105 lines
3.4 KiB
Python
105 lines
3.4 KiB
Python
|
|
import json
|
|
import os
|
|
import sys
|
|
from typing import Any, Dict
|
|
|
|
import pytumblr
|
|
|
|
|
|
def init_client() -> pytumblr.TumblrRestClient:
|
|
consumer_key = os.getenv('TUMBLR_CONSUMER_KEY')
|
|
consumer_secret = os.getenv('TUMBLR_CONSUMER_SECRET')
|
|
oauth_token = os.getenv('TUMBLR_OAUTH_TOKEN')
|
|
oauth_secret = os.getenv('TUMBLR_OAUTH_SECRET')
|
|
|
|
missing_vars = [name for name,
|
|
val in [('$TUMBLR_CONSUMER_KEY', consumer_key),
|
|
('$TUMBLR_CONSUMER_SECRET', consumer_secret),
|
|
('$TUMBLR_OAUTH_TOKEN', oauth_token),
|
|
('$TUMBLR_OAUTH_SECRET', oauth_secret)] if val is None]
|
|
|
|
if missing_vars:
|
|
print("Missing important environment variables:", missing_vars)
|
|
sys.exit(1)
|
|
|
|
return pytumblr.TumblrRestClient(
|
|
consumer_key=consumer_key, # type: ignore
|
|
consumer_secret=consumer_secret, # type: ignore
|
|
oauth_token=oauth_token, # type: ignore
|
|
oauth_secret=oauth_secret, # type: ignore
|
|
)
|
|
|
|
|
|
def main() -> None:
|
|
client = init_client()
|
|
post_map: Dict[str, Any] = {}
|
|
dumpster: Dict[str, Any] = {}
|
|
total = 0
|
|
offset = 0
|
|
limit = 20
|
|
while offset <= total:
|
|
# Begin LOOP
|
|
# Get me some posts! 😈🍪🍪🍪
|
|
data = client.posts('panda-pal.tumblr.com', **{'tag': 'inuyasha'}, \
|
|
offset = offset, \
|
|
limit = limit)
|
|
|
|
# Total check for the first iteration, but always checked for sanity.
|
|
if total == 0:
|
|
# Let's see what's in there,
|
|
total_posts = data['total_posts']
|
|
|
|
# Was nothing there?
|
|
if not total_posts:
|
|
print("Couldn't get total posts. We're outta here!")
|
|
sys.exit(1)
|
|
|
|
# Something was there, so we're good.
|
|
print(f"I'm working with {total_posts} total posts...")
|
|
total = total_posts
|
|
|
|
curr_posts = data['posts']
|
|
local_post_map: Dict[str, Any] = {}
|
|
for curr_post in curr_posts:
|
|
curr_key = curr_post['id_string']
|
|
if curr_key not in local_post_map:
|
|
local_post_map[curr_key] = curr_post
|
|
|
|
filtered_local_post_map = {}
|
|
for local_key in local_post_map:
|
|
local_post = local_post_map[local_key]
|
|
if 'parent_post_url' not in local_post:
|
|
filtered_local_post_map[local_key] = local_post
|
|
else:
|
|
dumpster[local_key] = local_post
|
|
|
|
# The sacred should we add, and if we should, DO ADD, if statement.
|
|
if any(post not in post_map for post in filtered_local_post_map):
|
|
post_map.update(filtered_local_post_map)
|
|
|
|
# The increment. Should always end the loop!
|
|
offset += limit
|
|
if offset <= total:
|
|
print(f"Processed batch {(offset // limit) - 1} of {total // 20}...")
|
|
else:
|
|
print(f"Processed final batch of {total} posts")
|
|
# End LOOP
|
|
|
|
if not post_map and not dumpster:
|
|
print('We found nothing, so we end with nothing. 🤷')
|
|
sys.exit(1)
|
|
else:
|
|
with open("./tumblr_data.json", "w") as f:
|
|
json.dump(post_map, f, indent=2, sort_keys=True)
|
|
with open("./tumblr_dumpster.json", "w") as f:
|
|
json.dump(dumpster, f, indent=2, sort_keys=True)
|
|
return
|
|
|
|
|
|
def retrieve_all_posts(tag: str):
|
|
return
|
|
|
|
if __name__ == '__main__':
|
|
main()
|
|
sys.exit(0) |