Update cloudiness with batching and multi-threading

[ifenni]

The goal of these updates is to significantly reduce the current computational cost of cloudiness calculations for both forecasted (future) and archived (historical) data.
The code updates mainly impacted utils/cloudiness.py, and included:

• batching the set of points into blocks to reduce the number of session calls and leverage the Open-Meteo API’s multi-point input capability.
• adding multithreading to the loop over these blocks inside the function "get_overpass_cloudines"
• adding a rate_limiter class to avoid API rate limit issues

We also updated the main py file next_pass.py to include a 1 minute waiting time between Sentinel-1 and Sentinel-2 when cloudiness is requested, to avoid hitting the weather API minute usage limits (600 calls / min).

Without cloudiness, the next-pass computation completes in 3 minutes and 55 seconds. Enabling the cloudiness option (-c) with the previous implementation increased the runtime to 15 minutes and 47 seconds. With the present updates, the runtime is reduced to 4 minutes and 52 seconds (including the mandatory 1-minute wait to avoid rate limiting), corresponding to a gain of approximately 11 minutes, or a ~69% reduction in runtime compared to the previous cloudiness-enabled version.