🖊️ Trim leading and trailing spaces in assignments (#5936)

With this PR we explicitly trim leading and trailing spaces in assignments in levels 2-11. Note that in these levels the assignment and list assignment statements do not quotes. The program `a is three spaces   ` used to be transpiled to `a = 'three spaces   '` but is now transpiled to `a = 'three spaces'`. Similarly, the list `a is  cat , dog ,  chicken ` is now transpiled to `a = ['cat', 'dog', 'chicken']`.

Fixes #5684

**How to test**
Ensure that the following snippet yields `[4]` twice in level 9. Ensure that there is no grammar ambiguity. 
```
i = 4   # met comment
print '[' i ']'
if 1 = 1
    j = 4    # met comment
    print '[' j ']'
```

Author: boryanagoncharenko

grammars/level2-Additions.lark

@@ -20,10 +20,11 @@ sleep: _SLEEP (INT | var_access)?
 
 //in level 2, forward may also depend on a variable
 turtle: _FORWARD (NUMBER | textwithoutspaces)? -> forward | _TURN ((NUMBER | textwithoutspaces))? -> turn | _COLOR ((black | blue | brown | gray | green | orange | pink | purple | red | white | yellow | textwithoutspaces))? -> color
-assign: var _IS text -> assign
+assign: var _IS textwithinnerspaces _SPACE? -> assign
 
 textwithoutspaces: /([^\n #]+)/ -> text
 text: /([^\n#]+)/ -> text
+textwithinnerspaces: /([^\n#]*[^\n# ])/ -> text
 
 var: NAME // used for variable definitions, e.g. a = 1
 var_access: NAME // used for variable references, e.g. for i in range. It parses the same as var, but does not result in a lookup table entry

grammars/level3-Additions.lark

@@ -1,8 +1,8 @@
 command:+= assign_list | add | remove | error_add_missing_to | error_remove_missing_from | error_add_missing_list | error_remove_missing_list >> error_invalid
 _print_ask_argument: (_SPACE | (list_access textwithoutspaces?) | textwithoutspaces)*
 
-assign: var _IS (list_access | text) -> assign
-assign_list: var _IS text_list (_COMMA text_list)+
+assign: var _IS (list_access | textwithinnerspaces) _SPACE? -> assign
+assign_list: var _IS textwithinnerspaces (_COMMA textwithinnerspaces)+ _SPACE?
 play: _PLAY (list_access | textwithoutspaces)
 
 list_access: var_access _AT (INT | random)
@@ -11,7 +11,8 @@ turtle: _FORWARD ((NUMBER | list_access | textwithoutspaces))? -> forward | _TUR
 sleep: _SLEEP (INT | list_access | var_access)?
 // lists are introduced and list separators (comma and arabic comma) have to excluded from text.
 text: /([^\n،,，、#]+)/ -> text
-text_list: /([^\n,،，、#]+)/ -> text // list elements are another exception since they can contain punctuation but not list separators
+// list elements are another exception since they can contain punctuation but not list separators. Also, they cannot end with space.
+textwithinnerspaces: /([^\n،,，、#]*[^\n،,，、# ])/ -> text
 
 // Values which are added or removed from lists can contain spaces, so we need to escape the 'to_list' and 'from' keywords
 // The part " (?!<expand_keyword to_list>|<expand_keyword from>)" will be processed to a space with a negative look ahead of all to_list and from values

grammars/level5-Additions.lark

@@ -12,9 +12,6 @@ _if_less_command: print | ask | play | turtle | assign_list | add | remove | sle
 // just to be sure that if Lark tries to resolve ambiguity, error_invalid will be considered last
 error_invalid.-100: textwithoutspaces _SPACE* (quoted_text | textwithspaces)?
 
-assign_list: var _IS textwithspaces (_COMMA textwithspaces)+
-assign: var _IS (list_access | textwithspaces)
-
 error_print_no_quotes: _PRINT (textwithoutspaces | list_access | var_access) (_SPACE (textwithoutspaces | list_access | var_access))* -> error_print_nq
 
 // new commands for level 5

grammars/level6-Additions.lark

@@ -13,8 +13,8 @@ play: _PLAY (list_access | expression | textwithoutspaces)
 equality_check: (INT | textwithoutspaces) (_IS | _EQUALS) (INT | quoted_text | textwithoutspaces)
 condition: >> equality_check
 
-assign_list: var (_IS | _EQUALS) (INT | textwithspaces) (_COMMA (INT | textwithspaces))+
-assign: var (_IS | _EQUALS) (INT | list_access | expression | textwithoutspaces | textwithspaces)
+assign_list: var (_IS | _EQUALS) (INT | textwithinnerspaces) (_COMMA (INT | textwithinnerspaces))+ _SPACE?
+assign: var (_IS | _EQUALS) (INT | list_access | expression | textwithoutspaces | textwithinnerspaces) _SPACE?
 
 add: _ADD_LIST (INT | text_add_remove_list) _TO_LIST _SPACE var_access
 remove: _REMOVE (INT | text_add_remove_list) _FROM _SPACE var_access

tests/Tester.py

@@ -388,10 +388,12 @@ def source_map_tester(self, code, expected_source_map: dict):
         result = hedy.transpile(code, self.level, 'en')
         self.assertDictEqual(result.source_map.get_compressed_mapping(), expected_source_map)
 
-    def assert_translated_code_equal(self, orignal, translation):
+    def assert_translated_code_equal(self, original, translation):
         # When we translate a program we lose information about the whitespaces of the original program.
         # So when comparing the original and the translated code, we compress multiple whitespaces into one.
-        self.assertEqual(re.sub('\\s+', ' ', orignal), re.sub('\\s+', ' ', translation))
+        # Also, we lose the trailing spaces, so we strip before comparing.
+        self.assertEqual(re.sub('\\s+', ' ', original).strip(),
+                         re.sub('\\s+', ' ', translation).strip())
 
     @staticmethod
     def validate_Python_code(parseresult):

tests/test_level/test_level_02.py

@@ -242,7 +242,7 @@ def test_print_comment(self):
 
     def test_assign_comment(self):
         code = "test is Welkom bij Hedy # This is a comment"
-        expected = "test = 'Welkom bij Hedy '"
+        expected = "test = 'Welkom bij Hedy'"
         self.multi_level_tester(
             max_level=3,
             code=code,

tests/test_level/test_level_03.py

@@ -324,6 +324,18 @@ def test_assign_var_to_var(self):
 
         self.multi_level_tester(max_level=5, code=code, expected=expected)
 
+    def test_assign_var_trims_spaces(self):
+        code = "answer is  This is long    "
+        expected = "answer = 'This is long'"
+
+        self.multi_level_tester(max_level=5, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_var_trims_spaces_with_comment(self):
+        code = "answer is  This is long    # comment"
+        expected = "answer = 'This is long'"
+
+        self.multi_level_tester(max_level=5, code=code, expected=expected, unused_allowed=True)
+
     def test_assign_list(self):
         code = "dieren is Hond, Kat, Kangoeroe"
         expected = "dieren = ['Hond', 'Kat', 'Kangoeroe']"
@@ -354,11 +366,21 @@ def test_assign_list_to_hungarian_var(self):
 
         self.multi_level_tester(max_level=5, code=code, expected=expected)
 
-    def test_assign_list_with_spaces(self):
-        # spaces are parsed in the text here, that is fine (could be avoided if we say text
-        # can't *end* (or start) in a space but I find this ok for now
-        code = "dieren is Hond , Kat , Kangoeroe"
-        expected = "dieren = ['Hond ', 'Kat ', 'Kangoeroe']"
+    def test_assign_list_trims_elements_trailing_spaces(self):
+        code = "dieren is Hond , Kat , Kangoeroe "
+        expected = "dieren = ['Hond', 'Kat', 'Kangoeroe']"
+
+        self.multi_level_tester(max_level=5, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_list_trims_elements_leading_spaces(self):
+        code = "dieren is   Hond,   Kat,   Kangoeroe"
+        expected = "dieren = ['Hond', 'Kat', 'Kangoeroe']"
+
+        self.multi_level_tester(max_level=5, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_list_trims_elements_spaces(self):
+        code = "dieren is   I am  ,  waiting for  ,  the summer  "
+        expected = "dieren = ['I am', 'waiting for', 'the summer']"
 
         self.multi_level_tester(max_level=5, code=code, expected=expected, unused_allowed=True)
 

tests/test_level/test_level_04.py

@@ -517,7 +517,7 @@ def test_print_comment(self):
 
     def test_assign_comment(self):
         code = 'test is "Welkom bij Hedy" # This is a comment'
-        expected = 'test = \'"Welkom bij Hedy" \''
+        expected = """test = '"Welkom bij Hedy"'"""
         self.multi_level_tester(
             max_level=5,
             unused_allowed=True,

tests/test_level/test_level_06.py

@@ -927,7 +927,7 @@ def test_assign_catalan_var_name(self):
 
     def test_assign_comment(self):
         code = 'test is "Welkom bij Hedy" # This is a comment'
-        expected = 'test = Value(\'"Welkom bij Hedy" \')'
+        expected = """test = Value('"Welkom bij Hedy"')"""
         self.multi_level_tester(
             max_level=11,
             unused_allowed=True,
@@ -954,6 +954,18 @@ def test_assign_var_to_var(self):
 
         self.multi_level_tester(max_level=11, code=code, expected=expected)
 
+    def test_assign_var_trims_spaces(self):
+        code = "answer is  This is long    "
+        expected = "answer = Value('This is long')"
+
+        self.multi_level_tester(max_level=11, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_var_trims_spaces_with_comment(self):
+        code = "answer is  This is long    # comment"
+        expected = "answer = Value('This is long')"
+
+        self.multi_level_tester(max_level=11, code=code, expected=expected, unused_allowed=True)
+
     def test_assign_text_with_inner_single_quote(self):
         code = "var is Hedy's"
         expected = "var = Value('Hedy\\'s')"
@@ -1042,11 +1054,21 @@ def test_assign_list_with_arabic_comma_and_is(self):
             lang='ar'
         )
 
-    def test_assign_list_with_spaces(self):
-        # spaces are parsed in the text here, that is fine (could be avoided if we say text
-        # can't *end* (or start) in a space but I find this ok for now
-        code = "dieren is Hond , Kat , Kangoeroe"
-        expected = "dieren = Value([Value('Hond '), Value('Kat '), Value('Kangoeroe')])"
+    def test_assign_list_trims_elements_trailing_spaces(self):
+        code = "dieren is Hond , Kat , Kangoeroe  "
+        expected = "dieren = Value([Value('Hond'), Value('Kat'), Value('Kangoeroe')])"
+
+        self.multi_level_tester(max_level=11, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_list_trims_elements_leading_spaces(self):
+        code = "dieren is   Hond,   Kat,   Kangoeroe"
+        expected = "dieren = Value([Value('Hond'), Value('Kat'), Value('Kangoeroe')])"
+
+        self.multi_level_tester(max_level=11, code=code, expected=expected, unused_allowed=True)
+
+    def test_assign_list_trims_elements_spaces(self):
+        code = "dieren is   I am  ,  waiting for  ,  the summer  "
+        expected = "dieren = Value([Value('I am'), Value('waiting for'), Value('the summer')])"
 
         self.multi_level_tester(max_level=11, code=code, expected=expected, unused_allowed=True)
 

tests/test_level/test_level_08.py

@@ -385,6 +385,32 @@ def test_equality_arabic_and_latin_vars(self):
 
         self.multi_level_tester(code=code, expected=expected, max_level=11, output='jahoor!')
 
+    def test_assign_var_in_if_trims_spaces(self):
+        code = self.dedent(
+            "if 1 = 1",
+            ("j = 4   ", "  "),
+            ("    print '[' j ']'", "  "))
+
+        expected = textwrap.dedent(f"""\
+        if localize('1') == localize('1'):
+          j = Value('4', num_sys='Latin')
+          print(f'[{{j}}]')""")
+
+        self.multi_level_tester(code=code, expected=expected, max_level=11, output='[4]')
+
+    def test_assign_var_in_if_trims_spaces_with_comment(self):
+        code = textwrap.dedent("""\
+        if 1 = 1
+            j = 4    # met comment
+            print '[' j ']'""")
+
+        expected = textwrap.dedent(f"""\
+        if localize('1') == localize('1'):
+          j = Value('4', num_sys='Latin')
+          print(f'[{{j}}]')""")
+
+        self.multi_level_tester(code=code, expected=expected, max_level=11, output='[4]')
+
     #
     # in/not in list
     #

tests/test_level/test_level_12.py

@@ -938,12 +938,7 @@ def test_print_comment(self):
     def test_assign_comment(self):
         code = 'test = "Welkom bij Hedy" # This is a comment'
         expected = "test = Value('Welkom bij Hedy')"
-        self.multi_level_tester(
-            max_level=18,
-            code=code,
-            expected=expected,
-            unused_allowed=True
-        )
+        self.multi_level_tester(code=code, expected=expected, unused_allowed=True)
 
     #
     # ask tests