* Python: Exclude CMAB experiments from user profile updates and add related tests

* linting fix

* chore: trigger tests

* Python: Exclude CMAB experiments from user profile updates and update related tests

* Python: Exclude CMAB experiments from user profile storage in decision service

* Python: Remove test for CMAB experiments with IGNORE_USER_PROFILE_SERVICE option

* Add Ruff configuration file

* Update target version in Ruff configuration to Python 3.9

* Replace Flake8 with Ruff for linting in CI and update contribution guidelines

* Remove Flake8 configuration file as part of replacing Flake8 with Ruff

* ruff autofix

* testing ruff fail in ci

* Fix formatting in ruff.toml by ensuring newline at end of file

* Update .github/workflows/python.yml

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* ruff fix

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

…figuration (#497)

* fix: update PollingConfigManager initialization requirements and enhance tests for url handling

* fix: clarify initialization requirements for PollingConfigManager

* chore: remove outdated source clear GitHub Actions workflow

* chore: add persist-credentials option to checkout steps in workflows

* [AI-FSSDK] [FSSDK-12337] Add Feature Rollout support to project config parsing

* [AI-FSSDK] [FSSDK-12337] Fix test structure, mypy and ruff compliance

- Move feature rollout tests from standalone test_feature_rollout.py into
  test_config.py following module-level testing convention
- Use base.BaseTest instead of unittest.TestCase for consistency
- Fix mypy strict type errors in Variation construction using cast
- All checks pass: ruff, mypy --strict, pytest (941/941)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [AI-FSSDK] [FSSDK-12337] Simplify feature rollout config parsing

- Inline _get_everyone_else_variation logic, remove unnecessary static method
- Use get_rollout_from_id() to match TDD pseudocode
- Remove isinstance check (rollout experiments are always dicts)
- Remove 3 unit tests for deleted helper method (edge cases already
  covered by integration-level tests)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [AI-FSSDK] [FSSDK-12337] Restore _get_everyone_else_variation as method

- Restore as instance method matching TDD pseudocode structure
- Takes flag (FeatureFlag) param, calls get_rollout_from_id internally
- Caller simplified to: everyone_else_variation = self._get_everyone_else_variation(flag)

* [AI-FSSDK] [FSSDK-12337] Add ExperimentTypes constants for type field

- Add ExperimentTypes enum in helpers/enums.py with AB, MAB, CMAB, FEATURE_ROLLOUT
- Use ExperimentTypes.FEATURE_ROLLOUT constant in config parsing instead of raw string

* [AI-FSSDK] [FSSDK-12337] Type-restrict Experiment.type to ExperimentType

- Add ExperimentType Literal type in helpers/types.py: 'a/b', 'mab', 'cmab', 'feature_rollout'
- Change Experiment.type from Optional[str] to Optional[ExperimentType]

* [AI-FSSDK] [FSSDK-12337] Remove ExperimentTypes class, simplify type check

- Remove redundant ExperimentTypes class from enums.py (ExperimentType Literal suffices)
- Simplify getattr(experiment, 'type', None) to experiment.type

* [AI-FSSDK] [FSSDK-12337] Return Variation entity from _get_everyone_else_variation

- Build Variation entity once in helper, derive dict from it in caller
- Addresses PR review comment from jaeopt

* [AI-FSSDK] [FSSDK-12337] Remove redundant tests, keep 6 essential ones

Removed 7 tests that were covered by other tests:
- test_experiment_type_field_parsed (covered by injection test)
- test_feature_rollout_with_empty_rollout_experiments (similar to no_rollout)
- test_feature_rollout_multiple_experiments_mixed_types (covered by injection + unchanged)
- test_feature_rollout_flag_variations_map_includes_injected (subset of maps test)
- test_experiment_type_ab (just string assignment)
- test_feature_rollout_with_variables_on_everyone_else (edge case)
- test_existing_datafile_not_broken (covered by none_when_missing + unchanged)

* [AI-FSSDK] [FSSDK-12337] Add ExperimentTypes constant and targeted_delivery type

- Add ExperimentTypes class in enums.py with ab, mab, cmab, td, fr
- Add 'targeted_delivery' to ExperimentType Literal in types.py
- Use enums.ExperimentTypes.fr constant in injection check
- Add test for type field parsing from datafile

* [AI-FSSDK] [FSSDK-12337] Remove test not in ticket spec

Remove test_feature_rollout_everyone_else_is_last_rollout_rule
to match updated Jira ticket test requirements.

* [FSSDK-12337] Fix experiment type values to match backend

Update ExperimentTypes and ExperimentType Literal to use actual
backend values: 'multi_armed_bandit' and 'contextual_multi_armed_bandit'
instead of shorthand 'mab' and 'cmab'.

* Format ExperimentType definition for ruff check

* [AI-FSSDK] [FSSDK-12337] Update experiment type values to short-form abbreviations

* [AI-FSSDK] [FSSDK-12337] Add validation for experiment type field

* trigger CI

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

…g parsing (#500)

…ons (#505)

Move error-handling logic out of finally block into normal control flow
and change bare except to except Exception. Per Python docs, a return
in finally suppresses any in-flight exception, which silently swallows
errors raised inside except blocks.

Fixes #439

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [FSSDK-12368] Remove legacy flag-level holdout fields

Remove deprecated flag-level holdout functionality as cleanup after
local holdouts (FSSDK-12369) implementation.

Changes:
- Removed includedFlags and excludedFlags from Holdout entity
- Removed includedFlags and excludedFlags from HoldoutDict type
- Removed get_holdouts_for_flag() method from ProjectConfig
- Removed flag-level holdout infrastructure:
  - global_holdouts list
  - included_holdouts dict
  - excluded_holdouts dict
  - flag_holdouts_map dict
  - Flag-level population logic in ProjectConfig.__init__
- Removed flag-level holdout checking from DecisionService.get_variation_for_feature()
- Removed 10 test methods testing removed functionality
- Removed includedFlags/excludedFlags from all test fixtures

Impact: 7 files modified, 310 lines deleted, 5 lines added (net: -305 lines)

Verification:
- grep for "includedFlags", "excludedFlags", "getHoldoutsForFlag" returns zero results
- grep for "global_holdouts", "included_holdouts", "excluded_holdouts", "flag_holdouts_map" returns zero results

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

* [FSSDK-12368] Restore holdout evaluation in decide flow, remove only flag-level include/exclude fields

The previous cleanup removed the entire holdout evaluation from the decide
flow. This restores holdout checking while keeping the removal of legacy
includedFlags/excludedFlags fields — all running holdouts now apply to all
flags uniformly.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
Co-authored-by: Farhan Anjum <Farhan.Anjum@optimizely.com>

* [FSSDK-12521] chore: preparing for release v5.5.0

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* chore: remove internal tooling entries from changelog

Removed ruff linter migration and Arnica code risk warning entries
from the v5.5.0 changelog as they are internal CI/tooling changes
not relevant to SDK users.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Update CHANGELOG.md

Co-authored-by: Muzahidul Islam <129880873+muzahidul-opti@users.noreply.github.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Muzahidul Islam <129880873+muzahidul-opti@users.noreply.github.com>

* [AI-FSSDK] [FSSDK-12369] Add local holdouts support to Python SDK

* [FSSDK-12369] Fix copyright year from 2025 to 2026 in new test files

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* [FSSDK-12369] Add mandatory forced-decision-beats-local-holdout enforcement test

* trigger CI

* trigger CI

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Farhan Anjum <Farhan.Anjum@optimizely.com>

…er (#513)

* [AI-FSSDK] [FSSDK-12670] Block ODP identify event for single identifier

* [FSSDK-12670] Address review feedback: fix log message

* [FSSDK-12670] Retrigger CI

* [FSSDK-12670] Add comment explaining the < 2 identifiers guard

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* [FSSDK-12670] retrigger CI

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Farhan Anjum <Farhan.Anjum@optimizely.com>