Open maaikelimper opened 12 months ago
Hey Maïke, I just integrated the proposed modifications. Now, the script runs faster. While dispatching to subprocesses, there was a blockage specifically at the database insertion. Indeed, there was a conflict between processes, particularly during the opening and closing of the connection (it's resolved now). I'm sharing with you the changes made on GitHub.
Thanks, I could not test the changes I proposed, good to hear that you managed to resolve the conflict between the processes relating to opening and closing of database connections
thank you Maaïke
Chems mentioned that processing was slow. This PR introduces batch-processing on the queue, with each batch being processed in a new multiprocessing.Process.
(no bulk-insert in the DB introduced yet, as I suspect the download+processing is the main bottle neck atm)