feat(api): update API spec from langfuse/langfuse 9c46a07

[langfuse-bot]

Greptile Summary

Updates the generated API client from the latest Langfuse API specification.

• Adds session_id filtering to synchronous and asynchronous observation-listing clients.
• Documents semantic-root filtering for metrics and evaluation rules.
• Refreshes blob-storage export, comment-author, SCIM password, and scores endpoint documentation.

Confidence Score: 5/5

The PR appears safe to merge because the behavioral addition is consistently propagated and the remaining changes affect generated documentation only.

The new observation session filter is keyword-only and maps directly to sessionId in both synchronous and asynchronous request paths; no concrete blocking failure remains.

_{Reviews (1): Last reviewed commit: "feat(api): update API spec from langfuse..." | Re-trigger Greptile}

Context used:

• Knowledge Base — API Client (langfuse/api/)

[github-actions]

@claude review

[claude]

LGTM — this is an auto-generated Fern API client update mirroring changes from langfuse/langfuse; no bugs found.

What was reviewed: docstring-only clarifications (blob storage export source, comment author, SCIM password field, scores deprecation removal), the new isRootObservation metrics dimension example (docs-only), and the new session_id passthrough parameter on observations.get_many — confirmed it's threaded consistently through client.py and raw_client.py and mapped to the sessionId query param exactly like the existing user_id parameter.

Extended reasoning...

Overview

This PR is an auto-generated update of the Langfuse API client (via Fern) to match the latest langfuse/langfuse OpenAPI spec (commit 9c46a07). All 16 changed files are generated client/raw_client/types modules. The substantive changes are: (1) docstring/comment wording updates across blob storage integrations, comments, SCIM, and scores endpoints — clarifying default export-source behavior, comment author membership requirements, SCIM password field being ignored, and removal of a stale deprecation notice on the v2 scores endpoints; (2) documentation additions for a new isRootObservation boolean dimension/filter on the metrics and evaluation-rule-filter types (the underlying EvaluationRuleFilter_Boolean type already existed in code — only descriptive text was added); (3) one functional addition: a new optional session_id parameter on observations.get_many (sync and async, client and raw_client), which is threaded through identically to the existing user_id parameter and serialized to the sessionId query param.

Security risks

None. No auth, crypto, or permission logic is touched. The only new parameter (session_id) is a plain optional string forwarded verbatim as a query parameter, consistent with how all other filter parameters are already handled by the generated HTTP client.

Level of scrutiny

Low. This is a mechanically generated SDK sync PR with no hand-written business logic — the diff is dominated by docstring text and one additive, pattern-following parameter. This class of change is inherently low-risk since it follows the exact same generation pattern as dozens of existing parameters in the same files.

Other factors

The new parameter follows the established code pattern exactly (declaration, docstring, pass-through in the public client, and query-param mapping in raw_client) with no divergence. No existing tests needed updating since this is additive and mirrors existing filter parameters; the bug-hunting pass also found nothing.