diff --git a/dataset-construction/requirements.txt b/dataset-construction/requirements.txt
index f0d23e4..cd44e01 100644
--- a/dataset-construction/requirements.txt
+++ b/dataset-construction/requirements.txt
@@ -1,3 +1,5 @@
 tqdm
 pymongo
-numpy
\ No newline at end of file
+numpy
+nltk==3.7
+strsim==0.0.3
diff --git a/dataset-construction/setup.sh b/dataset-construction/setup.sh
new file mode 100644
index 0000000..0738087
--- /dev/null
+++ b/dataset-construction/setup.sh
@@ -0,0 +1,4 @@
+#!/usr/bin/env bash
+
+pip3 install -r requirements.txt
+python3 -m nltk.downloader punkt
diff --git a/dataset-construction/src/ndb_data/construction/make_database_initial.py b/dataset-construction/src/ndb_data/construction/make_database_initial.py
index d24385e..23c71f2 100644
--- a/dataset-construction/src/ndb_data/construction/make_database_initial.py
+++ b/dataset-construction/src/ndb_data/construction/make_database_initial.py
@@ -31,8 +31,6 @@
 from similarity.normalized_levenshtein import NormalizedLevenshtein
 from tqdm import tqdm
 
-from ndb_data.util.log_helper import setup_logging
-
 detok = TreebankWordDetokenizer()
 
 logger = logging.getLogger(__name__)
@@ -71,7 +69,6 @@ def normalize_subject(subject_name, fact):
 
 
 if __name__ == "__main__":
-    setup_logging()
     parser = ArgumentParser()
     parser.add_argument("cache_dir")
     parser.add_argument("out_file")