Refactor array handling and add tests

kuksa-client/kuksa_client/grpc/__init__.py

@@ -315,6 +315,27 @@ def from_message(cls, message: types_pb2.Datapoint):
             ) if message.HasField('timestamp') else None,
         )
 
+
+    def cast_array_values(cast, array):
+
+        array = array.strip('[]')
+        pattern = r'(?:\\"|[^",])*"(?:\\"|[^"])*"|[^",]+'
+        values = re.findall(pattern, array)
+        for item in values:
+            if item == '':
+                #skip
+                pass
+            else:
+                yield cast(item)
+
+    def cast_bool(value) -> bool:
+        if value in ('False', 'false', 'F', 'f'):
+            value = 0
+        return bool(value)
+
+    def cast_str(value) -> str:
+        return str(value).replace('\"', '').replace('\\', '"').strip()
+
     def to_message(self, value_type: DataType) -> types_pb2.Datapoint:
         message = types_pb2.Datapoint()
 
@@ -323,27 +344,6 @@ def set_array_attr(obj, attr, values):
             array.Clear()
             array.values.extend(values)
 
-        def cast_array_values(cast, array):
-            array = array.strip('[]')
-            pattern = r'(?:\\.|[^",])*"(?:\\.|[^"])*"|[^",]+'
-            values = re.findall(pattern, array)
-            for item in values:
-                if item == '':
-                    #skip
-                    pass
-                else:
-                    if cast == str:
-                        item = item.replace('\"', '').replace('\\', '"').strip()
-                    yield cast(item)
-
-        def cast_bool(value):
-            if value in ('False', 'false', 'F', 'f'):
-                value = 0
-            return bool(value)
-
-        def cast_str(value):
-            return str(value).replace('\"', '').replace('\\', '"').strip()
-
 
         field, set_field, cast_field = {
             DataType.INT8: ('int32', setattr, int),
@@ -356,20 +356,32 @@ def cast_str(value):
             DataType.INT64: ('int64', setattr, int),
             DataType.FLOAT: ('float', setattr, float),
             DataType.DOUBLE: ('double', setattr, float),
-            DataType.BOOLEAN: ('bool', setattr, cast_bool),
-            DataType.STRING: ('string', setattr, cast_str),
-            DataType.INT8_ARRAY: ('int32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.INT16_ARRAY: ('int32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.INT32_ARRAY: ('int32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.UINT8_ARRAY: ('uint32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.UINT16_ARRAY: ('uint32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.UINT32_ARRAY: ('uint32_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.UINT64_ARRAY: ('uint64_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.INT64_ARRAY: ('int64_array', set_array_attr, lambda array: cast_array_values(int, array)),
-            DataType.FLOAT_ARRAY: ('float_array', set_array_attr, lambda array: cast_array_values(float, array)),
-            DataType.DOUBLE_ARRAY: ('double_array', set_array_attr, lambda array: cast_array_values(float, array)),
-            DataType.BOOLEAN_ARRAY: ('bool_array', set_array_attr, lambda array: cast_array_values(cast_bool, array)),
-            DataType.STRING_ARRAY: ('string_array', set_array_attr, lambda array: cast_array_values(str, array)),
+            DataType.BOOLEAN: ('bool', setattr, Datapoint.cast_bool),
+            DataType.STRING: ('string', setattr, Datapoint.cast_str),
+            DataType.INT8_ARRAY: ('int32_array', set_array_attr, 
+                                  lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.INT16_ARRAY: ('int32_array', set_array_attr, 
+                                   lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.INT32_ARRAY: ('int32_array', set_array_attr, 
+                                   lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.UINT8_ARRAY: ('uint32_array', set_array_attr, 
+                                   lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.UINT16_ARRAY: ('uint32_array', set_array_attr, 
+                                    lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.UINT32_ARRAY: ('uint32_array', set_array_attr, 
+                                    lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.UINT64_ARRAY: ('uint64_array', set_array_attr, 
+                                    lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.INT64_ARRAY: ('int64_array', set_array_attr, 
+                                   lambda array: Datapoint.cast_array_values(int, array)),
+            DataType.FLOAT_ARRAY: ('float_array', set_array_attr, 
+                                   lambda array: Datapoint.cast_array_values(float, array)),
+            DataType.DOUBLE_ARRAY: ('double_array', set_array_attr, 
+                                    lambda array: Datapoint.cast_array_values(float, array)),
+            DataType.BOOLEAN_ARRAY: ('bool_array', set_array_attr, 
+                                     lambda array: Datapoint.cast_array_values(Datapoint.cast_bool, array)),
+            DataType.STRING_ARRAY: ('string_array', set_array_attr, 
+                                    lambda array: Datapoint.cast_array_values(Datapoint.cast_str, array)),
         }.get(value_type, (None, None, None))
         if self.value is not None:
             if all((field, set_field, cast_field)):

kuksa-client/tests/test_datapoint.py

@@ -0,0 +1,110 @@
+# /********************************************************************************
+# * Copyright (c) 2023 Contributors to the Eclipse Foundation
+# *
+# * See the NOTICE file(s) distributed with this work for additional
+# * information regarding copyright ownership.
+# *
+# * This program and the accompanying materials are made available under the
+# * terms of the Apache License 2.0 which is available at
+# * http://www.apache.org/licenses/LICENSE-2.0
+# *
+# * SPDX-License-Identifier: Apache-2.0
+# ********************************************************************************/
+
+import pytest
+from kuksa_client.grpc import Datapoint
+
+#
+# Client rules:
+# For simple strings like abd it is optional to quote them ("abc") or not (abc)
+# Quotes are needed if you have commas ("ab, c")
+# If you have duoble quotes in strings you must escape them
+
+def test_array_parse_no_quote():
+    test_str = r'[say hello, abc]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say hello"
+    assert my_array[1] == "abc"
+
+def test_array_parse_no_inside_quote():
+    test_str = r'["say hello","abc"]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say hello"
+    assert my_array[1] == "abc"
+
+def test_array_parse_double_quote():
+    test_str = r'["say \"hello\"","abc"]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say \"hello\""
+    assert my_array[1] == "abc"
+
+def test_array_parse_double_quote_2():
+    test_str = r'[say \"hello\",abc]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say \"hello\""
+    assert my_array[1] == "abc"
+
+def test_array_parse_comma():
+    test_str = r'["say, hello","abc"]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say, hello"
+    assert my_array[1] == "abc"
+
+def test_array_square():
+    """No problem having square brackets as part of strings"""
+    test_str = r'[say hello[], abc]'
+    my_array = list(Datapoint.cast_array_values(Datapoint.cast_str,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == "say hello[]"
+    assert my_array[1] == "abc"
+
+def test_int_no_quote():
+    test_str = r'[123,456]'
+    my_array = list(Datapoint.cast_array_values(int,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == 123
+    assert my_array[1] == 456
+
+def test_int_quote():
+    """Quoting individual int values is not allowed"""
+    test_str = r'["123","456"]'
+    with pytest.raises(ValueError):
+        my_array = list(Datapoint.cast_array_values(int,test_str))
+
+
+def test_float_no_quote():
+    test_str = r'[123,456.23]'
+    my_array = list(Datapoint.cast_array_values(float,test_str))
+    assert len(my_array) == 2
+    assert my_array[0] == 123
+    assert my_array[1] == 456.23
+
+def test_cast_str():
+    """Unquoted quotation marks shall be removed, quoted kept without quotes"""
+    test_str = r'"say hello"'
+    assert Datapoint.cast_str(test_str) == r'say hello'
+    test_str = r'"say \"hello\""'
+    assert Datapoint.cast_str(test_str) == r'say "hello"'
+    test_str = r'say "hello"'
+    assert Datapoint.cast_str(test_str) == r'say hello'
+
+def test_cast_bool():
+    assert Datapoint.cast_bool("true") is True
+    assert Datapoint.cast_bool("True") is True
+    assert Datapoint.cast_bool("T") is True
+    assert Datapoint.cast_bool("t") is True
+    assert Datapoint.cast_bool("false") is False
+    assert Datapoint.cast_bool("False") is False
+    assert Datapoint.cast_bool("F") is False
+    assert Datapoint.cast_bool("f") is False
+
+    # And then some other, treated as true for now
+    assert Datapoint.cast_bool("Ja") is True
+    assert Datapoint.cast_bool("Nein") is True
+    assert Datapoint.cast_bool("Doch") is True
+