Fix #367 escape/unescape strings in simplemrs (#369)

Author: goodmami

CHANGELOG.md

@@ -6,6 +6,10 @@
 
 * Bump versions of `actions/setup-python` (and others) in actions ([#364])
 
+### Fixed
+
+* Escape/unescape quoted strings in SimpleMRS ([#367])
+
 
 ## [v1.8.0]
 
@@ -1592,3 +1596,4 @@ information about changes, except for
 [#357]: https://github.com/delph-in/pydelphin/issues/357
 [#360]: https://github.com/delph-in/pydelphin/issues/360
 [#364]: https://github.com/delph-in/pydelphin/issues/364
+[#367]: https://github.com/delph-in/pydelphin/issues/367

delphin/codecs/simplemrs.py

@@ -3,6 +3,7 @@
 """
 
 from pathlib import Path
+from typing import Optional
 
 from delphin.util import Lexer
 from delphin import predicate
@@ -184,7 +185,7 @@ def _decode_mrs(lexer):
     variables = {}
     lexer.expect_type(LBRACK)
     lnk = _decode_lnk(lexer)
-    surface = lexer.accept_type(DQSTRING)
+    surface = _decode_dqstring(lexer.accept_type(DQSTRING))
     feature = lexer.accept_type(FEATURE)
     while feature is not None:
         feature = feature.upper()
@@ -223,6 +224,12 @@ def _decode_lnk(lexer):
     return lnk
 
 
+def _decode_dqstring(dqstring: Optional[str]) -> Optional[str]:
+    if dqstring is not None:
+        dqstring = _unescape(dqstring)
+    return dqstring
+
+
 def _decode_variable(lexer, variables):
     var = lexer.expect_type(SYMBOL).lower()
     if var not in variables:
@@ -243,17 +250,16 @@ def _decode_rel(lexer, variables):
     args = {}
     surface = None
     lexer.expect_type(LBRACK)
-    pred = predicate.normalize(
-        lexer.choice_type(DQSTRING, SQSYMBOL, PREDICATE, SYMBOL)[1])
+    pred = _decode_predicate(lexer)
     lnk = _decode_lnk(lexer)
-    surface = lexer.accept_type(DQSTRING)
+    surface = _decode_dqstring(lexer.accept_type(DQSTRING))
     _, label = lexer.expect((FEATURE, 'LBL'), (SYMBOL, None))
     # any remaining are arguments or a constant
     role = lexer.accept_type(FEATURE)
     while role is not None:
         role = role.upper()
         if role == 'CARG':
-            value = lexer.expect_type(DQSTRING)
+            value = _decode_dqstring(lexer.expect_type(DQSTRING))
         else:
             value = _decode_variable(lexer, variables)
         args[role] = value
@@ -267,6 +273,15 @@ def _decode_rel(lexer, variables):
               base=None)
 
 
+def _decode_predicate(lexer) -> str:
+    predstring = lexer.accept_type(DQSTRING)
+    if predstring is not None:
+        predstring = _decode_dqstring(predstring)
+    else:
+        predstring = lexer.choice_type(SQSYMBOL, PREDICATE, SYMBOL)[1]
+    return predicate.normalize(predstring)
+
+
 def _decode_cons(lexer, cls, variables):
     lhs = _decode_variable(lexer, variables)
     relation = lexer.expect_type(SYMBOL).lower()
@@ -312,7 +327,7 @@ def _encode_surface_info(m, lnk):
         if m.lnk:
             tokens.append(str(m.lnk))
         if m.surface is not None:
-            tokens.append('"{}"'.format(m.surface))
+            tokens.append('"{}"'.format(_escape(m.surface)))
     return tokens
 
 
@@ -351,12 +366,12 @@ def _encode_rels(rels, varprops, lnk, indent):
             pred += str(rel.lnk)
         reltoks = ['[', pred]
         if lnk and rel.surface is not None:
-            reltoks.append('"{}"'.format(rel.surface))
+            reltoks.append('"{}"'.format(_escape(rel.surface)))
         reltoks.extend(('LBL:', rel.label))
         for role in sorted(rel.args, key=role_priority):
             arg = rel.args[role]
             if role == CONSTANT_ROLE:
-                arg = '"{}"'.format(arg)
+                arg = '"{}"'.format(_escape(arg))
             else:
                 arg = _encode_variable(arg, varprops)
             reltoks.extend((role + ':', arg))
@@ -383,3 +398,37 @@ def _encode_icons(icons, varprops):
     if tokens:
         tokens = ['ICONS: <'] + [' '.join(tokens)] + ['>']
     return tokens
+
+
+# Character Escaping
+
+
+_ESCAPES = {
+    '\\': '\\\\',
+    '"': '\\"',
+}
+
+
+_UNESCAPES = {
+    '\\\\': '\\',
+    '\\"': '"',
+}
+
+
+def _escape(s: str) -> str:
+    return "".join(_ESCAPES.get(c, c) for c in s)
+
+
+def _unescape(s: str) -> str:
+    if not s:
+        return s
+    cs = []
+    i = 0
+    while i < len(s):
+        if s[i] == '\\' and (i + 1) < len(s):
+            cs.append(s[i+1])
+            i += 2
+        else:
+            cs.append(s[i])
+            i += 1
+    return "".join(cs)

tests/codecs/simplemrs_test.py

@@ -60,3 +60,23 @@ def assert_predicate(p):
     assert_predicate(r'_24/7_n_1')
     assert_predicate(r'_foo<bar_n_1')
     assert_predicate(r'_foo_n_<1:3')
+
+
+def test_escapes_issue_367():
+    # https://github.com/delph-in/pydelphin/issues/367
+    m = simplemrs.decode(
+        '[ TOP: h0'
+        ' INDEX: e2 [ e SF: prop TENSE: pres MOOD: indicative PROG: - PERF: - ]'
+        ' RELS: < [ udef_q<0:6> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg ] RSTR: h5 BODY: h6 ]'
+        '         [ _blue_a_1<0:6> LBL: h7 ARG0: x3 ARG1: i8 ]'
+        '         [ _in_p_loc<10:12> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x9 [ x PERS: 3 NUM: sg IND: + ] ]'
+        '         [ _this_q_dem<13:17> LBL: h10 ARG0: x9 RSTR: h11 BODY: h12 ]'
+        '         [ _folder_n_of<18:25> LBL: h13 ARG0: x9 ARG1: i14 ] >'
+        ' HCONS: < h0 qeq h1 h5 qeq h7 h11 qeq h13 > ]'
+    )
+    m.surface = '"Blue" is in this folder.'
+    s = simplemrs.encode(m)
+    assert '\\"Blue\\" is in this folder.' in s
+    m2 = simplemrs.decode(s)
+    assert m == m2
+    assert m.surface == m2.surface