Merge branch '3825-crunch-pipe-to-arv-put-final'

[arvados.git] / crunch_scripts / run-command

diff --git a/crunch_scripts/run-command b/crunch_scripts/run-command
index c85e74a6e48b4bd23f85c90604c8a42df36a495d..c5fbcdfc024f5fd3c6dac20e2a8dc89ec5168a66 100755 (executable)
--- a/crunch_scripts/run-command
+++ b/crunch_scripts/run-command
@@ -1,12 +1,20 @@
 #!/usr/bin/env python
 
+import logging
+
+logger = logging.getLogger('run-command')
+log_handler = logging.StreamHandler()
+log_handler.setFormatter(logging.Formatter("run-command: %(message)s"))
+logger.addHandler(log_handler)
+logger.setLevel(logging.INFO)
+
 import arvados
 import re
 import os
 import subprocess
 import sys
 import shutil
-import subst
+import crunchutil.subst as subst
 import time
 import arvados.commands.put as put
 import signal
@@ -15,12 +23,10 @@ import copy
 import traceback
 import pprint
 import multiprocessing
-import logging
-import robust_put
-import vwd
+import crunchutil.robust_put as robust_put
+import crunchutil.vwd as vwd
 
 os.umask(0077)
-logging.basicConfig(format="run-command: %(message)s")
 
 t = arvados.current_task().tmpdir
 
@@ -102,7 +108,7 @@ def get_items(p, value):
     fn = subst.do_substitution(p, value)
     mode = os.stat(fn).st_mode
     prefix = fn[len(os.environ['TASK_KEEPMOUNT'])+1:]
-    if mode != None:
+    if mode is not None:
         if stat.S_ISDIR(mode):
             items = ["$(dir %s/%s/)" % (prefix, l) for l in os.listdir(fn)]
         elif stat.S_ISREG(mode):
@@ -114,6 +120,8 @@ def get_items(p, value):
 
 stdoutname = None
 stdoutfile = None
+stdinname = None
+stdinfile = None
 rcode = 1
 
 try:
@@ -121,8 +129,8 @@ try:
         if arvados.current_task()['sequence'] == 0:
             var = jobp["task.foreach"]
             items = get_items(jobp, jobp[var])
-            logging.info("parallelizing on %s with items %s" % (var, items))
-            if items != None:
+            logger.info("parallelizing on %s with items %s" % (var, items))
+            if items is not None:
                 for i in items:
                     params = copy.copy(jobp)
                     params[var] = i
@@ -154,20 +162,28 @@ try:
 
     cmd = expand_list(taskp, taskp["command"])
 
-    if "save.stdout" in taskp:
-        stdoutname = subst.do_substitution(taskp, taskp["save.stdout"])
-        stdoutfile = open(stdoutname, "wb")
+    if "task.stdin" in taskp:
+        stdinname = subst.do_substitution(taskp, taskp["task.stdin"])
+        stdinfile = open(stdinname, "rb")
 
-    logging.info("{}{}".format(' '.join(cmd), (" > " + stdoutname) if stdoutname != None else ""))
+    if "task.stdout" in taskp:
+        stdoutname = subst.do_substitution(taskp, taskp["task.stdout"])
+        stdoutfile = open(stdoutname, "wb")
 
+    logger.info("{}{}{}".format(' '.join(cmd), (" < " + stdinname) if stdinname is not None else "", (" > " + stdoutname) if stdoutname is not None else ""))
+except subst.SubstitutionError as e:
+    logger.error(str(e))
+    logger.error("task parameters were:")
+    logger.error(pprint.pformat(taskp))
+    sys.exit(1)
 except Exception as e:
-    logging.exception("caught exception")
-    logging.error("task parameters was:")
-    logging.error(pprint.pformat(taskp))
+    logger.exception("caught exception")
+    logger.error("task parameters were:")
+    logger.error(pprint.pformat(taskp))
     sys.exit(1)
 
 try:
-    sp = subprocess.Popen(cmd, shell=False, stdout=stdoutfile)
+    sp = subprocess.Popen(cmd, shell=False, stdin=stdinfile, stdout=stdoutfile)
     sig = SigHandler()
 
     # forward signals to the process.
@@ -178,14 +194,14 @@ try:
     # wait for process to complete.
     rcode = sp.wait()
 
-    if sig.sig != None:
-        logging.critical("terminating on signal %s" % sig.sig)
+    if sig.sig is not None:
+        logger.critical("terminating on signal %s" % sig.sig)
         sys.exit(2)
     else:
-        logging.info("completed with exit code %i (%s)" % (rcode, "success" if rcode == 0 else "failed"))
+        logger.info("completed with exit code %i (%s)" % (rcode, "success" if rcode == 0 else "failed"))
 
 except Exception as e:
-    logging.exception("caught exception")
+    logger.exception("caught exception")
 
 # restore default signal handlers.
 signal.signal(signal.SIGINT, signal.SIG_DFL)
@@ -195,11 +211,11 @@ signal.signal(signal.SIGQUIT, signal.SIG_DFL)
 for l in links:
     os.unlink(l)
 
-logging.info("the following output files will be saved to keep:")
+logger.info("the following output files will be saved to keep:")
 
 subprocess.call(["find", ".", "-type", "f", "-printf", "run-command: %12.12s %h/%f\\n"], stdout=sys.stderr)
 
-logging.info("start writing output to keep")
+logger.info("start writing output to keep")
 
 if "task.vwd" in taskp:
     if "task.foreach" in jobp:
@@ -209,7 +225,7 @@ if "task.vwd" in taskp:
         # Just a single task, so do merge with the original collection
         outcollection = vwd.checkin(subst.do_substitution(taskp, taskp["task.vwd"]), outdir, merge=True).manifest_text()
 else:
-    outcollection = robust_put.upload(outdir)
+    outcollection = robust_put.upload(outdir, logger)
 
 api.job_tasks().update(uuid=arvados.current_task()['uuid'],
                                      body={