Prepend colons to roles from parse_triples()

Fixes #80

CHANGELOG.md

@@ -12,6 +12,10 @@
 * `penman.interface` is removed from the public API but remains
   temporarily for backward compatibility ([#78])
 
+### Fixed
+
+* `parse_triples()` now prepends a colon to roles ([#80])
+
 ### Changed
 
 * Make `parse()`, `format()`, `interpret()`, and `configure()`
@@ -697,3 +701,4 @@ First release with very basic functionality.
 [#76]: https://github.com/goodmami/penman/issues/76
 [#77]: https://github.com/goodmami/penman/issues/77
 [#78]: https://github.com/goodmami/penman/issues/78
+[#80]: https://github.com/goodmami/penman/issues/80

penman/_parse.py

@@ -155,6 +155,8 @@ def _parse_triples(tokens: TokenIterator) -> List[BasicTriple]:
         role = tokens.expect('SYMBOL').text
         if strip_caret and role.startswith('^'):
             role = role[1:]
+        if not role.startswith(':'):
+            role = ':' + role
         tokens.expect('LPAREN')
         # SYMBOL may contain commas, so handle it here. If there
         # is no space between the comma and the next SYMBOL, they

tests/test_codec.py

@@ -18,7 +18,7 @@ def test_parse(self):
         assert codec.parse('(a / alpha)') == ('a', [('/', 'alpha')])
 
     def test_parse_triples(self):
-        assert codec.parse_triples('role(a, b)') == [('a', 'role', 'b')]
+        assert codec.parse_triples('role(a, b)') == [('a', ':role', 'b')]
 
     def test_format(self):
         assert codec.format(('a', [('/', 'alpha')])) == '(a / alpha)'