v3.5.1

What's Changed

• feat: allow objects for Categorical Attributes by @typotter in #55
• chore: move api to baseUrl by @typotter in #56

Full Changelog: v3.4.0...v3.5.1

v3.4.0

What's Changed

• chore: Update restsharp package by @typotter in #51
• feat: additional GetBanditAction method by @typotter in #52
• version bump by @typotter in #53

Full Changelog: v3.3.3...v3.4.0

v3.3.3

What's Changed

• fix/ff 3389 by @sameerank in #48
• version bump to 3.3.3 by @sameerank in #49

Full Changelog: v3.3.2...v3.3.3

v3.3.2

What's Changed

• Quality: Parameterize SDK and test-data branches in testing workflow by @typotter in #44
• fix: catch exception on polling by @typotter in #46
• chore:version bump by @typotter in #47

Full Changelog: v3.3.1...v3.3.2

v3.3.1

What's Changed

• feat: Only load bandits if they are referenced by flags by @typotter in #29
• store, send ETAG and only apply config if modified by @typotter in #30
• chore: reference README in CS project to be snarfed by Nuget by @typotter in #32
• chore: pack README in nuget archive by @typotter in #33
• Fix: hardcode better SDK name and trim version to valid SemVer format by @typotter in #37
• chore:Refactor configuration classes by @typotter in #35
• Use BanditReferences in the UFC response by @typotter in #39
• fix: tie-break top score by sorting by action name by @typotter in #41
• feat: Only fetch bandits if loaded models are insufficient for the references by @typotter in #40
• feat: Client Mode for on-demand configuration loading by @typotter in #42
• version bump to 3.3.1 by @typotter in #43

Full Changelog: v3.2.1...v3.3.1

v3.2.1

What's Changed

• fix: more precise bandit evaluation test by @typotter in #27
• feat: update EppoClient to handle empty actions by @typotter in #28
• version bump by @typotter in #31

Full Changelog: v3.2.0...v3.2.1

v3.2

What's Changed

• Accept non-numbers for numeric attributes but filter them and warn. Also Remove snake_case json attributes by @typotter in #22
• Allow simple list of action strings without context attributes by @typotter in #25

Full Changelog: v3.1.0...v3.2.0

v3.1

What's Changed

• chore: More robust handling of universal test cases by @typotter in #20
• bugfix: Clear caches on new fetch to ensure stale flags/bandits do not persist by @typotter in #19
• bugfix: fixed typo on BanditLogEvent.FlagKey :( by @typotter in #24

Full Changelog: v3.0...v3.1.0

v3.0

What's Changed

Summary

Version 3 uses a new format for flag configurations, the Universal Flag Configuration (UFC) and provides support for Multi-armed Contextual Bandits.

Breaking Changes

• Included in the change to UFC is the requirement of strongly-typed variations. Thus all Get<Type>Assignment() calls now require a non-null default value to return if nothing is assigned or a (handled) error is encountered.
• The flagKey argument now comes first in all Get<Type>Assignment() method signatures.
  *IAssignmentLogger interface has changed, Log event data objects are more portable for data warehouses like Segment

Changes

• Implement IS_NULL operator and INTEGER value type by @typotter in #8
• Implement JSON assignment type (FF-999) by @typotter in #10
• Implement UFC (FF-2235) by @typotter in #11
• ufc readme update by @vpai in #7
• NOT_MATCHES operator, more toString edge cases (FF-2380) by @typotter in #12
• Add back GetJsonStringAssignment method. by @typotter in #13
• DTOs and parsing tests for Bandits by @typotter in #14
• Loads bandit models and flag configuration by @typotter in #15
• Read, store, fetch and lookup bandit flags by @typotter in #18
• Bandit Evaluation (FF-2354) by @typotter in #17
• Version bump and README updates for Bandits. by @typotter in #21

New Contributors

• @typotter made their first contribution in #8
• @vpai made their first contribution in #7

Full Changelog: v2.1.0...v3.0

v2.1.0

What's Changed

• add feature_flag and allocation to assignment log event; construct experiment from a compound by @leoromanovsky in #4
• Make feature_flag event attribute camelCase by @sameerank in #5
• add support for semver targeting rule evaluation by @leoromanovsky in #6

New Contributors

• @sameerank made their first contribution in #5

Full Changelog: v2.0.0...v2.1.0