v1.18.16

What's Changed

• feat: support jsonschema option when new a sfn by @woorui in #921
• feat(llm-bridge): add x.ai provider by @fanweixiao in #927

Full Changelog: v1.18.15...v1.18.16

v1.18.15

What's Changed

• feat(llm-bridge): void provider's errors are sent directly to the end user by @woorui in #889
• feat(llm-bridge): return openai error message if request llm provider failed by @woorui in #896
• feat(llm-bridge): add cerebras llm provider by @venjiang in #895
• feat(llm-bridge): add anthropic llm provider by @venjiang in #897
• refactor(llm-bridge): ollama provider by @venjiang in #866
• fix(llm-bridge): correct usage in stream response by @woorui in #901
• feat(llm-bridge): Azure OpenAI Provider returns the token usage when requested with include_usage by @woorui in #906
• feat(cli): added cli run information by @venjiang in #907
• feat(llm-bridge): system prompt supports backend Overwrite and Prefix by @woorui in #917

Full Changelog: v1.18.14...v1.18.15

v1.18.14

What's Changed

• refactor: new caller from interfaces & new service from its option by @woorui in #883
• chore: add more docker platform support by @woorui in #886

Full Changelog: v1.18.13...v1.18.14

v1.18.13

What's Changed

• feat(ai-bridge): support Github Models by @fanweixiao in #885
• chore: developers can docker run ai bridge by @woorui in #855

Full Changelog: v1.18.12...v1.18.13

v1.8.12

What's Changed

• feat: Baselime as default OpenTracing provider to analysis TimeToFirstToken (TTFT) and TimeBetweenTokens (TBT) by @woorui in #872

BASELIME_API_KEY=xxxx yomo serve -c config.yaml for observability.

Full Changelog: v1.18.11...v1.18.12

v1.18.11

What's Changed

• feat(llm-bridge): add gemini provider by @woorui in #863
• refactor: use goroutinue instead of mutex to sync messages by @woorui in #842
• feat(llm-sfn): log function calling jsonschema when sfn start by @woorui in #847
• feat(log): 🌈 tinted console logs by @woorui in #851
• feat: function definition supports enum by @woorui in #848
• fix: return an error if data cannot be unmarshal as FunctionCall by @woorui in #849
• fix: error raised when function scheme required is null by @venjiang in #860

Full Changelog: v1.18.10...v1.18.11

v1.18.10

What's Changed

• feat(llm-bridge): add ollama provider with mistral model support by @wujunzhuo in #834
• fix(cli): serverless ctx mock pkg move bug by @woorui in #827
• fix(cli): yomo run sfn exit if init error by @woorui in #828
• chore(deps): bump github.com/sashabaranov/go-openai from 1.24.0 to 1.24.1 by @dependabot in #826
• fix: response should include token usage from previous function calls by @woorui in #830
• fix: ai bridge run sfn timeout after 60s by @woorui in #829
• fix: example by @venjiang in #838
• perf: optimize yomo build speed by @venjiang in #840
• feat(ai-bridge): handle llm-sfn return nothing or multiple results returned by @woorui in #833
• feat: improve warning log if llm-sfn not return anything by @woorui in #841

Full Changelog: v1.18.9...v1.18.10

v1.18.9

What's Changed

• refactor: serverless context by @venjiang in #822
• feat: client error handler accept reconnect error by @woorui in #823

Full Changelog: v1.18.8...v1.18.9

v1.18.8

What's Changed

• feat: completions api flush after first calling by @woorui in #807
• fix: improve GetChatCompletions by @woorui in #809
• feat: client cannot write tag that within reserved range by @woorui in #815
• fix: test-prompt command error by @venjiang in #816
• feat: add transID for ai function by @woorui in #817

Full Changelog: v1.18.7...v1.18.8

v1.18.7

What's Changed

• test: add unit test for ai package by @fanweixiao in #769
• refactor: cli test-prompt subcommand by @venjiang in #772
• fix: client close conn by @woorui in #773
• add tests for api_server and provider by @fanweixiao in #770
• fix: the yomo test-prompt command error when running multiple sfn instances by @venjiang in #774
• test: fix unittest by @woorui in #780
• fix: ctrl+c cannot exit client if the client is reconnecting by @woorui in #787
• refactor: llm api by @venjiang in #781
• fix: fix panic if cannot find a conn form the connector by @woorui in #797
• refactor: use openai pkg & support chat in stream by @woorui in #791
• feat: connector can be customized in NewZipper() by @woorui in #801
• fix: service system prompt init as an empty string by @woorui in #804
• refactor: handshake with ai function definition by @woorui in #798
• refactor: cli by @venjiang in #802

Full Changelog: v1.18.6...v1.18.7