From 2492473d48b57348d4b9d86be0d50467336a6efa Mon Sep 17 00:00:00 2001
From: Russell Jarvis <rjjarvis@asu.edu>
Date: Fri, 26 Jun 2020 15:44:33 +1000
Subject: [PATCH] for heroku
---
science_access/get_bmark_corpus.py | 5 +++++
science_access/scrape.py | 4 +---
2 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/science_access/get_bmark_corpus.py b/science_access/get_bmark_corpus.py
index 6fff696..f8e6933 100644
--- a/science_access/get_bmark_corpus.py
+++ b/science_access/get_bmark_corpus.py
@@ -37,6 +37,11 @@ def process(link):
pdf_file = requests.get(link, stream=True)
buffered = convert_pdf_to_txt(pdf_file)
urlDat = text_proc(buffered,urlDat)
+ driver.close()
+ driver.quit()
+ driver = None
+ del driver
+
return urlDat
#try:
diff --git a/science_access/scrape.py b/science_access/scrape.py
index 2eb0894..57f71ae 100644
--- a/science_access/scrape.py
+++ b/science_access/scrape.py
@@ -55,7 +55,6 @@ if 'DYNO' in os.environ:
heroku = False
else:
heroku = True
-'''
def get_driver():
if 'DYNO' in os.environ:
heroku = True
@@ -98,8 +97,7 @@ def get_driver():
return driver
-driver = get_driver()
-'''
+#driver = get_driver()
rsrcmgr = PDFResourceManager()
retstr = StringIO()
--
GitLab