From be21ee2fa20e0eb622103c5f15b13fef4e53d9e0 Mon Sep 17 00:00:00 2001
From: Bilal Al-Shahwany <bilal@split.io>
Date: Mon, 28 Aug 2023 13:38:39 -0700
Subject: [PATCH] added flag set validations to config

---
 splitio/client/config.py    | 37 ++++++++++++++++++++++++++++++++++---
 tests/client/test_config.py | 21 +++++++++++++++++++++
 2 files changed, 55 insertions(+), 3 deletions(-)

diff --git a/splitio/client/config.py b/splitio/client/config.py
index 4531e40a..cd171319 100644
--- a/splitio/client/config.py
+++ b/splitio/client/config.py
@@ -1,13 +1,14 @@
 """Default settings for the Split.IO SDK Python client."""
 import os.path
 import logging
+import re
 
 from splitio.engine.impressions import ImpressionsMode
 
 
 _LOGGER = logging.getLogger(__name__)
 DEFAULT_DATA_SAMPLING = 1
-
+_FLAG_SETS_REGEX = '^[a-z0-9][_a-z0-9]{0,49}$'
 
 DEFAULT_CONFIG = {
     'operationMode': 'standalone',
@@ -58,10 +59,10 @@
     'dataSampling': DEFAULT_DATA_SAMPLING,
     'storageWrapper': None,
     'storagePrefix': None,
-    'storageType': None
+    'storageType': None,
+    'FlagSets': None
 }
 
-
 def _parse_operation_mode(sdk_key, config):
     """
     Process incoming config to determine operation mode and storage type
@@ -119,6 +120,34 @@ def _sanitize_impressions_mode(storage_type, mode, refresh_rate=None):
     return mode, refresh_rate
 
 
+def _sanitize_flag_sets(flag_sets):
+    """
+    Check supplied flag sets list
+
+    :param flag_set: list of flag sets
+    :type flag_set: list[str]
+
+    :returns: Sanitized and sorted flag sets
+    :rtype: list[str]
+    """
+    sanitized_flag_sets = set()
+    for flag_set in flag_sets:
+        if flag_set != flag_set.strip():
+            _LOGGER.warning("SDK config: Flag Set name %s has extra whitespace, trimming" % (flag_set))
+            flag_set = flag_set.strip()
+
+        if flag_set != flag_set.lower():
+            _LOGGER.warning("SDK config: Flag Set name %s should be all lowercase - converting string to lowercase" % (flag_set))
+            flag_set = flag_set.lower()
+
+        if re.search(_FLAG_SETS_REGEX, flag_set) is None or re.search(_FLAG_SETS_REGEX, flag_set).group() != flag_set:
+            _LOGGER.warning("SDK config: you passed %s, Flag Set must adhere to the regular expressions %s. This means a Flag Set must start with a letter, be in lowercase, alphanumeric and have a max length of 50 characteres. %s was discarded.", flag_set, _FLAG_SETS_REGEX, flag_set)
+            continue
+
+        sanitized_flag_sets.add(flag_set.strip())
+
+    return sorted(list(sanitized_flag_sets))
+
 def sanitize(sdk_key, config):
     """
     Look for inconsistencies or ill-formed configs and tune it accordingly.
@@ -143,4 +172,6 @@ def sanitize(sdk_key, config):
         _LOGGER.warning('metricRefreshRate parameter minimum value is 60 seconds, defaulting to 3600 seconds.')
         processed['metricsRefreshRate'] = 3600
 
+    processed['FlagSets'] = _sanitize_flag_sets(processed['FlagSets']) if processed['FlagSets'] is not None else None
+
     return processed
diff --git a/tests/client/test_config.py b/tests/client/test_config.py
index 0d96b478..e9a1c284 100644
--- a/tests/client/test_config.py
+++ b/tests/client/test_config.py
@@ -1,5 +1,6 @@
 """Configuration unit tests."""
 # pylint: disable=protected-access,no-self-use,line-too-long
+import pytest
 
 from splitio.client import config
 from splitio.engine.impressions.impressions import ImpressionsMode
@@ -68,3 +69,23 @@ def test_sanitize(self):
         processed = config.sanitize('some', configs)
 
         assert processed['redisLocalCacheEnabled']  # check default is True
+
+    def test_sanitize_flag_sets(self):
+        """Test sanitization for flag sets."""
+        flag_sets = config._sanitize_flag_sets([' set1', 'set2 ', 'set3'])
+        assert flag_sets == ['set1', 'set2', 'set3']
+
+        flag_sets = config._sanitize_flag_sets(['1set', '_set2'])
+        assert flag_sets == ['1set']
+
+        flag_sets = config._sanitize_flag_sets(['Set1', 'SET2'])
+        assert flag_sets == ['set1', 'set2']
+
+        flag_sets = config._sanitize_flag_sets(['se\t1', 's/et2', 's*et3', 's!et4', 'se@t5', 'se#t5', 'se$t5', 'se^t5', 'se%t5', 'se&t5'])
+        assert flag_sets == []
+
+        flag_sets = config._sanitize_flag_sets(['set4', 'set1', 'set3', 'set1'])
+        assert flag_sets == ['set1', 'set3', 'set4']
+
+        flag_sets = config._sanitize_flag_sets(['w' * 50, 's' * 51])
+        assert flag_sets == ['w' * 50]