Fix for non-ascii chars in header

cni · Dec 19, 2018 · d712a4c · d712a4c
1 parent 6ecf80f
commit d712a4c
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 2 deletions.
diff --git a/manifest.json b/manifest.json
@@ -11,9 +11,9 @@
   "flywheel": {
     "suite": "Stanford CNI"
   },
-  "version": "1.6.0",
+  "version": "1.6.1",
   "custom": {
-    "docker-image": "stanfordcni/pfile-mr-classifier:1.6.0"
+    "docker-image": "stanfordcni/pfile-mr-classifier:1.6.1"
   },
   "config": {
     "timezone": {

diff --git a/pfile-mr-classifier.py b/pfile-mr-classifier.py
@@ -15,6 +15,13 @@
 logging.basicConfig()
 log = logging.getLogger('pfile-mr-classifier')
 
+def format_string(in_string):
+    formatted = re.sub(r'[^\x00-\x7f]',r'', str(in_string)) # Remove non-ascii characters
+    formatted = filter(lambda x: x in string.printable, formatted)
+    if len(formatted) == 1 and formatted == '?':
+        formatted = None
+    return formatted
+
 def get_pfile_classification(pfile):
     """
     Determine pfile classification from series description, etc.
@@ -185,6 +192,7 @@ def extract_pfile_header(pfile_header_csv):
         next(csvreader)
         for row in csvreader:
             if row[1]:
+                row[1] = format_string(row[1])
                 try:
                     row[1] = int(row[1])
                 except: