Merge branch 'fix-818' of https://github.com/MegaIng/lark into MegaIng-fix-818

lark/lark.py

@@ -185,7 +185,7 @@ class LarkOptions(Serialize):
 
 # Options that can be passed to the Lark parser, even when it was loaded from cache/standalone.
 # These option are only used outside of `load_grammar`.
-_LOAD_ALLOWED_OPTIONS = {'postlex', 'transformer', 'use_bytes', 'debug', 'g_regex_flags', 'regex', 'propagate_positions', 'tree_class'}
+_LOAD_ALLOWED_OPTIONS = {'postlex', 'transformer', 'lexer_callbacks', 'use_bytes', 'debug', 'g_regex_flags', 'regex', 'propagate_positions', 'tree_class'}
 
 _VALID_PRIORITY_OPTIONS = ('auto', 'normal', 'invert', None)
 _VALID_AMBIGUITY_OPTIONS = ('auto', 'resolve', 'explicit', 'forest')

lark/parser_frontends.py

@@ -43,9 +43,8 @@ class MakeParsingFrontend:
     def deserialize_lexer_conf(cls, data, memo, options):
         # We need lexer_conf earley to have the terminals that we need to produce the callback list for paser_conf
         # So we split deserialize into two methods
-        terminals = [item for item in memo.values() if isinstance(item, TerminalDef)]
         lexer_conf = LexerConf.deserialize(data['lexer_conf'], memo)
-        lexer_conf.callbacks = _get_lexer_callbacks(options.transformer, terminals)
+        lexer_conf.callbacks = options.lexer_callbacks or {}
         lexer_conf.re_module = regex if options.regex else re
         lexer_conf.use_bytes = options.use_bytes
         lexer_conf.g_regex_flags = options.g_regex_flags

tests/test_cache.py

@@ -55,6 +55,18 @@ class CustomLexer(Lexer):
             yield Token('A', obj)
 
 
+class TestT(Transformer):
+    def add(self, children):
+        return sum(children if isinstance(children, list) else children.children)
+
+    def NUM(self, token):
+        return int(token)
+
+
+def append_zero(t):
+    return t.update(value=t.value + '0')
+
+
 class TestCache(TestCase):
     def setUp(self):
         pass
@@ -80,7 +92,7 @@ class TestCache(TestCase):
             parser = Lark(g, parser='lalr', cache=True)
             assert parser.parse('a') == Tree('start', [])
 
-            parser = Lark(g+' "b"', parser='lalr', cache=True)
+            parser = Lark(g + ' "b"', parser='lalr', cache=True)
             assert len(mock_fs.files) == 2
             assert parser.parse('ab') == Tree('start', [])
 
@@ -100,7 +112,7 @@ class TestCache(TestCase):
             parser = Lark(g, parser="lalr", debug=True, cache=True)
             assert parser.options.options['debug']
 
-            # Test inline transformer (tree-less)
+            # Test inline transformer (tree-less) & lexer_callbacks
             mock_fs.files = {}
             g = """
             start: add+
@@ -109,19 +121,19 @@ class TestCache(TestCase):
             %ignore " "
             """
             text = "1+2 3+4"
-            expected = Tree('start', [3, 7])
+            expected = Tree('start', [30, 70])
 
-            parser = Lark(g, parser='lalr', transformer=TestT(), cache=True)
-            parser = Lark(g, parser='lalr', transformer=TestT(), cache=True)
+            parser = Lark(g, parser='lalr', transformer=TestT(), cache=True, lexer_callbacks={'NUM': append_zero})
+            res0 = parser.parse(text)
+            parser = Lark(g, parser='lalr', transformer=TestT(), cache=True, lexer_callbacks={'NUM': append_zero})
             assert len(mock_fs.files) == 1
             res1 = parser.parse(text)
-            res2 = TestT().transform( Lark(g, parser="lalr", cache=True).parse(text) )
-            assert res1 == res2
+            res2 = TestT().transform(Lark(g, parser="lalr", cache=True, lexer_callbacks={'NUM': append_zero}).parse(text))
+            assert res0 == res1 == res2 == expected
 
         finally:
             lark_module.FS = fs
 
 
-
 if __name__ == '__main__':
     main()