Merge pull request #67 from corytodd/yaml-support

Introduce YAML support

README.rst

@@ -68,17 +68,22 @@ Command Line Client
 
 Usage::
 
-    jdiff [-h] [-p] [-s SYNTAX] [-i INDENT] first second
+    jdiff [-h] [-p] [-s {compact,symmetric,explicit}] [-i INDENT] [-f {json,yaml}] first second
 
-    positional arguments:
-      first
-      second
+positional arguments:
+  first
+  second
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -p, --patch
+  -s {compact,symmetric,explicit}, --syntax {compact,symmetric,explicit}
+                        Diff syntax controls how differences are rendered (default: compact)
+  -i INDENT, --indent INDENT
+                        Number of spaces to indent. None is compact, no indentation. (default: None)
+  -f {json,yaml}, --format {json,yaml}
+                        Specify file format for input and dump (default: json)
 
-    optional arguments:
-      -h, --help            show this help message and exit
-      -p, --patch
-      -s SYNTAX, --syntax SYNTAX
-      -i INDENT, --indent INDENT
 
 Examples:
 
@@ -87,3 +92,5 @@ Examples:
     $ jdiff a.json b.json -i 2
 
     $ jdiff a.json b.json -i 2 -s symmetric
+
+    $ jdiff a.yaml b.yaml -f yaml -s symmetric

jsondiff/__init__.py

@@ -2,6 +2,7 @@
 
 import sys
 import json
+import yaml
 
 from .symbols import *
 from .symbols import Symbol
@@ -35,6 +36,20 @@ def __call__(self, obj, dest=None):
 default_dumper = JsonDumper()
 
 
+class YamlDumper(object):
+    """Write object as YAML string"""
+
+    def __init__(self, **kwargs):
+        self.kwargs = kwargs
+
+    def __call__(self,  obj, dest=None):
+        """Format obj as a YAML string and optionally write to dest
+        :param obj: dict to dump
+        :param dest: file-like object
+        :return: str
+        """
+        return yaml.dump(obj, dest, **self.kwargs)
+
 class JsonLoader(object):
     def __init__(self, **kwargs):
         self.kwargs = kwargs
@@ -49,6 +64,16 @@ def __call__(self, src):
 default_loader = JsonLoader()
 
 
+class YamlLoader(object):
+    """Load YAML data from file-like object or string"""
+
+    def __call__(self, src):
+        """Parse and return YAML data
+        :param src: str|file-like source
+        :return: dict parsed data
+        """
+        return yaml.safe_load(src)
+
 class JsonDiffSyntax(object):
     def emit_set_diff(self, a, b, s, added, removed):
         raise NotImplementedError()
@@ -641,4 +666,6 @@ def similarity(a, b, cls=JsonDiffer, **kwargs):
     "JsonDiffer",
     "JsonDumper",
     "JsonLoader",
+    "YamlDumper",
+    "YamlLoader"
 ]

jsondiff/cli.py

@@ -1,24 +1,35 @@
 import argparse
 import jsondiff
-import json
-import warnings
 import sys
 
+
 def main():
-    parser = argparse.ArgumentParser()
+    parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
 
     parser.add_argument("first")
     parser.add_argument("second")
     parser.add_argument("-p", "--patch", action="store_true", default=False)
-    parser.add_argument("-s", "--syntax", action="store", type=str, default="compact")
-    parser.add_argument("-i", "--indent", action="store", type=int, default=None)
+    parser.add_argument("-s", "--syntax", choices=(jsondiff.builtin_syntaxes.keys()), default="compact",
+                        help="Diff syntax controls how differences are rendered")
+    parser.add_argument("-i", "--indent", action="store", type=int, default=None,
+                        help="Number of spaces to indent. None is compact, no indentation.")
+    parser.add_argument("-f", "--format", choices=("json", "yaml"), default="json",
+                        help="Specify file format for input and dump")
 
     args = parser.parse_args()
 
+    # pyyaml _can_ load json but is ~20 times slower and has known issues so use
+    # the json from stdlib when json is specified.
+    serializers = {
+        "json": (jsondiff.JsonLoader(), jsondiff.JsonDumper(indent=args.indent)),
+        "yaml": (jsondiff.YamlLoader(), jsondiff.YamlDumper(indent=args.indent)),
+    }
+    loader, dumper = serializers[args.format]
+
     with open(args.first, "r") as f:
         with open(args.second, "r") as g:
-            jf = json.load(f)
-            jg = json.load(g)
+            jf = loader(f)
+            jg = loader(g)
             if args.patch:
                 x = jsondiff.patch(
                     jf,
@@ -34,7 +45,7 @@ def main():
                     syntax=args.syntax
                 )
 
-            json.dump(x, sys.stdout, indent=args.indent)
+            dumper(x, sys.stdout)
 
 
 if __name__ == '__main__':

pyproject.toml

@@ -17,6 +17,9 @@ classifiers = [
     'License :: OSI Approved :: MIT License',
     'Programming Language :: Python :: 3',
 ]
+dependencies = [
+    "pyyaml"
+]
 
 [project.optional-dependencies]
 test = [

tests/data/test_01.json

@@ -0,0 +1,6 @@
+{
+  "hello": "world",
+  "data": [
+    1,2,3
+  ]
+}

tests/data/test_01.yaml

@@ -0,0 +1,5 @@
+hello: world
+data:
+  - 1
+  - 2
+  - 3

tests/test_jsondiff.py

@@ -1,8 +1,11 @@
+import io
 import logging
+import os.path
 import sys
 import unittest
 import pytest
 
+import jsondiff
 from jsondiff import diff, replace, add, discard, insert, delete, update, JsonDiffer
 
 from .utils import generate_random_json, perturbate_json
@@ -208,3 +211,110 @@ def test_right_only_syntax(self):
             diff(c, d, syntax="rightonly", marshal=True),
             [4, 5, 6],
         )
+
+class TestLoaders(unittest.TestCase):
+
+    here = os.path.dirname(__file__)
+    data_dir = os.path.join(here, "data")
+
+    def test_json_string_loader(self):
+        json_str = '{"hello": "world", "data": [1,2,3]}'
+        expected = {"hello": "world", "data": [1, 2, 3]}
+        loader = jsondiff.JsonLoader()
+        actual = loader(json_str)
+        self.assertEqual(expected, actual)
+
+    def test_json_file_loader(self):
+        json_file = os.path.join(TestLoaders.data_dir, "test_01.json")
+        expected = {"hello": "world", "data": [1, 2, 3]}
+        loader = jsondiff.JsonLoader()
+        with open(json_file) as f:
+            actual = loader(f)
+        self.assertEqual(expected, actual)
+
+    def test_yaml_string_loader(self):
+        json_str = """---
+hello: world
+data:
+  - 1
+  - 2
+  - 3
+        """
+        expected = {"hello": "world", "data": [1, 2, 3]}
+        loader = jsondiff.YamlLoader()
+        actual = loader(json_str)
+        self.assertEqual(expected, actual)
+
+    def test_yaml_file_loader(self):
+        yaml_file = os.path.join(TestLoaders.data_dir, "test_01.yaml")
+        expected = {"hello": "world", "data": [1, 2, 3]}
+        loader = jsondiff.YamlLoader()
+        with open(yaml_file) as f:
+            actual = loader(f)
+        self.assertEqual(expected, actual)
+
+
+class TestDumpers(unittest.TestCase):
+
+    def test_json_dump_string(self):
+        data = {"hello": "world", "data": [1, 2, 3]}
+        expected = '{"hello": "world", "data": [1, 2, 3]}'
+        dumper = jsondiff.JsonDumper()
+        actual = dumper(data)
+        self.assertEqual(expected, actual)
+
+    def test_json_dump_string_indented(self):
+        data = {"hello": "world", "data": [1, 2, 3]}
+        expected = """{
+    "hello": "world",
+    "data": [
+        1,
+        2,
+        3
+    ]
+}"""
+        dumper = jsondiff.JsonDumper(indent=4)
+        actual = dumper(data)
+        self.assertEqual(expected, actual)
+
+    def test_json_dump_string_fp(self):
+        data = {"hello": "world", "data": [1, 2, 3]}
+        expected = """{
+    "hello": "world",
+    "data": [
+        1,
+        2,
+        3
+    ]
+}"""
+        dumper = jsondiff.JsonDumper(indent=4)
+        buffer = io.StringIO()
+        dumper(data, buffer)
+        self.assertEqual(expected, buffer.getvalue())
+
+    def test_yaml_dump_string(self):
+        data = {"hello": "world", "data": [1, 2, 3]}
+        expected = """data:
+- 1
+- 2
+- 3
+hello: world
+"""
+        # Sort keys just to make things deterministic
+        dumper = jsondiff.YamlDumper(sort_keys=True)
+        actual = dumper(data)
+        self.assertEqual(expected, actual)
+
+    def test_yaml_dump_string_fp(self):
+        data = {"hello": "world", "data": [1, 2, 3]}
+        expected = """data:
+- 1
+- 2
+- 3
+hello: world
+"""
+        # Sort keys just to make things deterministic
+        dumper = jsondiff.YamlDumper(sort_keys=True)
+        buffer = io.StringIO()
+        dumper(data, buffer)
+        self.assertEqual(expected, buffer.getvalue())