feat: implement additional config optimization api fields#196
Merged
andrewklatzke merged 3 commits intoJun 18, 2026
Merged
Conversation
Contributor
Author
|
Now additionally includes the content from here (fix): #197 |
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 509240f. Configure here.
jsonbailey
approved these changes
Jun 17, 2026
andrewklatzke
merged commit Jun 18, 2026
4ed30d8
into
aklatzke/AIC-2263/sdk-dx-improvements
5 checks passed
Merged
andrewklatzke
added a commit
that referenced
this pull request
Jul 17, 2026
🤖 I have created a release *beep* *boop* --- <details><summary>launchdarkly-server-sdk-ai: 1.2.0</summary> ## [1.2.0](launchdarkly-server-sdk-ai-1.1.0...launchdarkly-server-sdk-ai-1.2.0) (2026-07-17) ### Features * implements template/raw methods for fetching configs ([1c06d35](1c06d35)) * implements template/raw methods for fetching configs ([#204](#204)) ([5106866](5106866)) </details> <details><summary>ldai_optimizer: 0.2.0</summary> ## [0.2.0](ldai_optimizer-0.1.0...ldai_optimizer-0.2.0) (2026-07-17) ### ⚠ BREAKING CHANGES * Bump minimum LangChain version to 1.0.0 ### Features * ability to specify variation key on config or from_options for optimization package ([c5e39eb](c5e39eb)) * ability to specify variation key on config or from_options for optimization package ([#162](#162)) ([f0c4612](f0c4612)) * add auto-commit option ([4cb8859](4cb8859)) * add optimization for duration ([5d76276](5d76276)) * Add optimization package stub ([#109](#109)) ([ebd5166](ebd5166)) * add shared dataclass for calls so they can be handled by same handler ([31c8385](31c8385)) * add token limit handling ([e8c6692](e8c6692)) * adds ability to optimize for cost ([94de596](94de596)) * adds ability to optimize for cost ([#172](#172)) ([3b4baa3](3b4baa3)) * adds ability to use inverted judges ([0ed243e](0ed243e)) * adds ability to use inverted judges ([#168](#168)) ([23baeb4](23baeb4)) * Adds optimization package stub ([58b7731](58b7731)) * Adds optimization package stub ([cc85a05](cc85a05)) * adds reporting for cost and latency optimization failures ([365fa94](365fa94)) * adds reporting for cost and latency optimization failures ([#180](#180)) ([d267832](d267832)) * all logs -> debug ([2fd55e2](2fd55e2)) * Bump minimum LangChain version to 1.0.0 ([dc592c5](dc592c5)) * Drop support for python 3.9 ([#114](#114)) ([dc592c5](dc592c5)) * dx improvements for optimization package ([7074cfa](7074cfa)) * ground truth optimization path ([44c8c59](44c8c59)) * implement ability to use completions or agents for judge calls ([ea596a7](ea596a7)) * implement additional config optimization api fields ([#196](#196)) ([4ed30d8](4ed30d8)) * implement latency & token tracking for optimizations ([288336e](288336e)) * Implement optimization code paths and functionality for initial release ([#140](#140)) ([5204c47](5204c47)) * implementation of agent optimization + tests ([1712e4f](1712e4f)) * implements LD API client, optimize_from_config path ([2fecd54](2fecd54)) * implements optimize method in SDK, code moved ([9859d08](9859d08)) * partially implement optimize_from_config ([d3e1f96](d3e1f96)) * prevent overfitting via prompt changes and post-processing ([8f9f1e2](8f9f1e2)) * use judge_passed for all calcs ([a8f14de](a8f14de)) ### Bug Fixes * address cursor feedback ([f2f0894](f2f0894)) * adjust iteration logic so validation doesn't consume them ([3042984](3042984)) * better handling of params and custom params for optimization ([572a2aa](572a2aa)) * better handling of params and custom params for optimization ([#163](#163)) ([92f51fa](92f51fa)) * consistency with other makefiles ([b9a5601](b9a5601)) * don't only evaluate final input in GT results ([53f455f](53f455f)) * don't only evaluate final input in GT results ([9bedf9e](9bedf9e)) * ensure cost data is persisted ([4eb0bb0](4eb0bb0)) * fix for single-iteration optimizations ([96eadb7](96eadb7)) * fix for single-iteration optimizations ([#209](#209)) ([ba5aa63](ba5aa63)) * lint ([aee6aa7](aee6aa7)) * lint + missed variable rename ([f8e5509](f8e5509)) * lints + structured output tool rename ([8481690](8481690)) * move failure forward ([3c65a6e](3c65a6e)) * only strip known provider prefixes ([66bc1f0](66bc1f0)) * **optimization:** address Bugbot comments on PR [#140](#140) ([9e57d86](9e57d86)) * **optimization:** fix CI test failures and apply Bugbot-reported fixes ([ffad0cb](ffad0cb)) * **optimization:** match model config by key as well as id for pricing lookup ([a8415be](a8415be)) * **optimization:** remove dead persistence counter code ([c76cb16](c76cb16)) * **optimization:** retry LLM calls on transient provider errors (429/503/529) ([a131e4b](a131e4b)) * **optimization:** use run-chosen context for config judges; guard non-dict JSON ([64e12f2](64e12f2)) * pull model configs if available in options path ([4fc1ecf](4fc1ecf)) * remove unnecessary token path ([dc82818](dc82818)) * sort imports ([c032aaf](c032aaf)) * success path + add test, cursor feedback ([8f3468f](8f3468f)) </details> --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). <!-- CURSOR_SUMMARY --> --- > [!NOTE] > **Low Risk** > Mechanical version, changelog, and dependency-floor updates from Release Please; no runtime code changes in this PR. > > **Overview** > **Release Please** version bump only—no application logic changes in this diff. > > **`launchdarkly-server-sdk-ai` 1.2.0** — `pyproject.toml`, `ldai.__version__`, changelog, and provenance docs move from **1.1.0 → 1.2.0**. The new changelog section documents **template/raw config fetch APIs** (shipped in prior commits, e.g. `completion_config_template`). > > **`launchdarkly-ai-optimizer` 0.2.0** — same pattern (**0.1.0 → 0.2.0**) with an expanded **CHANGELOG** (including **LangChain ≥ 1.0** as a breaking note for optimizer consumers). > > **Provider packages** — `launchdarkly-server-sdk-ai-langchain` and `launchdarkly-server-sdk-ai-openai` now require **`launchdarkly-server-sdk-ai>=1.2.0`** instead of **≥1.1.0**. > > **`.release-please-manifest.json`** is updated to match the new package versions. > > <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit e86bad4. Bugbot is set up for automated code reviews on this repo. Configure [here](https://www.cursor.com/dashboard/bugbot).</sup> <!-- /CURSOR_SUMMARY -->
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.

Requirements
Describe the solution you've provided
We added first class support for some fields on the UI --
This PR pulls them into the SDK. token/latency optimization are using their previous paths for now in this PR. Rather than the regex approach we just use the flags from the API now. The latency/cost optimization paths will be updated in a subsequent PR.
Describe alternatives you've considered
The initial implementation of these two code paths for optimizations were kind of hacky to begin with (just using a dictionary to look up words that might mean they want to do it). This was the intended solution.
Note
Medium Risk
Changes when latency/cost gates and judge templates apply (explicit flags vs inferred text) and alters config-judge loading and auto-commit gating, which can shift optimization outcomes for existing runs.
Overview
Wires LaunchDarkly agent optimization API fields into the Python SDK:
latencyOptimization,tokenOptimization, andautoCommiton remote configs, plus optionalvariation_keyonOptimizationOptions/GroundTruthOptimizationOptionsto start from a specific AI config variation (REST fetch; requires API key andproject_key).Latency and token behavior no longer infer goals from acceptance-statement keyword regexes. Gates, judge prompt augmentations, variation prompts, and model-pricing warnings now key off
latency_optimizationandtoken_optimizationbooleans (from options or API). When unset/false, those paths stay off.Config judges resolve via raw flag
variation()and local{{key}}interpolation (includingmessage_history/response_to_evaluate) instead ofLDAIClient.judge_config. System-only judge templates get an auto-built user turn.optimize_from_configmaps the new API fields into built options; auto-commit runs only when both the fetched config’sautoCommitand caller options allow it. Tests drop regex helpers and cover the new flags, judge path, andvariation_keyvalidation.Reviewed by Cursor Bugbot for commit 509240f. Bugbot is set up for automated code reviews on this repo. Configure here.