new: [internal] Add support for orjson

orjson is much faster library for decoding and encoding JSON formats

.github/workflows/pytest.yml

@@ -36,5 +36,11 @@ jobs:
         poetry run pytest --cov=pymisp tests/test_*.py
         poetry run mypy tests/testlive_comprehensive.py tests/test_mispevent.py tests/testlive_sync.py pymisp
 
+    - name: Test with nosetests with orjson
+      run: |
+        pip3 install orjson
+        poetry run pytest --cov=pymisp tests/test_*.py
+        poetry run mypy tests/testlive_comprehensive.py tests/test_mispevent.py tests/testlive_sync.py pymisp
+
     - name: Upload coverage to Codecov
       uses: codecov/codecov-action@v3

pymisp/abstract.py

@@ -1,41 +1,33 @@
 #!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-
+import logging
 from datetime import date, datetime
-
 from deprecated import deprecated  # type: ignore
 from json import JSONEncoder
 from uuid import UUID
 from abc import ABCMeta
-
-try:
-    from rapidjson import load  # type: ignore
-    from rapidjson import loads  # type: ignore
-    from rapidjson import dumps  # type: ignore
-    HAS_RAPIDJSON = True
-except ImportError:
-    from json import load
-    from json import loads
-    from json import dumps
-    HAS_RAPIDJSON = False
-
-import logging
 from enum import Enum
 from typing import Union, Optional, Any, Dict, List, Set, Mapping
-
-from .exceptions import PyMISPInvalidFormat, PyMISPError
-
-
 from collections.abc import MutableMapping
 from functools import lru_cache
 from pathlib import Path
 
+try:
+    import orjson  # type: ignore
+    from orjson import loads, dumps  # type: ignore
+    HAS_ORJSON = True
+except ImportError:
+    from json import loads, dumps
+    HAS_ORJSON = False
+
+from .exceptions import PyMISPInvalidFormat, PyMISPError
+
 logger = logging.getLogger('pymisp')
 
+
 resources_path = Path(__file__).parent / 'data'
 misp_objects_path = resources_path / 'misp-objects' / 'objects'
-with (resources_path / 'describeTypes.json').open('r') as f:
-    describe_types = load(f)['result']
+with (resources_path / 'describeTypes.json').open('rb') as f:
+    describe_types = loads(f.read())['result']
 
 
 class MISPFileCache(object):
@@ -43,11 +35,11 @@ class MISPFileCache(object):
 
     @staticmethod
     @lru_cache(maxsize=150)
-    def _load_json(path: Path) -> Union[dict, None]:
+    def _load_json(path: Path) -> Optional[dict]:
         if not path.exists():
             return None
-        with path.open('r', encoding='utf-8') as f:
-            data = load(f)
+        with path.open('rb') as f:
+            data = loads(f.read())
         return data
 
 
@@ -249,6 +241,15 @@ class AbstractMISP(MutableMapping, MISPFileCache, metaclass=ABCMeta):
 
     def to_json(self, sort_keys: bool = False, indent: Optional[int] = None) -> str:
         """Dump recursively any class of type MISPAbstract to a json string"""
+        if HAS_ORJSON:
+            option = 0
+            if sort_keys:
+                option |= orjson.OPT_SORT_KEYS
+            if indent:
+                option |= orjson.OPT_INDENT_2
+
+            return dumps(self, default=pymisp_json_default, option=option).decode("utf-8")
+
         return dumps(self, default=pymisp_json_default, sort_keys=sort_keys, indent=indent)
 
     def __getitem__(self, key):
@@ -406,23 +407,13 @@ class MISPTag(AbstractMISP):
         return '<{self.__class__.__name__}(NotInitialized)>'.format(self=self)
 
 
-if HAS_RAPIDJSON:
-    def pymisp_json_default(obj: Union[AbstractMISP, datetime, date, Enum, UUID]) -> Union[Dict, str]:
-        if isinstance(obj, AbstractMISP):
-            return obj.jsonable()
-        elif isinstance(obj, (datetime, date)):
-            return obj.isoformat()
-        elif isinstance(obj, Enum):
-            return obj.value
-        elif isinstance(obj, UUID):
-            return str(obj)
-else:
-    def pymisp_json_default(obj: Union[AbstractMISP, datetime, date, Enum, UUID]) -> Union[Dict, str]:
-        if isinstance(obj, AbstractMISP):
-            return obj.jsonable()
-        elif isinstance(obj, (datetime, date)):
-            return obj.isoformat()
-        elif isinstance(obj, Enum):
-            return obj.value
-        elif isinstance(obj, UUID):
-            return str(obj)
+# UUID, datetime, date and Enum is serialized by ORJSON by default
+def pymisp_json_default(obj: Union[AbstractMISP, datetime, date, Enum, UUID]) -> Union[Dict, str]:
+    if isinstance(obj, AbstractMISP):
+        return obj.jsonable()
+    elif isinstance(obj, (datetime, date)):
+        return obj.isoformat()
+    elif isinstance(obj, Enum):
+        return obj.value
+    elif isinstance(obj, UUID):
+        return str(obj)

pymisp/api.py

@@ -1,13 +1,9 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-
 from typing import TypeVar, Optional, Tuple, List, Dict, Union, Any, Mapping, Iterable, MutableMapping
 from datetime import date, datetime
 import csv
 from pathlib import Path
 import logging
 from urllib.parse import urljoin
-import json
 import requests
 from requests.auth import AuthBase
 import re
@@ -18,6 +14,14 @@ import copy
 import urllib3  # type: ignore
 from io import BytesIO, StringIO
 
+try:
+    # orjson is optional dependency that speedups parsing and encoding JSON
+    from orjson import loads, dumps  # type: ignore
+    HAS_ORJSON = True
+except ImportError:
+    from json import loads, dumps
+    HAS_ORJSON = False
+
 from . import __version__, everything_broken
 from .exceptions import MISPServerError, PyMISPUnexpectedResponse, PyMISPError, NoURL, NoKey
 from .mispevent import MISPEvent, MISPAttribute, MISPSighting, MISPLog, MISPObject, \
@@ -297,7 +301,7 @@ class PyMISP:
         """Get the most recent version from github"""
         r = requests.get('https://raw.githubusercontent.com/MISP/MISP/2.4/VERSION.json')
         if r.status_code == 200:
-            master_version = json.loads(r.text)
+            master_version = loads(r.content)
             return {'version': '{}.{}.{}'.format(master_version['major'], master_version['minor'], master_version['hotfix'])}
         return {'error': 'Impossible to retrieve the version of the master branch.'}
 
@@ -3345,7 +3349,7 @@ class PyMISP:
         """
         query: Dict[str, Any] = {'setting': user_setting}
         if isinstance(value, dict):
-            value = json.dumps(value)
+            value = dumps(value).decode("utf-8") if HAS_ORJSON else dumps(value)
         query['value'] = value
         if user:
             query['user_id'] = get_uuid_or_id_from_abstract_misp(user)
@@ -3682,7 +3686,7 @@ class PyMISP:
         if 400 <= response.status_code < 500:
             # The server returns a json message with the error details
             try:
-                error_message = response.json()
+                error_message = loads(response.content)
             except Exception:
                 raise MISPServerError(f'Error code {response.status_code}:\n{response.text}')
 
@@ -3692,7 +3696,7 @@ class PyMISP:
         # At this point, we had no error.
 
         try:
-            response_json = response.json()
+            response_json = loads(response.content)
             logger.debug(response_json)
             if isinstance(response_json, dict) and response_json.get('response') is not None:
                 # Cleanup.
@@ -3721,7 +3725,7 @@ class PyMISP:
         if url[0] == '/':
             # strip it: it will fail if MISP is in a sub directory
             url = url[1:]
-        # Cake PHP being an idiot, it doesn't accepts %20 (space) in the URL path,
+        # Cake PHP being an idiot, it doesn't accept %20 (space) in the URL path,
         # so we need to make it a + instead and hope for the best
         url = url.replace(' ', '+')
         url = urljoin(self.root_url, url)
@@ -3733,7 +3737,7 @@ class PyMISP:
                 if isinstance(data, dict):
                     # Remove None values.
                     data = {k: v for k, v in data.items() if v is not None}
-                d = json.dumps(data, default=pymisp_json_default)
+                d = dumps(data, default=pymisp_json_default)
 
         logger.debug(f'{request_type} - {url}')
         if d is not None:

pymisp/mispevent.py

@@ -1,8 +1,5 @@
-# -*- coding: utf-8 -*-
-
 from datetime import timezone, datetime, date
 import copy
-import json
 import os
 import base64
 import sys
@@ -17,6 +14,12 @@ from pathlib import Path
 from typing import List, Optional, Union, IO, Dict, Any
 import warnings
 
+try:
+    # orjson is optional dependency that speedups parsing and encoding JSON
+    import orjson as json  # type: ignore
+except ImportError:
+    import json
+
 from .abstract import AbstractMISP, MISPTag
 from .exceptions import (UnknownMISPObjectTemplate, InvalidMISPGalaxy, InvalidMISPObject,
                          PyMISPError, NewEventError, NewAttributeError, NewEventReportError,
@@ -1090,7 +1093,7 @@ class MISPObject(AbstractMISP):
             self._validate()
         return super(MISPObject, self).to_dict(json_format)
 
-    def to_json(self, sort_keys: bool = False, indent: Optional[int] = None, strict: bool = False):
+    def to_json(self, sort_keys: bool = False, indent: Optional[int] = None, strict: bool = False) -> str:
         if strict or self._strict and self._known_template:
             self._validate()
         return super(MISPObject, self).to_json(sort_keys=sort_keys, indent=indent)
@@ -1760,7 +1763,7 @@ class MISPEvent(AbstractMISP):
             event.pop('Object', None)
         self.from_dict(**event)
         if validate:
-            warnings.warn('''The validate parameter is deprecated because PyMISP is more flexible at loading event than the schema''')
+            warnings.warn('The validate parameter is deprecated because PyMISP is more flexible at loading event than the schema')
 
     def __setattr__(self, name, value):
         if name in ['date']: