Skip to content

Releases: Arize-ai/phoenix

arize-phoenix: v5.5.0

19 Oct 17:52
6628b87
Compare
Choose a tag to compare

5.5.0 (2024-10-19)

Features

  • Refactor subscription streaming and plumb through anthropic support (#5105) (d405f61)

arize-phoenix: v5.4.0

18 Oct 23:26
33eac82
Compare
Choose a tag to compare

5.4.0 (2024-10-18)

Features

  • add support for secure wss connection (#5110) (9bf0740)
  • playground: return finalized span at end of subscription (#5089) (6ca4288)
  • playground: show span details when playground run ends (#5108) (acce9d9)

Bug Fixes

  • Derive playground values instead of storing them in state (#5067) (d935d4b)
  • playground: optional type in playground tool calls (#5080) (925b496)

Documentation

  • anthropic tools tracing tutorial and fixture (#5094) (51b0201)

arize-phoenix-evals: v0.17.2

18 Oct 22:01
ec65059
Compare
Choose a tag to compare

0.17.2 (2024-10-18)

Bug Fixes

arize-phoenix: v5.3.1

17 Oct 22:43
83dd769
Compare
Choose a tag to compare

5.3.1 (2024-10-17)

Bug Fixes

  • playground: add openai as a container dependency (#5074) (6bc625c)
  • Re-enable run button on subscription error (#5081) (1418472)

arize-phoenix: v5.3.0

17 Oct 21:27
491ac18
Compare
Choose a tag to compare

5.3.0 (2024-10-17)

Features

  • add litellm integrations UI (#5028) (5e895cb)
  • Add model listing (#4948) (94d92fb)
  • js: Scaffold ts phoenix client (#4847) (24f447e)
  • make model listing more generic (#5022) (fd0c6ce)
  • Playground UI prototype (#4778) (a0a83a2)
  • playground: accumulate errors while parsing span attributes (#4941) (c27dfb1)
  • playground: add / delete messages (#4947) (5d45e48)
  • playground: add credential storage (#4970) (b15f2a4)
  • playground: add tools ui (#5002) (767bd37)
  • playground: allow up to 4 instances, distinguish alphabetically (#4951) (240bfe8)
  • playground: copy messages to clipboard (#5000) (23e9d82)
  • playground: create llm span for playground runs (#4982) (39c9940)
  • playground: Extract and display variables from all message templates as "inputs" (#4994) (b8cd777)
  • playground: Implement codemirror based template string editor (#4943) (e20716f)
  • playground: Implement editable input variable textareas (#4987) (#5006) (ab09109)
  • playground: Implement message role picker callback (#4909) (b2a8bd0)
  • playground: make chat messages match style in traces (#4931) (ed6e05e)
  • playground: model selector (#4971) (025c33e)
  • playground: parse model name and infer provider form span (#5021) (45973b7)
  • playground: parse span attribute llm.input_messages for playground span replay (#4906) (f4b1d92)
  • playground: playground layout (#4978) (2c50441)
  • playground: plumb through and apply template variables (#5052) (d0b1641)
  • playground: provide a back to trace button from the span playgr… (#4954) (01227e2)
  • playground: re-ordering of messages using dnd (#4936) (ad27394)
  • playground: rudimentary graphql support for messages input (#4907) (ee1f85b)
  • playground: streaming chat completions (#4785) (5948ea3)
  • playground: support azure openai (#5065) (4df3f8a)
  • playground: toggle for the template language (#5004) (45755bb)
  • playground: tool call backend (#5027) (5485c03)
  • playground: wire up tool calling ui (#5029) (75b7000)

Bug Fixes

  • playground: authenticate websockets (#4924) (4a53e8e)
  • playground: ignore keyboard events in template messages for drag and drop (#4945) (2b6529f)
  • playground: invalidate cache for playground span project to ensure new span is refetched (#4991) (985042c)
  • playground: make messages collapsible so they can be dragged (#5062) (30d95b4)
  • playground: plumb through and record invocation parameters (#5005) (9d375e5)
  • playground: plumb through credentials (#5003) (0fa0c87)
  • playground: plumb through model name and providers (#4999) (23958bd)
  • playground: remove toolChoice if no tools provided (#5079) (c883b06)
  • playground: Remove unused statements in language grammars (#4992) (bbbfd22)

Documentation

  • add environment variable PHOENIX_CSRF_TRUSTED_ORIGINS (GITBOOK-885) (38fc7ba)
  • Add Evaluating Phoenix Traces nbconverted page (GITBOOK-889) (38912b4)
  • Adding Auth Video (GITBOOK-896) (11e5a07)
  • auth-related environment variables (GITBOOK-887) (8a8ec8d)
  • DSPy doc updates (GITBOOK-886) (b00def9)
  • Fixing Quickstart: Tracing (GITBOOK-888) (e0f9fba)
  • how to add evaluations to experiments after the fact (GITBOOK-895) (7b584ee)
  • initial playground PRD (#4777) (2a0ddd9)
  • Mistral 1.0 Changes (GITBOOK-883) (4e11e86)
  • Navbar changes (GITBOOK-884) (62798ad)
  • No subject (GITBOOK-890) (54556a1)...
Read more

arize-phoenix-evals: v0.17.1

17 Oct 22:59
fd64f40
Compare
Choose a tag to compare

0.17.1 (2024-10-17)

Bug Fixes

  • increase python upper bound to include python 3.13 for arize-phoenix-evals and arize-phoenix-otel (#5077) (ef5c893)

arize-phoenix-otel: v0.6.1

17 Oct 22:55
eb983ca
Compare
Choose a tag to compare
Pre-release

0.6.1 (2024-10-17)

Bug Fixes

  • increase python upper bound to include python 3.13 for arize-phoenix-evals and arize-phoenix-otel (#5077) (ef5c893)

arize-phoenix-otel: v0.6.0

16 Oct 20:44
c3257b6
Compare
Choose a tag to compare
Pre-release

0.6.0 (2024-10-16)

Features

  • Add environment variable setting for structured logging (#4635) (a50ca10)

arize-phoenix: v5.2.2

09 Oct 20:44
852dcb1
Compare
Choose a tag to compare

5.2.2 (2024-10-09)

Bug Fixes

  • add Starlette middleware for checking Cross-Site Request Forgery (CSRF) when trusted origins are specified via environment variable (#4916) (26f8e4b)
  • playwright (#4935) (199d0eb)

Reverts

arize-phoenix-evals: v0.17.0

09 Oct 18:49
81fa394
Compare
Choose a tag to compare

0.17.0 (2024-10-09)

Features

  • Always prompt as system for OpenAI Models (#4937) (5f28ef2)