-formatted-string',
{'class': 'choice-text'})
for choice in a:
post["votes"]["vote_options"].append(choice.text)
# getting likes
a = block.find('span', id='vote-count-middle')
if a:
count = utils.beautifyText(a.text)
if count.isnumeric():
post["post_likes"] = int(count)
# getting links
a = []
srcs = []
b = block.find('div', id='content')
c = block.find('div', id='content-attachment')
for y in [b, c]:
if y:
a += y.find_all('a')
srcs += y.find_all('img')
for link in a:
link = utils.completeYoutubeLink(link.get('href', ""))
if "/watch?" in link:
post["post_type"]["is_video"] = True
post["video_links"].append(link)
elif link not in [
post["post_link"], self.url, '', self.origin
]:
post["other_links"].append(link)
for img in srcs:
post["post_type"]["is_picture"] = True
src = img.get('src', '')
if src != '':
post["picture_links"].append(src)
# removing duplicate links
for option in ["picture_links", "video_links", "other_links"]:
post[option] = list(set(post[option]))
# getting comments
for com in block.find_all('ytd-comment-thread-renderer'):
comment = self.initiateCommentSkeleton()
a = com.find('div', id='author-thumbnail')
if a:
a = a.find('a')
if a:
comment["comment_author"][
"link"] = utils.completeYoutubeLink(
a.get('href', ''))
img = a.find('img')
if img:
comment["comment_author"][
"media_directory"] = img.get('src', "")
a = com.find('a', id='author-text')
if a:
comment["comment_author"]["name"] = utils.beautifyText(
a.text)
a = com.find('yt-formatted-string',
{'class': 'published-time-text'})
if a:
a = a.find('a')
if a:
comment["timestamp"] = a.text
comment["comment_url"] = utils.completeYoutubeLink(
a.get('href', ''))
a = com.find('yt-formatted-string', id='content-text')
if a:
comment["comment_text"] = a.text.replace('\ufeff', '')
for link in a.find_all('a'):
comment["comment_links"].append(
utils.completeYoutubeLink(link.get('href', '')))
a = com.find('span', id='vote-count-middle')
if a:
count = utils.beautifyText(a.text)
if count.isnumeric():
comment["comment_likes"] = int(count)
post["post_comments"].append(comment)
posts.append(post)
self.profile["posts"] = posts
return posts
