Merge pull request #152 from jacebrowning/pyyaml-support

Add support for using PyYAML as the formatter

CHANGELOG.md

@@ -1,3 +1,7 @@
+# 0.7 (unreleased)
+
+- Added a `YAML_LIBRARY` setting to control the underlying YAML library.
+
 # 0.6 (2020-01-25)
 
 - Added a registration system for custom formatter classes.

datafiles/formats.py

@@ -1,3 +1,5 @@
+# pylint: disable=import-outside-toplevel
+
 import json
 from abc import ABCMeta, abstractmethod
 from contextlib import suppress
@@ -6,8 +8,6 @@
 from typing import IO, Any, Dict, List
 
 import log
-import tomlkit
-from ruamel import yaml
 
 from . import settings
 
@@ -64,22 +64,28 @@ def extensions(cls):
 
     @classmethod
     def deserialize(cls, file_object):
+        import tomlkit
+
         return tomlkit.loads(file_object.read()) or {}
 
     @classmethod
     def serialize(cls, data):
+        import tomlkit
+
         return tomlkit.dumps(data)
 
 
-class YAML(Formatter):
-    """Formatter for (safe, round-trip) YAML Ain't Markup Language."""
+class RuamelYAML(Formatter):
+    """Formatter for (round-trip) YAML Ain't Markup Language."""
 
     @classmethod
     def extensions(cls):
         return {'.yml', '.yaml'}
 
     @classmethod
     def deserialize(cls, file_object):
+        from ruamel import yaml
+
         try:
             return yaml.YAML(typ='rt').load(file_object) or {}
         except NotImplementedError as e:
@@ -88,6 +94,8 @@ def deserialize(cls, file_object):
 
     @classmethod
     def serialize(cls, data):
+        from ruamel import yaml
+
         if settings.INDENT_YAML_BLOCKS:
             f = StringIO()
             y = yaml.YAML()
@@ -100,6 +108,44 @@ def serialize(cls, data):
         return "" if text == "{}\n" else text
 
 
+class PyYAML(Formatter):
+    """Formatter for YAML Ain't Markup Language."""
+
+    @classmethod
+    def extensions(cls):
+        return {'.yml', '.yaml'}
+
+    @classmethod
+    def deserialize(cls, file_object):
+        import yaml
+
+        data = yaml.safe_load(file_object)
+
+        return data
+
+    @classmethod
+    def serialize(cls, data):
+        import yaml
+
+        def represent_none(self, _):
+            return self.represent_scalar('tag:yaml.org,2002:null', '')
+
+        yaml.add_representer(type(None), represent_none)
+
+        class Dumper(yaml.Dumper):
+            def increase_indent(self, flow=False, indentless=False):
+                return super().increase_indent(
+                    flow=flow,
+                    indentless=False if settings.INDENT_YAML_BLOCKS else indentless,
+                )
+
+        text = yaml.dump(  # type: ignore
+            data, Dumper=Dumper, sort_keys=False, default_flow_style=False
+        )
+
+        return text
+
+
 def deserialize(path: Path, extension: str) -> Dict:
     formatter = _get_formatter(extension)
     with path.open('r') as file_object:
@@ -112,6 +158,9 @@ def serialize(data: Dict, extension: str = '.yml') -> str:
 
 
 def _get_formatter(extension: str):
+    if settings.YAML_LIBRARY == 'PyYAML':
+        register('.yml', PyYAML)
+
     with suppress(KeyError):
         return _REGISTRY[extension]
 

datafiles/settings.py

@@ -1,5 +1,9 @@
 """Shared configuration flags."""
 
 HIDE_TRACEBACK_IN_HOOKS = True
+
 HOOKS_ENABLED = True
+
 INDENT_YAML_BLOCKS = True
+
+YAML_LIBRARY = 'ruamel.yaml'

datafiles/tests/test_formats.py

@@ -11,9 +11,9 @@ def describe_serialize():
     def data():
         return {'key': "value", 'items': [1, 'a', None]}
 
-    def describe_yaml():
+    def describe_ruamel_yaml():
         def it_indents_blocks_by_default(expect, data):
-            text = formats.YAML.serialize(data)
+            text = formats.RuamelYAML.serialize(data)
             expect(text) == dedent(
                 """
             key: value
@@ -26,7 +26,33 @@ def it_indents_blocks_by_default(expect, data):
 
         def it_can_render_lists_inline(expect, data, monkeypatch):
             monkeypatch.setattr(settings, 'INDENT_YAML_BLOCKS', False)
-            text = formats.YAML.serialize(data)
+            text = formats.RuamelYAML.serialize(data)
+            expect(text) == dedent(
+                """
+            key: value
+            items:
+            - 1
+            - a
+            -
+            """
+            )
+
+    def describe_pyyaml():
+        def it_indents_blocks_by_default(expect, data):
+            text = formats.PyYAML.serialize(data)
+            expect(text) == dedent(
+                """
+            key: value
+            items:
+              - 1
+              - a
+              -
+            """
+            )
+
+        def it_can_render_lists_inline(expect, data, monkeypatch):
+            monkeypatch.setattr(settings, 'INDENT_YAML_BLOCKS', False)
+            text = formats.PyYAML.serialize(data)
             expect(text) == dedent(
                 """
             key: value

docs/settings.md

@@ -52,3 +52,10 @@ items:
 - 2
 - 3
 ```
+
+# `YAML_LIBRARY`
+
+This setting controls the underlying YAML library used to read and write files. The following options are available:
+
+- `'ruamel.yaml'` (default)
+- `'PyYAML'`

pyproject.toml

@@ -1,7 +1,7 @@
 [tool.poetry]
 
 name = "datafiles"
-version = "0.7b1"
+version = "0.7b2"
 description = "File-based ORM for dataclasses."
 
 license = "MIT"
@@ -42,6 +42,7 @@ classifiers = [
 python = "^3.7"
 
 # Formats
+PyYAML = "^5.3"
 "ruamel.yaml" = "^0.16.7"
 tomlkit = "^0.5.3"
 
@@ -70,7 +71,7 @@ pydocstyle = "*"
 # Testing
 pytest = "^5.3.2"
 pytest-describe = "*"
-pytest-expecter = "*"
+pytest-expecter = "^2.0"
 pytest-mock = "*"
 pytest-random = "*"
 pytest-repeat = "*"