Fix issues #48 and #98 using Lark's interactive parser.

geographika · Dec 3, 2022 · 7f1c625 · 7f1c625
1 parent a7350ac
commit 7f1c625
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 11 deletions.
diff --git a/mappyfile/mapfile.lark b/mappyfile/mapfile.lark
@@ -49,9 +49,8 @@ _composite_item: (composite|attr|points|projection|pattern|values|config)
 !validation: "VALIDATION"i string_pair* _END
 !connectionoptions: "CONNECTIONOPTIONS"i string_pair* _END
 
-attr: (UNQUOTED_STRING | composite_type) (value | UNQUOTED_STRING)
-//attr: (UNQUOTED_STRING | SYMBOL) (value | UNQUOTED_STRING)
-// SYMBOL is listed in composite_type but is also an attribute name
+attr: (UNQUOTED_STRING | composite_type) (value | UNQUOTED_STRING | UNQUOTED_STRING_VALUE)
+%declare UNQUOTED_STRING_VALUE  // generated using the interactive parser. See issues #48, #98
 
 ?value: string | int | float | expression | not_expression | attr_bind | path
 | regexp | runtime_var | list | NULL | true | false | extent | rgb | hexcolor

diff --git a/mappyfile/parser.py b/mappyfile/parser.py
@@ -41,6 +41,7 @@
 
 log = logging.getLogger("mappyfile")
 
+SYMBOL_ATTRIBUTES = {'ANCHORPOINT', 'ANTIALIAS', 'FILLED', 'FONT', 'IMAGE', 'NAME', 'COLOR', 'TYPE', 'FONT', 'CHARACTER', 'POINTS', 'TRANSPARENT'}
 
 class Parser(object):
 
@@ -221,7 +222,18 @@ def parse(self, text, fn=None):
 
         try:
             self._comments[:] = []  # clear any comments from a previous parse
-            tree = self.lalr.parse(text)
+            ip = self.lalr.parse_interactive(text)
+            for t in ip.iter_parse():
+                if t.type == 'UNQUOTED_STRING':
+                    # Unquoted strings after SYMBOL can only be values, not attributes
+                    if ip.parser_state.value_stack[-1] == 'SYMBOL' and t.value.upper() not in SYMBOL_ATTRIBUTES:
+                        t.type = 'UNQUOTED_STRING_VALUE'
+                elif t.type == 'GRID':
+                    # Unquoted 'GRID' coming after NAME is always a value, not a composite type
+                    if ip.parser_state.value_stack[-1] == 'NAME':
+                        t.type = 'UNQUOTED_STRING_VALUE'
+
+            tree = ip.resume_parse()
             if self.include_comments:
                 self.assign_comments(tree, self._comments)
             return tree

diff --git a/mappyfile/transformer.py b/mappyfile/transformer.py
@@ -281,7 +281,7 @@ def attr(self, tokens):
 
         if isinstance(key_token, (list, tuple)):
             key_token = key_token[0]
-            assert self.key_name(key_token) in ("style", "symbol")
+            assert self.key_name(key_token) in ("style", "symbol"), self.key_name(key_token)
 
         key_name = self.key_name(key_token)
         value_tokens = tokens[1:]

diff --git a/tests/test_snippets.py b/tests/test_snippets.py
@@ -687,7 +687,6 @@ def test_symbol_style():
     assert(output(s, schema_name="style") == exp)
 
 
-@pytest.mark.xfail
 def test_symbol_style2():
     """
     barb_warm is not in quotes
@@ -700,8 +699,9 @@ def test_symbol_style2():
         SIZE 8
     END
     """
-    exp = "STYLE SYMBOL 'barb_warm' END"
-    assert(output(s, schema_name="style") == exp)
+    exp = "STYLE SYMBOL 'barb_warm' GAP -45 SIZE 8 END"
+    res = output(s, schema_name="style")
+    assert(res == exp)
 
 
 def test_extent():
@@ -981,9 +981,7 @@ def test_cluster2():
     assert(output(s, schema_name="layer") == exp)
 
 
-@pytest.mark.xfail
 def test_outputformat_unquoted_keyword():
-
     s = u"""
     MAP
         OUTPUTFORMAT
@@ -993,7 +991,6 @@ def test_outputformat_unquoted_keyword():
     END
     """
 
-    print(output(s, schema_name="map"))
     exp = u"MAP OUTPUTFORMAT NAME 'grid' IMAGEMODE INT16 END END"
     assert(output(s, schema_name="map") == exp)