corssposter

SETUP "GUIDE"

• get signal cli (portable is good) and write path (STARTING FROM YOUR HOME DIR) to config.py
  • use config_template.py for crearing config.py
  • for example "signalcli/signal-cli-0.14.8/bin/signal-cli",
• get ublock-lite and unpack it to this_folder/extension/uBoLite.......
• create a venv here or somewhere & activate it and install atproto, seleniumbase :

python -m venv somename
source somename/bin/active
pip install atproto seleniumbase 

• use this venv to run both skripts (ig you could use 2 different ones too)
• setup signalcli
• get a group id from signal-cli , write it to config.py (mby you can also do it by name by now not sure)
• get bsky acc access code and write with bsky-handle to config.py
• select an instagram acc by writing it in config.py
• should work now (atl the signal part)
• if signalcli hasnt been initialized, it still works
• FOR ASPECT RATIO DETECTION: install identify and ffmpeg/ffprobe (pre installed with some distros)

structure:

seltest.py

• contains the 'main' program: gets recent posts, checks if theyre new, then downloads them, posts to signal and STAGES bsky posting via a .json file

postd.py

contains postd class used by seltest.py (in the Post class) and directly by nksyte2.py

config.py {has to be created manually}

contains logins, insta-target, signalcli path & setup. use config_template.py as reference

nskyte2.py

• posts the staged posts to bsky (not implemented yet)
  • rn it checks if its connected through tor. so call it with torify
    • not sure if some stuff still leaks. hostname-lookups could go past torify. easiest way to be sure is some wireguard analysis ig
• basic process:
  • read bsky_Staged_Posts_TMP.json and turn back into post objects
  • post them & add to hashset
  • delete bsky_Staged_Posts_TMP.json

burn_sky.py

• deletes all post from bsky. useful for testing

TODOS:

(seltest.py) better Server Error handling:

Sometimes imginn returns a server error. seems to be rare up to now and resolved within 24h.

current state:

• detection is implemented. see Post.open_with_check() method.

• Post.download() method returns false if it happened with multi-img/vid post:

  • with single vid postsites it doesnt matter, the overview cdnlink is already all the quality we can get
  • with single img post, the overview cdnlink is also used (lower quality but acceptable)
  • with multiimg it prints the error and skips { mby stage for retrying}

at overview the error is currently unhandled. ig we just rescedule then

(nskyte2.py) compress/resize image to get under 10^6 BYTE bsky server limit (for imgs)

• rn it skips those

(nskyte2.py) pass image/video aspect ratio to server (IMPLEMENTED,LIMITED TESTING DONE, PLS DO MORE)

• otherwise its shown as 1:1 with padding (might look better tho)
• currently this uses 'identify' bash calls. might be easier/more reliable with a python image lib that could need for further features anw

(seltest.py) implement stories

• not sure how to post them, but might still be nice as a backup

nskyte2.py