v2.50.3

What's Changed

• fix(serializer): support numpy scalars by @hassiebp in #936

Full Changelog: v2.50.2...v2.50.3

v2.50.2

What's Changed

• feat(prompts): allow passing kwargs to precompile langchain prompt by @hassiebp in #934

Full Changelog: v2.50.1...v2.50.2

v2.50.1

What's Changed

• fix(langchain): link prompts to nested generations as well by @hassiebp in #933

Full Changelog: v2.50.0...v2.50.1

v2.50.0

What's Changed

• chore: upgrade dependencies by @hassiebp in #926
• feat(decorator): allow setting parent trace or observation id by @hassiebp in #932

Full Changelog: v2.49.0...v2.50.0

v2.49.0

What's Changed

• feat(langchain): link langfuse prompts to langchain executions by @hassiebp in #928

Full Changelog: v2.48.1...v2.49.0

v2.48.1

What's Changed

• fix(openai): multiple streamed tool calls by @hassiebp in #922
• fix(openai): time to first token by @hassiebp in #923

Full Changelog: v2.48.0...v2.48.1

v2.48.0

What's Changed

• feat(decorator): Enhance observe decorator to support usage without parentheses by @hassiebp in #918
• fix(llama-index): initialize OpenAI model serializers by @hassiebp in #916
• fix(langchain): ollama usage parsing by @hassiebp in #917
• fix(langchain/AzureChatOpenai): unnecessary logs by @hassiebp in #919
• fix(langchain): batch run to update trace outputs by @hassiebp in #920

Full Changelog: v2.47.1...v2.48.0

v2.47.1

What's Changed

• feat: improve error messages by @maxdeichmann in #913

Full Changelog: v2.47.0...v2.47.1

v2.47.0

What's Changed

• feat(prompts): optionally disable prompt caching when cache ttl is 0 by @marcklingen in #914
  • Docs: https://langfuse.com/docs/prompts/get-started#disable-caching

Full Changelog: v2.46.3...v2.47.0

v2.46.3

What's Changed

• fix: prevent exceptions while parsing HTTP response error bodies by @maxdeichmann in #911

Full Changelog: v2.46.2...v2.46.3