src/nominatim_db/indexer/progress.py

   1 # SPDX-License-Identifier: GPL-3.0-or-later
   2 #
   3 # This file is part of Nominatim. (https://nominatim.org)
   4 #
   5 # Copyright (C) 2024 by the Nominatim developer community.
   6 # For a full list of authors see the git log.
   7 """
   8 Helpers for progress logging.
   9 """
  10 import logging
  11 from datetime import datetime
  12
  13 LOG = logging.getLogger()
  14
  15 INITIAL_PROGRESS = 10
  16
  17 class ProgressLogger:
  18     """ Tracks and prints progress for the indexing process.
  19         `name` is the name of the indexing step being tracked.
  20         `total` sets up the total number of items that need processing.
  21         `log_interval` denotes the interval in seconds at which progress
  22         should be reported.
  23     """
  24
  25     def __init__(self, name: str, total: int, log_interval: int = 1) -> None:
  26         self.name = name
  27         self.total_places = total
  28         self.done_places = 0
  29         self.rank_start_time = datetime.now()
  30         self.log_interval = log_interval
  31         self.next_info = INITIAL_PROGRESS if LOG.isEnabledFor(logging.WARNING) else total + 1
  32
  33     def add(self, num: int = 1) -> None:
  34         """ Mark `num` places as processed. Print a log message if the
  35             logging is at least info and the log interval has passed.
  36         """
  37         self.done_places += num
  38
  39         if self.done_places < self.next_info:
  40             return
  41
  42         now = datetime.now()
  43         done_time = (now - self.rank_start_time).total_seconds()
  44
  45         if done_time < 2:
  46             self.next_info = self.done_places + INITIAL_PROGRESS
  47             return
  48
  49         places_per_sec = self.done_places / done_time
  50         eta = (self.total_places - self.done_places) / places_per_sec
  51
  52         LOG.warning("Done %d in %d @ %.3f per second - %s ETA (seconds): %.2f",
  53                     self.done_places, int(done_time),
  54                     places_per_sec, self.name, eta)
  55
  56         self.next_info += int(places_per_sec) * self.log_interval
  57
  58     def done(self) -> int:
  59         """ Print final statistics about the progress.
  60         """
  61         rank_end_time = datetime.now()
  62
  63         if rank_end_time == self.rank_start_time:
  64             diff_seconds = 0.0
  65             places_per_sec = float(self.done_places)
  66         else:
  67             diff_seconds = (rank_end_time - self.rank_start_time).total_seconds()
  68             places_per_sec = self.done_places / diff_seconds
  69
  70         LOG.warning("Done %d/%d in %d @ %.3f per second - FINISHED %s\n",
  71                     self.done_places, self.total_places, int(diff_seconds),
  72                     places_per_sec, self.name)
  73
  74         return self.done_places