feat(api): update API spec from langfuse/langfuse 632aafa

[langfuse-bot]

Greptile Summary

This PR updates the docstring description of the matches operator across the observations API client files to reflect a behavior change in how the operator works — from "token-based full-text search" to "indexed literal search with token-boundary pruning."

• The description of input and output filter fields is updated from "accelerated token search" to "accelerated indexed literal search" in all four client classes (ObservationsClient, AsyncObservationsClient, RawObservationsClient, AsyncRawObservationsClient).
• The operator's prose explanation is expanded to clarify that % and _ are treated as literal characters (not SQL LIKE wildcards) and that contains is now described as offering "legacy substring semantics."

Confidence Score: 5/5

This PR is safe to merge — all changes are confined to docstrings and introduce no runtime behavior modifications.

All four changed hunks are docstring updates describing how the matches operator works. No function signatures, logic, or data flow is altered. The updated text is consistent across both the sync and async variants of the two client modules.

No files require special attention.

Sequence Diagram

sequenceDiagram
    participant Caller
    participant ObservationsClient
    participant API as Langfuse API

    Caller->>ObservationsClient: "get_many(filter=[{column:"input", operator:"matches", value:"..."}])"
    Note over ObservationsClient: Validates filter params<br/>(v2: must pair matches/= with input/output)
    ObservationsClient->>API: "GET /api/public/observations?filter=..."
    API-->>ObservationsClient: ObservationsViews (paginated)
    ObservationsClient-->>Caller: ObservationsViews

Loading

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

[github-actions]

@claude review

[claude]

LGTM — docstring-only update in auto-generated Fern API client.

Extended reasoning...

Overview

This PR updates docstrings in two auto-generated Fern API client files (langfuse/api/observations/client.py and langfuse/api/observations/raw_client.py). The changes only modify the description of the matches filter operator — from "accelerated token search / token-based full-text search" to "accelerated indexed literal search with token-boundary pruning", and add a clarifying note that % and _ are treated as literal characters (unlike SQL LIKE).

Security risks

None. No executable code paths, parameters, request bodies, or auth logic are touched. The diff is confined to docstring text inside generated client methods.

Level of scrutiny

Minimal. This is a routine API-spec regeneration (feat(api): update API spec from langfuse/langfuse 632aafa) that touches only auto-generated documentation strings. The repo already has prior commits of this exact form on main.

Other factors

No bugs were flagged by the bug hunting system. The four updated docstrings (sync + async × client + raw_client) are consistent with each other. Nothing here warrants human review.