From 829fc6672ba1d66b32de492c764d71e7aba0cfb3 Mon Sep 17 00:00:00 2001
From: Luis Salamanca <luis.salamanca@sdsc.ethz.ch>
Date: Fri, 30 Nov 2018 13:12:33 +0000
Subject: [PATCH] Corrections to scripts
---
src/python/run_correctxml.py | 6 +++++-
src/python/run_extract_origxml.py | 6 +++++-
src/python/utils_proc.py | 15 +++++++++++++++
3 files changed, 25 insertions(+), 2 deletions(-)
diff --git a/src/python/run_correctxml.py b/src/python/run_correctxml.py
index 87e74b7b..45543847 100644
--- a/src/python/run_correctxml.py
+++ b/src/python/run_correctxml.py
@@ -50,5 +50,9 @@ for infile in files_proc:
#print('Corrected %s' % infile)
except:
print("File to correct %s prompted an error" % infile)
-
+
+# Commands to get the compressed version of the file
+#data/AB/${year}/02_extractedxml.tar.gz
+utils_proc.compress_tar(output_file)
+
print('Total time for correction of year %d: %f' % (int(year_tocomp) ,(time.time() - t1)))
\ No newline at end of file
diff --git a/src/python/run_extract_origxml.py b/src/python/run_extract_origxml.py
index 8af12f7e..e283788c 100644
--- a/src/python/run_extract_origxml.py
+++ b/src/python/run_extract_origxml.py
@@ -49,5 +49,9 @@ for infile in files_proc:
print('Extracted %s' % infile)
except:
print("File %s prompted an error" % infile)
-
+
+# Commands to get the compressed version of the file
+#data/AB/${year}/02_extractedxml.tar.gz
+utils_proc.compress_tar(output_file)
+
print('Total time for year %d: %f' % (int(year_tocomp) ,(time.time() - t1)))
\ No newline at end of file
diff --git a/src/python/utils_proc.py b/src/python/utils_proc.py
index 5cf163fa..7348be81 100644
--- a/src/python/utils_proc.py
+++ b/src/python/utils_proc.py
@@ -144,6 +144,21 @@ def addto_tar(input_file, folder_database, name_file):
tf.close()
return name_tar
+def compress_tar(infile, outname = ''):
+ if len(outname) == 0:
+ outname = infile
+
+ # tar -xf data/AB/${year}/02_extractedxml.tar.gz
+ # tar -czvf data/AB/${year}/02_extractedxml.tar.gz ./${year}/
+ # rm -rf ${year}
+ year = infile.split('/')[-2]
+ c1 = 'tar -xf ' + infile
+ call_with_out(c1)
+ c2 = 'tar -czvf ' + outname + ' ./' + str(year) + '/'
+ call_with_out(c2)
+ c3 = 'rm -rf ' + str(year)
+ call_with_out(c3)
+
def correct_metadata(year, id_doc, flag_end):
--
GitLab