I added a batch_size to the extract file, allowing for batch processi…

…ng of projects instead of loading them all at once. During batch processing, I also incorporated timeout handling.

JavaExtractor/extract.py

@@ -3,30 +3,22 @@
 import itertools
 import multiprocessing
 import os
-import sys
 import shutil
 import subprocess
 from threading import Timer
-import sys
 from argparse import ArgumentParser
-from subprocess import Popen, PIPE, STDOUT, call
-
 
 def get_immediate_subdirectories(a_dir):
  return [(os.path.join(a_dir, name)) for name in os.listdir(a_dir)
  if os.path.isdir(os.path.join(a_dir, name))]
 
-
 def ParallelExtractDir(args, tmpdir, dir_):
- ExtractFeaturesForDir(args,tmpdir, dir_, "")
-
+ ExtractFeaturesForDir(args, tmpdir, dir_, "")
 
 def ExtractFeaturesForDir(args, tmpdir, dir_, prefix):
  command = ['java', '-cp', args.jar, 'JavaExtractor.App',
  '--max_path_length', str(args.max_path_length), '--max_path_width', str(args.max_path_width),
  '--dir', dir_, '--num_threads', str(args.num_threads)]
- # print command
- # os.system(command)
  kill = lambda process: process.kill()
  outputFileName = tmpdir + prefix + dir_.split('/')[-1]
  failed = False
@@ -39,36 +31,35 @@ def ExtractFeaturesForDir(args, tmpdir, dir_, prefix):
  finally:
  timer.cancel()
 
- if sleeper.poll() == 0:
- if len(stderr) > 0:
- print(sys.stderr, stderr, file=sys.stdout)
- else:
- print(sys.stderr, 'dir: ' + str(dir_) + ' was not completed in time', file=sys.stdout, flush=True)
+ if sleeper.poll() != 0:
  failed = True
  subdirs = get_immediate_subdirectories(dir_)
  for subdir in subdirs:
  ExtractFeaturesForDir(args, subdir, prefix + dir_.split('/')[-1] + '_')
- if failed:
- if os.path.exists(outputFileName):
- os.remove(outputFileName)
-
+
+ if failed and os.path.exists(outputFileName):
+ os.remove(outputFileName)
 
 def ExtractFeaturesForDirsList(args, dirs):
  tmp_dir = f"./tmp/feature_extractor{os.getpid()}/"
  if os.path.exists(tmp_dir):
  shutil.rmtree(tmp_dir, ignore_errors=True)
  os.makedirs(tmp_dir)
- try:
- p = multiprocessing.Pool(4)
- p.starmap(ParallelExtractDir, zip(itertools.repeat(args),itertools.repeat(tmp_dir), dirs))
- #for dir in dirs:
- # ExtractFeaturesForDir(args, dir, '')
+
+ for i in range(0, len(dirs), args.batch_size):
+ batch_dirs = dirs[i:i + args.batch_size]
+ timeout_seconds = 60 # timeout setting
+ try:
+ with multiprocessing.Pool(4) as p:
+ result = p.starmap_async(ParallelExtractDir, zip(itertools.repeat(args), itertools.repeat(tmp_dir), batch_dirs))
+ result.get(timeout=timeout_seconds)
+ except multiprocessing.TimeoutError:
+ continue
+
  output_files = os.listdir(tmp_dir)
  for f in output_files:
  os.system("cat %s/%s" % (tmp_dir, f))
- finally:
- shutil.rmtree(tmp_dir, ignore_errors=True)
-
+ os.remove(os.path.join(tmp_dir, f))
 
 if __name__ == '__main__':
  parser = ArgumentParser()
@@ -78,6 +69,9 @@ def ExtractFeaturesForDirsList(args, dirs):
  parser.add_argument("-j", "--jar", dest="jar", required=True)
  parser.add_argument("-dir", "--dir", dest="dir", required=False)
  parser.add_argument("-file", "--file", dest="file", required=False)
+ # add a new parameter batch_size
+ parser.add_argument("-batch_size", "--batch_size", dest="batch_size", required=False, default=3, type=int)
+
  args = parser.parse_args()
 
  if args.file is not None:
@@ -86,9 +80,5 @@ def ExtractFeaturesForDirsList(args, dirs):
  os.system(command)
  elif args.dir is not None:
  subdirs = get_immediate_subdirectories(args.dir)
- to_extract = subdirs
- if len(subdirs) == 0:
- to_extract = [args.dir.rstrip('/')]
+ to_extract = subdirs if subdirs else [args.dir.rstrip('/')]
  ExtractFeaturesForDirsList(args, to_extract)
-
-