fix(scores): support CORRECTION score data type

[DavidTraina]

What does this PR do?

The score methods' type hints omitted the CORRECTION score data type, so the example in the Corrections docs — langfuse.create_score(..., value="...", data_type="CORRECTION") — failed static type checking (mypy/pyright report [call-overload]), even though the runtime already accepted it.

This adds CORRECTION to:

• the public ScoreDataType alias (langfuse/types.py);
• the string-value @overloads and internal casts of create_score, score_current_span, and score_current_trace (langfuse/_client/client.py);
• the string-value overloads and casts of LangfuseObservationWrapper.score / score_trace (langfuse/_client/span.py);

and updates the corresponding data_type: / value: docstrings. CORRECTION is string-valued, so it is added only to the string overloads; the numeric overloads continue to reject it. There is no runtime behavior change — data_type already passes straight through to ScoreBody, whose generated ScoreDataType enum already includes CORRECTION.

Fixes: no tracking issue — corrects the documented example linked above.

Type of change

• Bug fix
• New feature
• Breaking change
• Refactor
• Documentation update
• Tooling, CI, or repo maintenance

Verification

List the main commands you ran:

uv run --frozen mypy langfuse --no-error-summary
uv run --frozen ruff check .
uv run --frozen ruff format --check langfuse/types.py langfuse/_client/client.py langfuse/_client/span.py
uv run --frozen pytest -n auto --dist worksteal tests/unit

Checklist

• I self-reviewed the diff using code_review.md.
• I added or updated tests for behavior changes — N/A: type-hint-only change with no runtime behavior change. A runtime test would pass identically before and after, and a get_args(ScoreDataType) membership test would be a change-detector. The type checker is the effective verification: it now accepts the documented create_score(..., data_type="CORRECTION") example and still rejects numeric CORRECTION.
• I updated docs, examples, or .env.template if needed — updated the inline data_type: / value: docstrings.
• I did not hand-edit generated files; if generated files changed, I used the upstream regeneration path — langfuse/api/ untouched.
• I did not commit secrets or credentials.

🤖 Generated with Claude Code

Greptile Summary

This PR adds CORRECTION to the public ScoreDataType alias and to all string-value @overloads (and their internal casts) of create_score, score_current_span, score_current_trace, LangfuseObservationWrapper.score, and score_trace. The generated API enum in langfuse/api/commons/types/score_data_type.py already included CORRECTION, so there is no runtime behaviour change.

• langfuse/types.py: ScoreDataType Literal type alias now includes "CORRECTION", bringing it in sync with the generated ScoreDataType StrEnum.
• langfuse/_client/client.py and langfuse/_client/span.py: All string-value overloads and cast expressions are updated consistently; numeric overloads are untouched; docstrings are refreshed to list the new type.

Confidence Score: 5/5

Safe to merge — this is a type-hint-only change that brings public aliases in line with the already-shipped generated API enum; no runtime logic is altered.

All three changed files make consistent, surgical additions of CORRECTION to string-value overloads and casts. The underlying API enum already carried the value, so there is no surface for new runtime breakage. The single observation is a documentation comment that could be more precise.

No files require special attention; changes are localised to type annotations and docstrings.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Caller passes value and data_type] --> B{value type?}
    B -- float --> C["Overload: NUMERIC or BOOLEAN"]
    B -- str --> D["Overload: CATEGORICAL, TEXT, or CORRECTION"]
    C --> E[create_score / score / score_trace]
    D --> E
    E --> F[cast data_type to internal Literal]
    F --> G[ScoreBody to Langfuse API]

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
    A[Caller passes value and data_type] --> B{value type?}
    B -- float --> C["Overload: NUMERIC or BOOLEAN"]
    B -- str --> D["Overload: CATEGORICAL, TEXT, or CORRECTION"]
    C --> E[create_score / score / score_trace]
    D --> E
    E --> F[cast data_type to internal Literal]
    F --> G[ScoreBody to Langfuse API]

Loading

Prompt To Fix All With AI

Fix the following 1 code review issue. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 1
langfuse/types.py:40-41
The comment on this line only mentions "Text scores" but `CORRECTION` is also a string-valued type that is intentionally absent from `ExperimentScoreType`. Updating the comment to cover both keeps future readers from wondering whether the omission of `CORRECTION` from `ExperimentScoreType` was accidental.

```suggestion
# Text and Correction scores are not supported for evals and experiments
ExperimentScoreType = Literal["NUMERIC", "CATEGORICAL", "BOOLEAN"]
```

_{Reviews (1): Last reviewed commit: "fix(scores): support CORRECTION score da..." | Re-trigger Greptile}

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[CLAassistant]

All committers have signed the CLA.

[hassiebp]

Thanks for your contribution, @DavidTraina !