Releases: foundation-model-stack/fms-guardrails-orchestrator
Releases · foundation-model-stack/fms-guardrails-orchestrator
0.9.0 - Client OpenTelemetry, bidirectional streaming content, deps update
What's Changed
- 📝 Document detector rules by @evaline-ju in #255
- ✨ Introduce tower service into clients by @evaline-ju in #259
- ✨ Client metric trace layers by @evaline-ju in #260
- 📝 OpenTelemetry docs by @evaline-ju in #264
- Expose modules and types publicly for integration tests, add ServiceConfig::new by @declark1 in #267
- ♻️ Combine tracing info logs by @evaline-ju in #269
- ♻️🔊 Update grpc client trace layer by @evaline-ju in #270
- Update chat_completions to stream ChatCompletionChunk messages by @declark1 in #268
- 🔥 Remove legacy tokenization client by @evaline-ju in #273
- Stream content endpoint by @mdevino in #272
- Bug fix - added json header to detector client response by @resoluteCoder in #279
- ⬆️ Upgrade Rust and project dependencies by @evaline-ju in #281
- ⬆️♻️ Update open telemetry dependencies by @evaline-ju in #282
- Add detector_id in response for orchestrator apis by @swith004 in #278
New Contributors
API update
detector_id
is now included in orchestrator responses, to allow users to easily attribute each detection to a detector
Full Changelog: 0.8.0...0.9.0
0.8.0
What's Changed
- 🐛 Fix threshold getting passed through beyond orchestrator proces… by @gkumbhat in #236
- ✨🐛 Pass on non-threshold detector parameters by @evaline-ju in #235
- ✨ Integrate telemetry client instrumentation and traceparent propagation in orchestrator by @pscoro in #231
- ✨ Add OpenAiClient stream handling, update types by @declark1 in #230
- added error messages to various unwraps by @resoluteCoder in #243
- ✨ Implement Chat Completions API by @declark1 in #240
- 📝 Streaming content API by @evaline-ju in #246
- 🐛 Update stream content API with intended response by @evaline-ju in #251
- 🐛 added ability to cancel streaming generation when client disconnects by @resoluteCoder in #254
- 🔧 Migrate clients implementation from
reqwest
tohyper
by @pscoro in #252
New Contributors
- @resoluteCoder made their first contribution in #243
Full Changelog: 0.7.0...0.8.0
0.7.0
What's Changed
- 🔧 Remove context from chat endpoint in detector API by @evaline-ju in #222
- 🔧 Add detector_params as optional for every detectors endpoint by @evaline-ju in #212
- Client refactor by @declark1 in #220
- Initial opentelemetry setup for orchestrator by @pscoro in #221
- 🐛 Fix whole_doc_chunker missing from client list by @gkumbhat in #229
- Add missing licence comments on recent new files by @pscoro in #232
- Chat standalone endpoint by @mdevino in #233
Main config change
detectors:
# Detector ID/name to be used in user requests
hap-en:
# Detector type (text_contents, text_generation, text_chat, text_context_doc)
type: text_contents # <----- (1) Add a detector type!
service:
hostname: localhost # <--- (2) No longer requires the full URL, just the base hostname
port: 8080
Full Changelog: 0.6.0...0.7.0
0.6.0
What's Changed
- 💡 Add config description by @evaline-ju in #186
- 🧪 Add axum-test package for integration tests by @mdevino in #188
- ✨ Detection result ordering by start span by @evaline-ju in #190
- 🔊 Add http error source logging by @gkumbhat in #199
- ♻️ Re-order endpoints and add tags by @evaline-ju in #201
- ✅ 🧪 Prepare integration tests for stateful request handling without mocking orchestrator by @pscoro in #200
- 📝 Document APIs and server config in README by @evaline-ju in #204
- 🦯 🧑⚕️ Orchestrator HTTP/gRPC health checks and health readiness probe by @pscoro in #183
- 📝 Document
/health
detector endpoint by @mdevino in #207 - 📝✨ Add API for detections on chat completions and ADR by @evaline-ju in #209
- 🔧 Update standalone chat detections API by @evaline-ju in #211
- 🔧 Turnoff default health check by @gkumbhat in #216
- 🐛 Fix the config path CLI arg's corresponding env variable to match the documentation by @pscoro in #219
- ✨ Add header passthrough for NLP and detector clients by @gkumbhat in #214
Full Changelog: 0.5.0...0.6.0
0.5.0 - Updated chunker API, improved streaming
What's Changed
- Orchestrator Service Config Validation by @pscoro in #176
- 🐛 whole doc chunker streaming hang by @declark1 in #182
- ♻️ Refactor aggregator architecture by @declark1 in #154
- 🏷️ Update based on open-sourced types by @evaline-ju in #185
New Contributors
Full Changelog: 0.4.1...0.5.0
0.4.1
0.4.0 - Evidence support and deps update
What's Changed
- Update Rust & cargo dependencies, cleanup models.rs by @declark1 in #155
- Configure explicit tls roots for tonic client by @declark1 in #157
- Update DetectorParams to newtype wrapper by @declark1 in #156
- 📝 Update evidence for detector API and context docs by @evaline-ju in #160
- Update Document Context detector endpoint on ADR by @mdevino in #168
- 🐛 Fix detector params not getting passthrough to detector for context type by @gkumbhat in #170
- 🏷️ Make evidence value optional by @evaline-ju in #174
Full Changelog: 0.3.0...0.4.0
0.3.0 - New endpoints
What's Changed
- ✨ Generation detection unary endpoint by @mdevino in #116
- 🔧 Allow for configuration of request timeouts per service by @evaline-ju in #140
- 🐛 Fix edge case validation for masks that reach end of inputs by @evaline-ju in #148
- ✨ Implementation of /context endpoint by @mdevino in #147
- ✨ Implement /detection/generated by @mdevino in #152
Full Changelog: 0.2.0...0.3.0
0.2.0 - Error handling and fixes
What's Changed
- 🔧 Increase default timeouts by @gkumbhat in #134
- ✨Stream Error Handling by @declark1 in #120
- 🐛 Fix presence of seed, input_token and token count in streaming resp by @gkumbhat in #137
- ✨ Implement
/api/v1/text/task/detection/content
endpoint by @mdevino in #125 - 🛠️ Error propagation for 404 and 503 by @mdevino in #139
- 🚚 Rename orchestrator API to v2 by @gkumbhat in #141
- 🐛 Fix handle detector client error deserialization error by @declark1 in #143
- 🐛 Fix whitespace bug #130 by @flaviabeo in #135
New Contributors
- @flaviabeo made their first contribution in #135
Full Changelog: 0.1.0...0.2.0