aboutsummaryrefslogtreecommitdiffstats
path: root/covid19_tool.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2020-04-03 15:06:11 -0700
committerBryan Newbold <bnewbold@archive.org>2020-04-03 15:06:11 -0700
commit7bf8ae73b8b5dfca4d17f353cdbec669e69bbbec (patch)
tree8a0ef776a0d44339d31c099d0e85e7da7a66dd24 /covid19_tool.py
parent913a3d7b2fe585602124e2092dd44afa760086e8 (diff)
downloadfatcat-covid19-7bf8ae73b8b5dfca4d17f353cdbec669e69bbbec.tar.gz
fatcat-covid19-7bf8ae73b8b5dfca4d17f353cdbec669e69bbbec.zip
refactor elastic transform into CLI tool
Diffstat (limited to 'covid19_tool.py')
-rwxr-xr-xcovid19_tool.py14
1 files changed, 14 insertions, 0 deletions
diff --git a/covid19_tool.py b/covid19_tool.py
index 23a2c6c..1cf8dce 100755
--- a/covid19_tool.py
+++ b/covid19_tool.py
@@ -10,6 +10,8 @@ import sys
import argparse
from fatcat_covid19.webface import app
+from fatcat_covid19.derivatives import enrich_derivatives_file
+from fatcat_covid19.transform import transform_es_file
def main():
@@ -58,6 +60,16 @@ def main():
help="directory to look for files (in 'pdf' subdirectory)",
default="fulltext_web")
+ sub_transform_es = subparsers.add_parser('transform-es',
+ help="transform fulltext JSON to elasticsearch schema JSON")
+ sub_transform_es.add_argument('json_file',
+ help="input JSON rows file (fulltext)",
+ type=argparse.FileType('r'))
+ sub_transform_es.add_argument('--json-output',
+ help="file to write to",
+ type=argparse.FileType('r'),
+ default=sys.stdout)
+
args = parser.parse_args()
if args.action == 'webface':
@@ -65,6 +77,8 @@ def main():
if args.action == 'derivatives':
enrich_derivatives_file(args.json_file, args.json_output,
args.base_dir)
+ if args.action == 'transform-es':
+ transform_es_file(args.json_file, args.json_output)
else:
print("tell me what to do!")
sys.exit(-1)