fix(media): normalize urlsafe base64 for media uploads

[MJRuskin]

What does this PR do?

langfuse/langfuse#15179

Fixes compatibility issues with openinference package, which uses urlsafe base64 in certain output fields.
Normalizes the base64 string during the parse function.

Type of change

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

Verification

List the main commands you ran:

Checklist

• I self-reviewed the diff using code_review.md.
• I added or updated tests for behavior changes.
• I updated docs, examples, or .env.template if needed.
• I did not hand-edit generated files; if generated files changed, I used the upstream regeneration path.
• I did not commit secrets or credentials.

Greptile Summary

This PR fixes a compatibility bug where media uploads would fail when the base64 payload used the URL-safe alphabet (- and _ instead of + and /), as produced by the openinference package. The fix normalizes the incoming base64 string before decoding.

• Normalization logic: actual_data.replace("-", "+").replace("_", "/") is applied before base64.b64decode, converting URL-safe base64 to standard base64. This is functionally correct; an equally valid (and more idiomatic) alternative is base64.urlsafe_b64decode.
• No padding handling added: Both the old and new code assume the base64 string is properly padded with =; if openinference ever omits padding the call will still raise and be swallowed by the outer except block.

Confidence Score: 4/5

Safe to merge — the change is a one-line normalization that fixes a real compatibility gap and degrades gracefully on failure.

The normalization logic is correct and the error path is already guarded by the outer try/except. The only gaps are a missed opportunity to use the stdlib urlsafe_b64decode and the absence of unit tests for the new code path.

langfuse/media.py — worth adding a unit test in tests/unit/test_media.py that covers URL-safe encoded inputs.

Sequence Diagram

%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
    participant Caller
    participant LangfuseMedia
    participant base64

    Caller->>LangfuseMedia: _parse_base64_data_uri(data_uri)
    LangfuseMedia->>LangfuseMedia: "validate 'data:' prefix & split header/body"
    LangfuseMedia->>LangfuseMedia: check 'base64' in header parts
    Note over LangfuseMedia: NEW: normalize URL-safe chars<br/>replace('-','+').replace('_','/')
    LangfuseMedia->>base64: b64decode(normalized_data)
    base64-->>LangfuseMedia: raw bytes
    LangfuseMedia-->>Caller: (bytes, content_type)

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"}}}%%
sequenceDiagram
    participant Caller
    participant LangfuseMedia
    participant base64

    Caller->>LangfuseMedia: _parse_base64_data_uri(data_uri)
    LangfuseMedia->>LangfuseMedia: "validate 'data:' prefix & split header/body"
    LangfuseMedia->>LangfuseMedia: check 'base64' in header parts
    Note over LangfuseMedia: NEW: normalize URL-safe chars<br/>replace('-','+').replace('_','/')
    LangfuseMedia->>base64: b64decode(normalized_data)
    base64-->>LangfuseMedia: raw bytes
    LangfuseMedia-->>Caller: (bytes, content_type)

Loading

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
langfuse/media.py:294-296
Python's standard library already provides `base64.urlsafe_b64decode`, which performs exactly this alphabet substitution (`-`→`+`, `_`→`/`) internally. Using it directly is more idiomatic and makes the intent self-documenting.

```suggestion
            return base64.urlsafe_b64decode(actual_data), cast(MediaContentType, content_type)
```

### Issue 2 of 2
langfuse/media.py:294-296
**Missing test coverage for URL-safe base64 inputs.** The PR checklist acknowledges no tests were added. `tests/unit/test_media.py` has no case exercising the `-` and `_` substitution path, so a future regression (e.g. re-introduction of `b64decode` without normalisation) would go undetected. Consider adding a unit test that passes a URL-safe encoded data URI and asserts the decoded bytes match the expected output.

_{Reviews (1): Last reviewed commit: "fix(media): normalize urlsafe base64 for..." | Re-trigger Greptile}

Greptile also left 1 inline comment on this PR.

[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.

[greptile-apps]

Missing test coverage for URL-safe base64 inputs. The PR checklist acknowledges no tests were added. tests/unit/test_media.py has no case exercising the - and _ substitution path, so a future regression (e.g. re-introduction of b64decode without normalisation) would go undetected. Consider adding a unit test that passes a URL-safe encoded data URI and asserts the decoded bytes match the expected output.

Prompt To Fix With AI

This is a comment left during a code review.
Path: langfuse/media.py
Line: 294-296

Comment:
**Missing test coverage for URL-safe base64 inputs.** The PR checklist acknowledges no tests were added. `tests/unit/test_media.py` has no case exercising the `-` and `_` substitution path, so a future regression (e.g. re-introduction of `b64decode` without normalisation) would go undetected. Consider adding a unit test that passes a URL-safe encoded data URI and asserts the decoded bytes match the expected output.

How can I resolve this? If you propose a fix, please make it concise.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!