diff options
| -rwxr-xr-x | please | 24 |
1 files changed, 24 insertions, 0 deletions
@@ -257,6 +257,27 @@ def run_keysmissingcol(args): env=args.env) subprocess.call(cmd, shell=True) +def run_dumpungrobided(args): + if args.rebuild: + rebuild_scalding() + print("Starting dumpungrobided job...") + output = "{}/output-{}/{}-dumpungrobided".format( + HDFS_DIR, + args.env, + datetime.strftime(datetime.now(), "%Y-%m-%d-%H%M.%S")) + cmd = """hadoop jar \ + scalding/target/scala-2.11/sandcrawler-assembly-0.2.0-SNAPSHOT.jar \ + com.twitter.scalding.Tool sandcrawler.DumpUnGrobidedJob \ + --hdfs \ + --app.conf.path scalding/ia_cluster.conf \ + --hbase-table wbgrp-journal-extract-0-{env} \ + --zookeeper-hosts {zookeeper_hosts} \ + --output {output}""".format( + output=output, + zookeeper_hosts=ZOOKEEPER_HOSTS, + env=args.env) + subprocess.call(cmd, shell=True) + def main(): parser = argparse.ArgumentParser() @@ -320,6 +341,9 @@ def main(): sub_keysmissingcol.add_argument('column', help="column to SCAN for missing keys") + sub_dumpungrobided = subparsers.add_parser('dump-ungrobided') + sub_dumpungrobided.set_defaults(func=run_dumpungrobided) + args = parser.parse_args() if not args.__dict__.get("func"): print("tell me what to do! (try --help)") |