Reduce API calls by downloading avatars directly

IvanGlinkin / CCTV

Close-Circuit Telegram Vision revolutionizes location tracking with its open-source design and Telegram API integration. Offering precise tracking within 50-100 meters, users can monitor others in real-time for logistics or safety, redefining how we navigate our surroundings

GNU General Public License v3.0

2.31k stars 319 forks source link

Reduce API calls by downloading avatars directly #1

Closed ask0n closed 2 months ago

ask0n commented 2 months ago

did some cleanup:

[x] _no_avatar.jpg was renamed into .no_avatar.jpg to survive rm commands
[x] download avatars action is triggered against generated json, to save time on actual data collection and reduce API calls. Avatars won't change quickly but users can leave location while script is running
[x] image name stores user id, too keep fancy photos for users who changes nicknames and avatars often ;)
[x] small restructure to not repeat Start harvesting data: in a loop and countdown timer to indicate actual pause
[x] added .gitignore
[x] linter
[x] introduced speed_kmh parameter to calculate dynamic timeout between GetLocatedRequest operations. TG reduce to return modified data if distance between points is too high and time between requests is too low. It looks like save "travel speed" should be less than 100km/h
[x] added exception for FloodWaitError

IvanGlinkin commented 2 months ago

hey man. let me check it tomorrow and if it's working well will pull you request

IvanGlinkin commented 2 months ago

Man, something wrong with data collection - some areas of collecting data are missing hence no actual coordinates

ask0n commented 2 months ago

Sure will take a look. Which data was missed?

IvanGlinkin commented 2 months ago

just launch the script and see results on the map and compare the same with my out original output. Seems like the combine of coordinates isn't working or the points of collection don't collect it all. if you see the original output - you can see the proper square but right now it like chaotic

ask0n commented 2 months ago

oh, I see you mean this polygon figures?

it seems I was confused by this logic https://github.com/IvanGlinkin/CCTV/blob/main/start.py#L178-L187 my plan was to move html generation outside of the loop and use combined json for it. Lemme check.

IvanGlinkin commented 2 months ago

yeah, the second one screenshot is what i meant - it's showing more accurate locations. I liked what you did regarding frontend - great job) thanks for allocating your time and efforts. Really appreciate it But due to the application is not working - there is no point of frontend. let's find out together what has been broken in logic, fix it and put all of your addons to the master branch :)

IvanGlinkin commented 2 months ago

any luck so far?

ask0n commented 2 months ago

any luck so far?

hey, yes, lemme prepare a new PR. Collected data looks identical now but I discovered that https://t.me url also throttles requests and returns a dummy page. Need to update logic to not fetch default logo instead of avatar.