14715: keepproxy.service checks for cluster config

[arvados.git] / crunch_scripts / hash

diff --git a/crunch_scripts/hash b/crunch_scripts/hash
index 6d6c88f64d555aef6990340256bd4129b8d42c92..56eec7a5ffd985834009016138f4b5ac99f5cb87 100755 (executable)
--- a/crunch_scripts/hash
+++ b/crunch_scripts/hash
@@ -1,32 +1,37 @@
-#!/usr/bin/env python
+#!/usr/bin/env python                                                                                                                                                                            
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
 
 import arvados
 import hashlib
-import re
+import os
 
-arvados.job_setup.one_task_per_input_file(if_sequence=0, and_end_task=True)
+arvados.job_setup.one_task_per_input_file(if_sequence=0, and_end_task=True, input_as_path=True)
 
 this_job = arvados.current_job()
 this_task = arvados.current_task()
-this_task_input = this_task['parameters']['input']
 
 if 'algorithm' in this_job['script_parameters']:
-    alg = this_job['script_parameters']
+    alg = this_job['script_parameters']['algorithm']
 else:
     alg = 'md5'
 digestor = hashlib.new(alg)
 
-input_stream = arvados.DataReader(this_task_input)
-while True:
-    buf = input_stream.read(2**20)
-    if len(buf) == 0:
-        break
-    digestor.update(buf)
+input_file = arvados.get_task_param_mount('input')
 
-hexdigest = digestor.hexdigest()
-file_name = re.sub(r'^[^/]+/', '', this_task_input)
+with open(input_file) as f:
+    while True:
+        buf = f.read(2**20)
+        if len(buf) == 0:
+            break
+        digestor.update(buf)
 
-input_stream.close()
+hexdigest = digestor.hexdigest()
 
-this_task.set_output("%s %s\n" % (hexdigest, file_name))
+file_name = '/'.join(this_task['parameters']['input'].split('/')[1:])
 
+out = arvados.CollectionWriter()
+out.set_current_file_name("md5sum.txt")
+out.write("%s %s\n" % (hexdigest, file_name))
+this_task.set_output(out.finish())