feat: Introduce ManagedModel and ModelRunner (PR-3)#104
Merged
jsonbailey merged 8 commits intomainfrom Mar 26, 2026
Merged
Conversation
jsonbailey
changed the title
jsonbailey
force-pushed
the
jb/aic-1664/managed-model
branch
from
39bac26 to
3fc5d47
Compare
jsonbailey
force-pushed
the
jb/aic-1664/adapter-executor-factory
branch
from
4bef65c to
e10be52
Compare
feat: Add ModelRunner ABC with invoke_model() and invoke_structured_model() feat: Add ManagedModel replacing Chat; expose get_model_runner() escape hatch feat!: Rename ChatResponse to ModelResponse in providers/types.py feat!: Extract OpenAIModelRunner from OpenAIRunnerFactory; factory is now model-creation-only feat!: Extract LangChainModelRunner from LangChainRunnerFactory; factory is now model-creation-only feat: Add OpenAIHelper with shared utilities for model and future agent runners feat: Add LangChainHelper with shared utilities for model and future agent runners feat!: LangChainRunnerFactory is now a no-arg factory; static helpers moved to LangChainHelper fix: LDClient.create_chat() is deprecated in favour of create_model() fix: Chat alias in ldai.chat is deprecated in favour of ManagedModel fix: Rename ai_provider param to model_runner in Judge and ManagedModel
jsonbailey
force-pushed
the
jb/aic-1664/managed-model
branch
from
3fc5d47 to
fc3880c
Compare
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
packages/ai-providers/server-ai-langchain/src/ldai_langchain/langchain_model_runner.py
Outdated
Show resolved
Hide resolved
packages/ai-providers/server-ai-langchain/src/ldai_langchain/langchain_helper.py
Outdated
Show resolved
Hide resolved
packages/ai-providers/server-ai-langchain/src/ldai_langchain/langchain_helper.py
Outdated
Show resolved
Hide resolved
jsonbailey
changed the title
…heck, and Bedrock provider param
packages/ai-providers/server-ai-openai/src/ldai_openai/openai_runner_factory.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes 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.
packages/ai-providers/server-ai-openai/src/ldai_openai/openai_model_runner.py
Outdated
Show resolved
Hide resolved
…ve unused import - Restore mutable structured_response pattern from f951dac so raw_response and usage are preserved even when an exception occurs after partial extraction - Rename LangChainHelper.convert_messages -> convert_messages_to_langchain - Rename OpenAIHelper.convert_messages -> convert_messages_to_openai - Remove unused TokenUsage import from openai_model_runner.py - Fix runner_factory.create_model return type to Optional[ModelRunner] - Fix isort in langchain/openai runner and helper files - Fix isort in ldai/__init__.py and client.py Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
keelerm84
approved these changes
Mar 26, 2026
packages/ai-providers/server-ai-langchain/src/ldai_langchain/langchain_helper.py
Outdated
Show resolved
Hide resolved
packages/ai-providers/server-ai-openai/src/ldai_openai/openai_helper.py
Outdated
Show resolved
Hide resolved
…vel functions Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Merged
jsonbailey
added a commit
that referenced
this pull request
Apr 2, 2026
🤖 I have created a release *beep* *boop* --- <details><summary>launchdarkly-server-sdk-ai: 0.17.0</summary> ## [0.17.0](launchdarkly-server-sdk-ai-0.16.1...launchdarkly-server-sdk-ai-0.17.0) (2026-04-02) ### ⚠ BREAKING CHANGES * Bump minimum LangChain version to 1.0.0 * Split track_metrics_of into sync and async variants ([#112](#112)) * Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead * Extract shared utilities to openai_helper * Extract shared utilities to langchain_helper * Restructure provider factory and support additional create methods ([#102](#102)) ### Features * Add get_ai_usage_from_response to openai_helper ([4fab18f](4fab18f)) * Add get_tool_calls_from_response and sum_token_usage_from_messages to langchain_helper ([4fab18f](4fab18f)) * Add LangGraphAgentGraphRunner ([56ce0fd](56ce0fd)) * Add LDAIClient.create_agent() returning ManagedAgent ([53fd95e](53fd95e)) * Add ManagedAgentGraph support ([#111](#111)) ([56ce0fd](56ce0fd)) * Add ModelRunner ABC with invoke_model() and ([453c71c](453c71c)) * Add OpenAIAgentGraphRunner ([56ce0fd](56ce0fd)) * Add OpenAIAgentRunner with agentic tool-calling loop ([53fd95e](53fd95e)) * add optimization package stub ([872e81e](872e81e)) * Add optional graph_key to all LDAIConfigTracker track_* methods for graph correlation ([4fab18f](4fab18f)) * Add track_tool_call/track_tool_calls to LDAIConfigTracker ([4fab18f](4fab18f)) * Adds optimization package stub ([58b7731](58b7731)) * Bump minimum LangChain version to 1.0.0 ([dc592c5](dc592c5)) * Deprecated Chat object in favor of ManagedModel ([453c71c](453c71c)) * Deprecated create_chat(), use create_model() on the LDAIClient ([453c71c](453c71c)) * Drop support for python 3.9 ([#114](#114)) ([dc592c5](dc592c5)) * Extract shared utilities to langchain_helper ([453c71c](453c71c)) * Extract shared utilities to openai_helper ([453c71c](453c71c)) * Introduce ManagedAgent and AgentRunner implementations ([#110](#110)) ([53fd95e](53fd95e)) * Introduce ManagedModel and ModelRunner (PR-3) ([#104](#104)) ([453c71c](453c71c)) * Restructure provider factory and support additional create methods ([#102](#102)) ([e6e4907](e6e4907)) * Split track_metrics_of into sync and async variants ([#112](#112)) ([4fab18f](4fab18f)) ### Bug Fixes * make AIGraphTracker.track_total_tokens accept Optional[TokenUsage], skip when None or total <= 0 ([4fab18f](4fab18f)) * Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead ([4fab18f](4fab18f)) * use time.perf_counter_ns() for sub-millisecond precision in duration calculations ([4fab18f](4fab18f)) </details> <details><summary>launchdarkly-server-sdk-ai-langchain: 0.4.0</summary> ## [0.4.0](launchdarkly-server-sdk-ai-langchain-0.3.2...launchdarkly-server-sdk-ai-langchain-0.4.0) (2026-04-02) ### ⚠ BREAKING CHANGES * Bump minimum LangChain version to 1.0.0 * Split track_metrics_of into sync and async variants ([#112](#112)) * Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead * Extract shared utilities to openai_helper * Extract shared utilities to langchain_helper * Restructure provider factory and support additional create methods ([#102](#102)) ### Features * Add get_ai_usage_from_response to openai_helper ([4fab18f](4fab18f)) * Add get_tool_calls_from_response and sum_token_usage_from_messages to langchain_helper ([4fab18f](4fab18f)) * Add LangGraphAgentGraphRunner ([56ce0fd](56ce0fd)) * Add LDAIClient.create_agent() returning ManagedAgent ([53fd95e](53fd95e)) * Add ManagedAgentGraph support ([#111](#111)) ([56ce0fd](56ce0fd)) * Add ModelRunner ABC with invoke_model() and ([453c71c](453c71c)) * Add OpenAIAgentGraphRunner ([56ce0fd](56ce0fd)) * Add OpenAIAgentRunner with agentic tool-calling loop ([53fd95e](53fd95e)) * add optimization package stub ([872e81e](872e81e)) * Add optional graph_key to all LDAIConfigTracker track_* methods for graph correlation ([4fab18f](4fab18f)) * Add track_tool_call/track_tool_calls to LDAIConfigTracker ([4fab18f](4fab18f)) * Adds optimization package stub ([58b7731](58b7731)) * Bump minimum LangChain version to 1.0.0 ([dc592c5](dc592c5)) * Deprecated Chat object in favor of ManagedModel ([453c71c](453c71c)) * Deprecated create_chat(), use create_model() on the LDAIClient ([453c71c](453c71c)) * Drop support for python 3.9 ([#114](#114)) ([dc592c5](dc592c5)) * Extract shared utilities to langchain_helper ([453c71c](453c71c)) * Extract shared utilities to openai_helper ([453c71c](453c71c)) * Introduce ManagedAgent and AgentRunner implementations ([#110](#110)) ([53fd95e](53fd95e)) * Introduce ManagedModel and ModelRunner (PR-3) ([#104](#104)) ([453c71c](453c71c)) * Restructure provider factory and support additional create methods ([#102](#102)) ([e6e4907](e6e4907)) * Split track_metrics_of into sync and async variants ([#112](#112)) ([4fab18f](4fab18f)) ### Bug Fixes * make AIGraphTracker.track_total_tokens accept Optional[TokenUsage], skip when None or total <= 0 ([4fab18f](4fab18f)) * Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead ([4fab18f](4fab18f)) * use time.perf_counter_ns() for sub-millisecond precision in duration calculations ([4fab18f](4fab18f)) </details> <details><summary>launchdarkly-server-sdk-ai-openai: 0.3.0</summary> ## [0.3.0](launchdarkly-server-sdk-ai-openai-0.2.1...launchdarkly-server-sdk-ai-openai-0.3.0) (2026-04-02) ### ⚠ BREAKING CHANGES * Bump minimum LangChain version to 1.0.0 * Extract shared utilities to openai_helper * Extract shared utilities to langchain_helper * Restructure provider factory and support additional create methods ([#102](#102)) ### Features * Add LangGraphAgentGraphRunner ([56ce0fd](56ce0fd)) * Add LDAIClient.create_agent() returning ManagedAgent ([53fd95e](53fd95e)) * Add ManagedAgentGraph support ([#111](#111)) ([56ce0fd](56ce0fd)) * Add ModelRunner ABC with invoke_model() and ([453c71c](453c71c)) * Add OpenAIAgentGraphRunner ([56ce0fd](56ce0fd)) * Add OpenAIAgentRunner with agentic tool-calling loop ([53fd95e](53fd95e)) * add optimization package stub ([872e81e](872e81e)) * Adds optimization package stub ([58b7731](58b7731)) * Bump minimum LangChain version to 1.0.0 ([dc592c5](dc592c5)) * Deprecated Chat object in favor of ManagedModel ([453c71c](453c71c)) * Deprecated create_chat(), use create_model() on the LDAIClient ([453c71c](453c71c)) * Drop support for python 3.9 ([#114](#114)) ([dc592c5](dc592c5)) * Extract shared utilities to langchain_helper ([453c71c](453c71c)) * Extract shared utilities to openai_helper ([453c71c](453c71c)) * Introduce ManagedAgent and AgentRunner implementations ([#110](#110)) ([53fd95e](53fd95e)) * Introduce ManagedModel and ModelRunner (PR-3) ([#104](#104)) ([453c71c](453c71c)) * Restructure provider factory and support additional create methods ([#102](#102)) ([e6e4907](e6e4907)) </details> <details><summary>launchdarkly-server-sdk-ai-optimization: 0.1.0</summary> ## 0.1.0 (2026-04-02) ### ⚠ BREAKING CHANGES * Bump minimum LangChain version to 1.0.0 ### Features * Add optimization package stub ([#109](#109)) ([ebd5166](ebd5166)) * Adds optimization package stub ([58b7731](58b7731)) * Adds optimization package stub ([cc85a05](cc85a05)) * Bump minimum LangChain version to 1.0.0 ([dc592c5](dc592c5)) * Drop support for python 3.9 ([#114](#114)) ([dc592c5](dc592c5)) ### Bug Fixes * consistency with other makefiles ([b9a5601](b9a5601)) </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** > Low risk because this PR only updates release metadata (versions, manifests, provenance docs, and changelogs) without changing runtime code paths. > > **Overview** > Bumps package versions for the release: `launchdarkly-server-sdk-ai` to `0.17.0`, LangChain provider to `0.4.0`, OpenAI provider to `0.3.0`, and introduces/bumps `launchdarkly-server-sdk-ai-optimization` to `0.1.0` (via `.release-please-manifest.json` and each package’s `pyproject.toml`/`__init__.__version__`). > > Updates `CHANGELOG.md` entries and `PROVENANCE.md` version references to match the new release numbers. > > <sup>Written by [Cursor Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit ae3e183. This will update automatically on new commits. Configure [here](https://cursor.com/dashboard?tab=bugbot).</sup> <!-- /CURSOR_SUMMARY --> --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: jsonbailey <jbailey@launchdarkly.com>
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
2 participants
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.
feat: Add ModelRunner ABC with invoke_model() and invoke_structured_model()
feat: Add ManagedModel replacing Chat; expose get_model_runner() escape hatch
feat!: Rename ChatResponse to ModelResponse in providers/types.py
feat!: Extract OpenAIModelRunner from OpenAIRunnerFactory; factory is now model-creation-only
feat!: Extract LangChainModelRunner from LangChainRunnerFactory; factory is now model-creation-only
feat!: Extract shared utilities to openai_helper
feat!: Extract shared utilities to langchain_helper
feat: Deprecated create_chat(), use create_model() on the LDAIClient
Note
High Risk
High risk because it changes core SDK abstractions and public APIs (e.g.,
Chat/ChatResponse->ManagedModel/ModelResponse) and refactors provider invocation paths, which can break downstream integrations despite tests updates.Overview
Introduces a new runtime abstraction,
ModelRunner, and a new managed wrapper,ManagedModel, shifting the SDK’s model execution path from provider instances to configured runner objects.Renames
ChatResponsetoModelResponseacross the SDK, updatesRunnerFactory.create_model()to return aModelRunner, and updatesJudgeto invoke structured output through a runner instead of anAIProvider.Refactors the OpenAI and LangChain provider packages so their
*RunnerFactoryclasses are model-creation-only, extracting invocation logic intoOpenAIModelRunner/LangChainModelRunnerand shared helper modules; updates exports and tests accordingly. AddsLDAIClient.create_model()and deprecatescreate_chat()andChat(now a compatibility alias).Written by Cursor Bugbot for commit 9285ffd. This will update automatically on new commits. Configure here.