diff austalk-select-hVd-words.py @ 14:a38315ecf593 draft

planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty

--- a/austalk-select-hVd-words.py	Wed Feb 01 22:34:24 2017 -0500
+++ b/austalk-select-hVd-words.py	Wed Nov 01 01:18:15 2017 -0400
@@ -2,10 +2,10 @@
 import argparse
 import pyalveo
 import sys
+from util import API_URL
 
-API_URL = 'https://app.alveo.edu.au/'
 PREFIXES = """
-PREFIX dc:<http://purl.org/dc/terms/>
+PREFIX dcterms:<http://purl.org/dc/terms/>
 PREFIX austalk:<http://ns.austalk.edu.au/>
 PREFIX olac:<http://www.language-archives.org/OLAC/1.1/>
 PREFIX ausnc:<http://ns.ausnc.org.au/schemas/ausnc_md_model/>
@@ -22,14 +22,17 @@
 PREFIX iso: <http://purl.org/iso25964/skos-thes#>
 PREFIX dada: <http://purl.org/dada/schema/0.2#>"""
 
+
 def parser():
     parser = argparse.ArgumentParser(description="Retrieves Alveo Item Lists")
     parser.add_argument('--api_key', required=True, action="store", type=str, help="Alveo API key")
     parser.add_argument('--speaker', required=True, action="store", type=str, help="Speaker identifier")
-    parser.add_argument('--words', required=False, default='all', action="store", type=str, help="Word group (all, monopthongs, dipthongs)")
+    parser.add_argument('--words', required=False, default='all', action="store", type=str,
+                        help="Word group (all, monopthongs, dipthongs)")
     parser.add_argument('--output', required=True, action="store", type=str, help="output file name")
     return parser.parse_args()
 
+
 def find_hVd_words(api_key, speakerid, output, words='all'):
     """Find words in the Austalk corpus
     """
@@ -42,14 +45,14 @@
   ?item a ausnc:AusNCObject .
   ?item olac:speaker ?speaker .
   ?speaker austalk:id "%s" .
-  ?item austalk:prompt ?prompt .
+  ?item austalk:prototype ?prot .
+  ?prot austalk:prompt ?prompt .
   ?item austalk:componentName ?compname .
  """ % speakerid
 
-    hVdWords = {
-        'monopthongs': ['head', 'had', 'hud', 'heed', 'hid', 'hood', 'hod', 'whod', 'herd', 'haired', 'hard', 'horde'],
-        'dipthongs': ['howd', 'hoyd', 'hide', 'hode', 'hade', 'heared'] 
-        }
+    hVdWords = dict(monopthongs=['head', 'had', 'hud', 'heed', 'hid', 'hood', 'hod',
+                                 'whod', 'herd', 'haired', 'hard', 'horde'],
+                    dipthongs=['howd', 'hoyd', 'hide', 'hode', 'hade', 'heared'])
 
     if words == 'all':
         words = hVdWords['monopthongs'] + hVdWords['dipthongs']
@@ -62,6 +65,7 @@
 
     query += filterclause + "}"
 
+    print(query)
     result = client.sparql_query('austalk', query)
 
     items = []
@@ -71,11 +75,7 @@
     with open(output, 'w') as out:
         out.write("Speaker\tPrompt\tItemURL\n")
         for item in items:
-            # TODO: fix this once the RDF data is fixed in alveo
-            # need to modify the item URL
-            itemurl = item[1].replace('http://id.austalk.edu.au/item/', 'https://app.alveo.edu.au/catalog/austalk/')
-
-            out.write(speakerid + "\t" + item[0] + "\t" + itemurl + "\n")
+            out.write(speakerid + "\t" + item[0] + "\t" + item[1] + "\n")
 
 
 def main():
@@ -87,5 +87,6 @@
         print("ERROR: " + str(e), file=sys.stderr)
         sys.exit(1)
 
+
 if __name__ == '__main__':
     main()