From 430e209d0a80186deaf0b5cc5e723f7b4bb26d1a Mon Sep 17 00:00:00 2001
From: jgw4sq <~jgw4sq@virginia.edu>
Date: Fri, 11 Mar 2016 13:20:41 -0500
Subject: [PATCH] New commit for fixing formating of nist harvester tags

---
 scrapi/harvesters/nist.py | 28 +++++++++++++++++++++++++++-
 1 file changed, 27 insertions(+), 1 deletion(-)

diff --git a/scrapi/harvesters/nist.py b/scrapi/harvesters/nist.py
index 3192d2ec..defd5676 100644
--- a/scrapi/harvesters/nist.py
+++ b/scrapi/harvesters/nist.py
@@ -6,13 +6,39 @@
 from __future__ import unicode_literals
 
 from scrapi.base import OAIHarvester
+from scrapi.base.helpers import updated_schema
 
 
 class NistHarvester(OAIHarvester):
     short_name = 'nist'
     long_name = 'NIST MaterialsData'
     url = 'https://materialsdata.nist.gov'
-
     base_url = 'https://materialsdata.nist.gov/dspace/oai/request'
     property_list = ['relation', 'rights', 'identifier', 'type', 'date', 'setSpec']
     timezone_granularity = True
+
+    @property
+    def schema(self):
+        return updated_schema(self._schema, {'subjects': ('//dc:subject/node()', format_tags)})
+
+
+def format_tags(all_tags):
+    tags = []
+    for tag in all_tags:
+        tag = tag.replace("Computational File Repository Categories", '')
+        tag = tag.replace("Computational File Repository", '')
+        tag = tag.replace("File Repository Categories", '')
+
+        if "::" in tag:
+            tags.extend(tag.split("::"))
+        if "," in tag:
+            tags.extend(tag.split(","))
+        elif "::" not in tag:
+            tags.append(tag)
+
+    for tag in tags:
+        if "::" in tag:
+            tags.remove(tag)
+        if tag == "":
+            tags.remove(tag)
+    return tags