able to load pickled structures

src/syntactic_structure.py

@@ -1,5 +1,6 @@
 from xml.etree import ElementTree
 import logging
+import pickle
 
 from component import Component, ComponentType
 from lemma_features import get_lemma_features
@@ -92,7 +93,12 @@ class SyntacticStructure:
         return [] if matches is None else matches
 
 
-def build_structures(filename):
+def build_structures(args):
+    filename = args.structures
+    if args.pickled_structures:
+        with open(filename, 'rb') as fp:
+            return pickle.load(fp)
+
     max_num_components = -1
     with open(filename, 'r') as fp:
         et = ElementTree.XML(fp.read())

src/wani.py

@@ -107,8 +107,8 @@ def match_file(words, structures):
     return matches
 
 
-def main(structures_file, args):
-    structures, lemma_msds, max_num_components = build_structures(structures_file)
+def main(args):
+    structures, lemma_msds, max_num_components = build_structures(args)
 
     match_store = MatchStore(args)
     word_stats = WordStats(lemma_msds)
@@ -218,11 +218,13 @@ if __name__ == '__main__':
                         help='Tag for separators, usually pc or c', default="pc")
     parser.add_argument('--parallel',
                         help='Run in multiple processes, should speed things up')
+
     parser.add_argument('--match-to-file', help='Do not use!')
+    parser.add_argument('--pickled-structures', help='Do not use!', action='store_true')
 
     args = parser.parse_args()
     logging.basicConfig(stream=sys.stderr, level=args.verbose.upper())
 
     start = time.time()
-    main(args.structures, args)
+    main(args)
     logging.info("TIME: {}".format(time.time() - start))