aboutsummaryrefslogtreecommitdiffstats
path: root/python/pdfextract_tool.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2021-10-26 13:35:36 -0700
committerBryan Newbold <bnewbold@archive.org>2021-10-26 13:35:36 -0700
commit600ad67925a748200ddf21d5aeabd157d2bb3664 (patch)
tree89ae6bc24e6eb3821c03efd7d781430345c68aa0 /python/pdfextract_tool.py
parent05bd7cbcc62588e431c5efd533189e246b2a997e (diff)
downloadsandcrawler-600ad67925a748200ddf21d5aeabd157d2bb3664.tar.gz
sandcrawler-600ad67925a748200ddf21d5aeabd157d2bb3664.zip
start handling trivial lint cleanups: unused imports, 'is None', etc
Diffstat (limited to 'python/pdfextract_tool.py')
-rwxr-xr-xpython/pdfextract_tool.py3
1 files changed, 0 insertions, 3 deletions
diff --git a/python/pdfextract_tool.py b/python/pdfextract_tool.py
index 717b743..f3df6e3 100755
--- a/python/pdfextract_tool.py
+++ b/python/pdfextract_tool.py
@@ -4,11 +4,8 @@ KNOWN ISSUE: thumbnails are not published to kafka in multi-processing mode
"""
import argparse
-import datetime
-import json
import sys
-from grobid2json import teixml2json
from sandcrawler import *