|
@ -3,105 +3,71 @@ from .archive import archiver |
|
|
from .common import SELENIUM_EXCEPTIONS, click_button, wait_xpath, force_mobile |
|
|
from .common import SELENIUM_EXCEPTIONS, click_button, wait_xpath, force_mobile |
|
|
from .config import settings |
|
|
from .config import settings |
|
|
from selenium.webdriver.common.action_chains import ActionChains |
|
|
from selenium.webdriver.common.action_chains import ActionChains |
|
|
|
|
|
from calendar import month_name |
|
|
|
|
|
|
|
|
|
|
|
months = [m for m in month_name if m] |
|
|
|
|
|
|
|
|
import time |
|
|
import time |
|
|
|
|
|
|
|
|
# Used as a threshold to avoid running forever |
|
|
# Used as a threshold to avoid running forever |
|
|
MAX_POSTS = settings["MAX_POSTS"] |
|
|
MAX_POSTS = settings["MAX_POSTS"] |
|
|
|
|
|
|
|
|
def delete_activity(driver, |
|
|
def get_load_more(driver): |
|
|
year=None): |
|
|
|
|
|
""" |
|
|
""" |
|
|
Deletes or hides all activity related to posts. |
|
|
Click the "Load more from X" button repeatedly |
|
|
|
|
|
|
|
|
Args: |
|
|
|
|
|
driver: seleniumrequests.Chrome Driver instance |
|
|
|
|
|
year: optional int YYYY year |
|
|
|
|
|
""" |
|
|
""" |
|
|
|
|
|
|
|
|
driver.get("https://m.facebook.com/allactivity/?category_key=statuscluster") |
|
|
button_expr = f"//div[contains(text(), 'Load more from')]" |
|
|
|
|
|
|
|
|
time.sleep(1000) |
|
|
|
|
|
|
|
|
|
|
|
finished = False |
|
|
|
|
|
|
|
|
|
|
|
with archiver("wall") as archive_wall_post: |
|
|
|
|
|
for _ in range(MAX_POSTS): |
|
|
|
|
|
if finished: |
|
|
|
|
|
break |
|
|
|
|
|
post_button_sel = "_4s19" |
|
|
|
|
|
|
|
|
|
|
|
post_content_sel = "userContent" |
|
|
|
|
|
post_timestamp_sel = "timestampContent" |
|
|
|
|
|
|
|
|
|
|
|
confirmation_button_exp = "//div[contains(@data-sigil, 'undo-content')]//*/a[contains(@href, 'direct_action_execute')]" |
|
|
|
|
|
|
|
|
|
|
|
# Cannot return a text node, so it returns the parent. |
|
|
|
|
|
# Tries to be pretty resilient against DOM re-organizations |
|
|
|
|
|
timestamp_exp = "//article//*/header//*/div/a[contains(@href, 'story_fbid')]//text()/.." |
|
|
|
|
|
|
|
|
|
|
|
button_types = ["Delete post", "Remove tag", "Hide from timeline"] |
|
|
print("Trying to load more") |
|
|
|
|
|
|
|
|
while True: |
|
|
while True: |
|
|
try: |
|
|
try: |
|
|
try: |
|
|
wait_xpath(driver, button_expr) |
|
|
timeline_element = driver.find_element_by_xpath("//div[@data-sigil='story-popup-causal-init']/a") |
|
|
driver.find_element_by_xpath(button_expr).click() |
|
|
except SELENIUM_EXCEPTIONS: |
|
|
except SELENIUM_EXCEPTIONS: |
|
|
print("Could not find any posts") |
|
|
|
|
|
finished = True |
|
|
|
|
|
break |
|
|
break |
|
|
|
|
|
|
|
|
post_content_element = driver.find_element_by_xpath("//article/div[@class='story_body_container']/div") |
|
|
def get_timeslices(driver): |
|
|
post_content_ts = driver.find_element_by_xpath(timestamp_exp) |
|
|
""" |
|
|
|
|
|
Get a list of the time slices Facebook is going to let us click. |
|
|
if not (post_content_element or post_content_ts): |
|
|
""" |
|
|
break |
|
|
|
|
|
|
|
|
|
|
|
# Archive the post |
|
|
slice_expr = "//header" |
|
|
archive_wall_post.archive( |
|
|
|
|
|
Post( |
|
|
|
|
|
content=post_content_element.text, |
|
|
|
|
|
date=post_content_ts.text |
|
|
|
|
|
) |
|
|
|
|
|
) |
|
|
|
|
|
|
|
|
|
|
|
actions = ActionChains(driver) |
|
|
wait_xpath(driver, slice_expr) |
|
|
actions.move_to_element(timeline_element).click().perform() |
|
|
for ts in driver.find_elements_by_xpath(slice_expr): |
|
|
|
|
|
if any(w in months for w in ts.text.strip().split()): |
|
|
|
|
|
yield ts |
|
|
|
|
|
try: |
|
|
|
|
|
int(ts.text.strip()) |
|
|
|
|
|
if len(ts.text.strip()) == 4: # it's a year |
|
|
|
|
|
yield ts |
|
|
|
|
|
except ValueError: |
|
|
|
|
|
continue |
|
|
|
|
|
|
|
|
# Wait until the buttons show up |
|
|
def delete_activity(driver, |
|
|
wait_xpath(driver, "//*[contains(@data-sigil, 'removeStoryButton')]") |
|
|
year=None): |
|
|
|
|
|
""" |
|
|
|
|
|
Deletes or hides all activity related to posts. |
|
|
|
|
|
|
|
|
delete_button = None |
|
|
Args: |
|
|
|
|
|
driver: seleniumrequests.Chrome Driver instance |
|
|
|
|
|
year: optional int YYYY year |
|
|
|
|
|
""" |
|
|
|
|
|
|
|
|
# Search for visible buttons in priority order |
|
|
driver.get("https://m.facebook.com/allactivity/?category_key=statuscluster") |
|
|
# Delete -> Untag -> Hide |
|
|
|
|
|
for button_type in button_types: |
|
|
|
|
|
try: |
|
|
|
|
|
delete_button = driver.find_element_by_xpath("//*[text()='{0}']".format(button_type)) |
|
|
|
|
|
if not delete_button.is_displayed(): |
|
|
|
|
|
continue |
|
|
|
|
|
break |
|
|
|
|
|
except SELENIUM_EXCEPTIONS as e: |
|
|
|
|
|
print(e) |
|
|
|
|
|
continue |
|
|
|
|
|
|
|
|
|
|
|
if not delete_button: |
|
|
#print(get_load_more(driver)) |
|
|
print("Could not find anything to delete") |
|
|
|
|
|
break |
|
|
|
|
|
|
|
|
|
|
|
click_button(driver, delete_button) |
|
|
actions = ActionChains(driver) |
|
|
wait_xpath(driver, confirmation_button_exp) |
|
|
|
|
|
confirmation_button = driver.find_element_by_xpath(confirmation_button_exp) |
|
|
|
|
|
|
|
|
|
|
|
print(confirmation_button) |
|
|
for ts in get_timeslices(driver): |
|
|
click_button(driver, confirmation_button) |
|
|
# Need to figure out how to ignore the ones with nothing in them |
|
|
|
|
|
print(ts.text) |
|
|
|
|
|
actions.move_to_element(ts) |
|
|
|
|
|
get_load_more(driver) |
|
|
|
|
|
|
|
|
except SELENIUM_EXCEPTIONS as e: |
|
|
time.sleep(1000) |
|
|
print(e) |
|
|
|
|
|
continue |
|
|
|
|
|
else: |
|
|
|
|
|
break |
|
|
|
|
|
|
|
|
|
|
|
# Required to sleep the thread for a bit after using JS to click this button |
|
|
#with archiver("activity") as archive_wall_post: |
|
|
time.sleep(5) |
|
|
|
|
|
driver.refresh() |
|
|
|
|
|