Skip to content

Latest commit

 

History

History
475 lines (339 loc) · 30.3 KB

CHANGELOG.md

File metadata and controls

475 lines (339 loc) · 30.3 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

0.9.2 - 2024-08-14

Fixed

  • Stack fails when Call Audio Source is set to Genesys Audio Hook - regression in Genesys nested stack introduced in v0.9.0 #178

0.9.1 - 2024-08-12

Fixed

  • Stack fails when using a CustomLanguageModel due to logic error in ValidateParams in v0.9.0 #176

0.9.0 - 2024-08-08

Fixed

  • Stack deployment now fails fast if required Bedrock models are not available or enabled in the account/region
  • Mic input role (CALLER vs. AGENT) assignment not working #169
  • Missing call transcript download button #155

Added

  • Support for Bedrock anthropic-claude-3-haiku and anthropic-claude-3-sonnet models (removed support for older models) #152
  • Knowledge bases for Amazon Bedrock replaces Amazon Kendra for Agent Assist #157
  • Option for single language auto-detection - using Amazon Transcribe's 'Identify Language' feature with optional list of possible language codes. #153
  • Option for multiple language auto-detection - using Amazon Transcribe's 'Identify Multiple Languages' feature with optional list of possible language codes. #153
  • Option to integrate Talkdesk Audio stream as a call data source (see Talkdesk) #174
  • Ability to override default Transcribe language code using SIPREC Call initialization lambda hook #165
  • Ability to specify multiple language-specific CustomVocabulary or CLM and associated logic to select the appropriate one based on dynamic LanguageCode when using SIPREC Call initialization lambda hook #165
  • Simplified and more robust default demo agent recording and demo script for demos using Asterisk.

0.8.13 - 2024-05-22

Fixed

  • #138: Very delayed transcripts when using CallTranscriberLambda
  • #140: Very delayed transcripts when using Connect-KVS audio ingestion

0.8.12 - 2024-05-20

Fixed

  • Expanded redaction support to include all languages (and dialects) on the Amazon Transcribe supported languages page.

  • Call Transcriber Lambda

    • Refactor call end handling, to explicitly process ENDED messages from VoiceConnector, rather than imputing call end from stream data ending - adds robustness when stream can end (temporarily) for other reasons (such as call hold, etc.)
    • Identified a situation where if a call chained across multiple lambda invocations (lasted longer than 12 minutes), the new timer we introduced that checks if the call ends would leak into the next lambda invocation. This would have no effect for a single long running call other than duplicate checks to DynamoDB—however—if the call ends and a NEW call uses that warm Lambda, the new call would inherit the timers and within 5 seconds mark the call as ended. To fix this, we made modifications by adding proper clean-up of timers when the call ends, and also double checking when the next call begins that the new Lambda did not inherit any calls.
    • Added the callId to the log output, so that it is easier to trace calls.
    • Added the request id of the next Lambda invocation in the output of the log, so that it makes it easier to trace sequential lambda invocations (chaining).

  • Websocket server and Stream Audio client

    • Fix state management in Websocket server - add connection specific state variables to a global dictionary to maintain state for active connections
    • Fix reconnection logic in Stream Audio websocket client - i.e. retrofit the fix from LMA.
    • Clean up and streamline log formats in Websocket Server and Stream Audio client
    • Reduce the amount of logs due to health checks. Health check is done every 10 seconds, logging status for each health check which creates huge amounts of log records in cloudwatch. To fix this - Log first health check message from each new source and then log every 2 minutes (interval controlled by a new environment variable which is not exposed in cloudformation, but can be directly updated in the ECS configuration by developers).
    • Fix authentication code in Websocket client utility (utilites/webosocket-client)

0.8.11 - 2024-02-21

Fixed

  • Added custom resource in the lca-websocket-stack that sets the ECS cluster's desired service count to zero, then deletes the service, before deleting the cluster. This will prevent errors in deleting the stack when disabling websockets or deleting LCA.
  • Updated NodeJS version to 18
  • Removed VPC flow logs from being enabled
  • Fixed CloudFormation condition that incorrectly deployed PCA integration functions
  • Fixed issue with call event processor not correctly parsing ChimeSDK Voice Connector metadata updates

Added

  • New Cloudscape user interface
  • Support for Amazon Connect Kinesis Video Streams as an audio ingestion source. See Amazon Connect KVS README for more details.
  • Added names to VPCs for ease of locating them

0.8.10 - 2023-11-25

Added

  • New! Websocket audio ingestion server (optional). Use it to enable audio streaming from a custom desktop client application, from the sample streaming app (below), or from the new LCA Web UI audio streaming client. See Websocket server.
  • New! Sample audio streaming application which streams a stereo audio file from your desktop to LCA via the new websocket server. Use this to replay recordings, e.g. for testing transcription accuracy after CV/CLM changes, for testing agent assist features, etc. Or use the code as a reference implementation for building your own custom streaming app for LCA. See Websocket client app.
  • New! Web UI audio streaming client built into the LCA UI. Use this to stream audio into LCA from (a) your computer microphone, and (b) from a local broswer tab that is running a softphone or meeting app, or playing an audio recording. See Web UI streaming client.

Fix

  • publish.sh script now runs on MacOS (now works with GNU or BSD sed command).

0.8.9 - 2023-11-01

Added

  • Allow customer to specify their own Voice Connector as a CloudFormation Template parameter. If the customer provided Voice Connector is provided, LCA will not deploy a new VC. (#102)

Fixed

  • Fix LCA/PCA integration when using Chime SDK Call Analytics (PR #108, #102)
  • Call recording player now appears when recording file is available 1-2 minutes after the call when using Chime SDK Call Analytics (#108)
  • Prevent START_CALL_PROCESSING if call has already ended. This removes possibility that START CALL PROCESSING picks up re-purposed KVS streams for a different callId (PR #114)
  • Reduce agent assist latency when using Connect / Contact Lens by processing Utterance rather than Transcript events (PR #113)
  • Improved markdown formatting for demo script agent assist responses
  • Addressed stack update issues with Chime SDK Voice Connector, demo Asterisk instance, & Bedrock boto3 layer
  • Use latest QnAbot on AWS release (5.4.4) for Agent Assist
  • Dependabot security updates

0.8.8 - 2023-10-10

Added

  • Agent Assist bot can now generate a draft email from agent to caller, summarizing call and next steps

Fixed

0.8.7 - 2023-10-03

Added

  • Generative transcript summarization and LLM powered Agent Assist are now enabled by default using Amazon Bedrock.
  • Agent Assist generative answers are now enabled by default, using an Amazon Bedrock LLM model.
  • Agent Assist now enables semantic similarity item matching by default, using an Amazon Bedrock Embeddings model.
  • Generative transcript summarization now generates multiple call insight types.
  • Generative transcript summarization and insights now customizable with configurable prompt templates.
  • In-progress call summarization using the Agent Assist bot - quickly bring new agent or supervisor up to speed in transfer and escalation scenarios. Prompt templates are customizable using QnABot's content designer.

Fixed

  • Race condition between Asterisk Server and Chime SDK Voice Connector setup
  • Flawed timeout logic for long running summaries - PR#100
  • Hung calls when using Chime Call Analytics when callid was modified using call initialization lambda - PR#97
  • ContactLens categories not displayed when using Contact Lens integration

0.8.6 - 2023-09-14

Fixed

  • Fix LCA 0.8.5 Stack updates fail when changing Agent Assist parameters - Issue #96

0.8.5 - 2023-09-03

Added

  • Update QnABot to latest version 5.4.1
  • Added support for Amazon Bedrock (preview) for QnABot agent assist and call transcript summarization

Fixed

  • Fix Chime SDK VoiceConnector problems when updating stack
  • Fix Demo Asterisk configuration preventing use of softphone
  • Misc improvements to build/publish script

0.8.4 - 2023-08-11

Fixed

  • Fix issue with deploying new Chime Voice Connector if in a new AWS account
  • Fix Contact Lens summaries

0.8.3 - 2023-07-18

Fixed

  • Fix regression issue that prevented persistence of custom metadata json in DynamoDB call record
  • Update QnAbot submodule branch/commit to pull in deployment fixes from QnAbot v5.3.5

0.8.2 - 2023-07-11

Fixed

  • Fix stack update problems related to Chime Voice Connector configuration when updating to v0.8.1 from v0.8.0 with Chime Call Analytics enabled

0.8.1 - 2023-07-10

Fixed

  • Add compatibility with Start Call Processing feature, when using Chime SDK Call Analytics audio processor.

0.8.0 - 2023-05-30

Added

Changed

  • Update Code Build version
  • Improve throughput by decoupling agent assist inference latency from main call event processor - now runs asynchronously in parallel with transcript processing.
  • Refactored call event processor code to normalize and merge handling of Contact Lens, Transcribe, Transcribe Call Analytics, and Amazon Chime SDK Call Analytics messages.

0.7.2 - 2023-05-03

Fixed

  • Hot fix for S3 bucket ACL changes
  • Changed default function memory for Call Event Processor Lambda

0.7.1 - 2023-03-11

Added

  • Transcribe Call Analytics "POST_CALL" categories now displayed in the LCA UI after call ends when using Transcribe 'analytics' mode.
  • New download buttons on call details page to save call summary or call transcript to local Excel file.
  • Add configurable sentiment score thresholds for fine grain control over negative and positive sentiment scoring when using Transcribe 'standard' mode.
  • Update Agent Assist QnAbot version to v5.3.0 with optional semantic search support for FAQs.

Changed

  • Improved performance and scalability changes (tested to 300 concurrent calls):
    • Refactor call aggregate logic to make call event processor fully stateless - removed dependency on Kinesis tumbling window to enable multiple concurrent invocations per shard.
    • enable AppSync resolver caching to enable fast/efficient call state queries from call event processor.
    • enable Kinesis Enhanced Fan Out to reduce message read latency.
    • increase call event processor Lambda memory to 5120M to reduce start time and enable greater message handling concurrency in each invocation.
    • avoid multiple mutation retries on 'put item condition failure' - where retries will not resolve the condition and merely extend function duration.
    • call event processor now invokes transcript summarization asynchronously, avoiding blocking and choking execution concurrency limits.
    • retries for Start Transcription Stream if exceptions thrown, to allow improved tolerance for call bursts resulting in temporary TPS limit exceeded errors.
  • Fix issue with START_CALL_PROCESSING event rule ( introduced in 0.7.0 with the multiple LCA stack fix).
  • Fix issue introduced in 0.7.0 preventing invocation of custom transcript processing Lambda Hook function
  • Miscellaneous improvements to test scripts - see README
  • Additional minor fixes - see commit history

0.7.0 - 2023-02-12

Added

  • Experimental generative transcript summarization to provide a short paragraph with a synopsis of each completed call; use the built-in summarization model or experiment with custom language models or APIs of your choice. See Transcript Summarization.
  • Utility Lambda function that retrieves call transcription from DynamoDB. See Fetch Transcription Lambda.
  • Optional translation of live or completed call transcripts into language of choice, using Amazon Translate.
  • Ability to disable display of agent channel transcription in the call transcript pane.
  • Test scripts for simulating phone calls. See Asterisk Test Scripts.
  • LCA client utility to make it easier to test Call Event Processors and LCA UI without having to actually make a phone call. See LCA Client.
  • Download button on Calls page to save call list to local Excel file.
  • Default audio recording used by demo Asterisk server now plays the agent side of the Agent Assist demo script.

Changed

  • Fix bug in Call Transcriber Lambda that caused double transcription ifAmazon Chime SDK Voice Connector can't differentiate between caller and agent streams (caused by SBC not configured to use RFC 7865 metadata).
  • Fixed issue where multiple LCA stacks in the same region processed calls from each other's Chime VS streams. Now each stack uses only it's own Chime VC instance, so you can run multiple stacks in the same account and region.
  • Miscellaneous dependabot updates

0.6.0 - 2022-11-27

Added

  • Supports new Amazon Transcribe Real-time Call Analytics streaming API
  • Real-time Detected Issues, Call Categories, and Alerts
  • Real-time Category and Alert Notifications via Amazon SNS subscriptions. See Category Notifications.
  • Post call analytics (without additional transcription costs) through integration with the companion Post Call Analytics (PCA) solution.
  • QnABot designer markdown answers enable rich text and media in Agent Assist messages

Changed

  • Latest QnABot (v5.2.4) now used for agent assist
  • Improved logging in Call Transcriber lambda
  • Extend payload for Agent Assist Lambda to include dynamodb_table_name and dynamodb_pk, so Lambdas can query the LCA Call Event table in DynamoDB to retrieve call metadata.

0.5.2 - 2022-10-20

Added

  • Support for using Transcribe Custom Language Models.
  • Option to suppress partial transcription segments.
  • Support for custom transcript processing logic via a user provided Lambda Hook function. See Transcript Lambda Hook Function.
  • ChimeVC call initialization Lambda hook can attach optional arbitrary metadata json object to call record, for use by downstream custom applications via the LCA graphQL API or DynamoDB event sourcing table. See ChimeVC call initialization Lambda hook.
  • Ability to selectively disable call recordings using ChimeVC call initialization Lambda hook.

Fixed

  • Agent Assist configuration is now correctly maintained during stack updates when CallEventProcessor function is replaced.

0.5.1 - 2022-09-30

Added

  • CallTranscriber now has the ability to delay start of call processing until a (new) START_CALL_PROCESSING event is received - useful when call is streaming is initiated when call is established but transcribing should be delayed until after IVR navigation and hold time and triggered only when agent and caller are connected. See Start call processing for a in-progress call.
  • Support markdown and html rendering in LCA UI
  • Use QnABot markdown response when available, to take advantage of new markdown rendering feature for Agent Assist.

Changed

  • CallTranscriber Lambda now has it's own DynamoDB table and no longer shares a table with the AISTACK components.

Fixed

  • Add ExpiresAfter fields to AGENT_ASSIST messages, used as TTL in DynamoDB
  • Check StackName length and fail fast if length would cause downstream stack failures caused by long resource names in nested stacks. Max Stack Name is now 25 characters.
  • Provide 'reason' message for stack custom resource failures.
  • Expand Troubleshooting README.

0.5.0 - 2022-09-11

Added

  • AgentID call attribute and associated API support for setting, displaying, sorting, and filtering calls by Agent. See Setting AgentId.
  • AgentId field automatically assigned from Amazon Connect contact events when using Amazon Connect Contact Lens as the Call Audio Source.
  • Support for custom logic via a user provided Lambda function to selectively choose which calls to process, toggle agent/caller streams, assign AgentId to call, and/or modify values for CallId and displayed phone numbers. See Lambda Hook Function for SIPREC Call Initialization.
  • Configurable retention period for call records (default 90 days). Records and transcripts that are older than this number of days are permanently deleted.
  • UI supports new 'Load: 2 hrs' option for improved performance in high volume contact centers.

Changed

  • Moved transcriber Lambda out of AI stack and into ChimeVC stack.
  • Remove code for call event stream processing lambda no longer used since LCA v0.4.0.
  • Rename TranscriptProcessorLambda to CallEventProcessorLambda to reflect that it will process call analytics and contact metadata events in addition to transcripts.
  • Rename lca-ai-stack CF template.
  • Asterisk demo server reinstalled on instance reboot such as during stack updates containing Asterisk configuration or version changes.
  • Asterisk demo installation script is no longer dependent on hardcoded Asterisk version.
  • Asterisk demo server is reloaded each hour to resolve observed busy tones in previous releases.
  • Default Asterisk demo server version is now v19.
  • DynamoDB event sourcing table now maintains only one item per transcript segment and no longer maintains partial segments.
  • ChimeVC CallTranscriber Lambda function now emits one Call START event when both call streams are ready (as opposed to one per call stream), eliminating 'item already exists' errors in the CallEventProcessorLambda lambda
  • ChimeVC CallTranscriber Lambda function memory footprint reduced to 768MB to improve cost efficiency with minimal latency tradeoff.
  • README updates

0.4.1 - 2022-08-22

Changed

  • Remove E.164 type enforcement on CustomerPhoneNumber and SystemPhoneNumber. Any string value is now allowed, enabling calls to be processed when either/both CustomerPhoneNumber and SystemPhoneNumber fields are non E.164 strings.

0.4.0 - 2022-07-15

Added

Changed

  • Latest Asterisk version (18) for demo PBX
  • Solution title is now "Live Call Analytics with Agent Assist"

0.3.0 - 2022-05-18

Added

  • Support for Genesys AudioHook. You can now stream call audio into the Live Call Analytics solution from a Genesys AudioHook. See the details in the README file

Changed

  • Changed the audio stream consumer and transcriber component to run on AWS Lambda instead of AWS Fargate. This new Call Transcriber provides the following benefits:
    • Reduced Amazon Transcribe processing cost. The agent and caller audio streams are now automatically merged into a single stereo stream. Both the agent and caller audio streams are transcribed in a single session
    • Longer call duration handling. The Fargate consumer had a timeout of 42 minutes. The new Lambda based transcriber has a mechanism that allows to transparently handoff the call processing from one Lambda invocation to another
    • Faster scaling on sudden call volume spikes
    • Simplified stereo recording process. The stereo recordings are now produced by the Call Transcriber instead of merging recording files after the call was done
    • Serverless!
  • Additional language support
  • The Call Event Stream Processor Lambda function now runs on the arm64 architecture. The new Call Transcriber Lambda function also runs on arm64. The build was updated to enable arm64 emulation when using the SAM CLI in Amazon Linux 2
  • Partial transcript events are now only persisted for 1 day in DynamoDB. Final transcript events are persisted for 90 days by default
  • Separated the build and development python virtual environments to avoid development dependencies interfere with the SAM CLI
  • Updated the Python GraphQL client in the Call Event Stream Processor Lambda function to the released stable version gql v3.2.0. This version is now out of pre-release which removed the need for the Makefile based SAM CLI build of the Lambda layer. The library also added direct support for AppSync which removed the need for the custom AppSync transport and authentication code. The Makefile build and custom AppSync code has been deleted
  • Updated dependency versions of various components including:
    • Web UI
    • Call Event Processor Lambda function
    • Project build and development
  • Updated nodejs and npm versions used in CodeBuild to build the UI
  • Fixed demo agent recording download issue

Removed

  • The resources associated with the Fargate consumer such as Fargate cluster/service/task definition, VPC, SQS queues and autoscaling were removed in favor of the new Lambda based Call Transcriber
  • The resources associated with the creation of post call stereo recordings (including Lambda functions, S3 bucket and SQS queue) were removed in favor of the new Call Transcriber that merges the audio into stereo recordings at call time

0.2.1 - 2022-02-02

Fixed

  • Fixed semantic version in main cloudformation template

0.2.0 - 2022-02-02

Added

  • Added script to update semantic versions in source files. scripts/update-version.sh
  • Added TROUBLESHOOTING.md for instructions on how to check for errors
  • Added web UI admin user creation via CloudFormation. The email address of the admin user is passed via the AdminEmail CloudFormation parameter. An initial temporary password is automatically sent to this user via email. This email also includes the link to the web UI
  • Added CloudFormation parameter to enable or disable Sentiment Analysis. See the IsSentimentAnalysisEnabled parameter
  • Added CloudFormation mapping to configure the Amazon Comprehend language from the selected Amazon Transcribe language
  • Added support to select the Spanish language (es-US) in the CloudFormation template. The CloudFormation template now allows to select either English (en-US) or Spanish (es-US) using the TranscribeLanguageCode parameter. NOTE: Content redaction is only available when using the English language (en-US). It is automatically disabled when using other languages
  • Added the CloudFrontAllowedGeos CloudFormation parameter to control the CloudFront geographic restrictions. You can specify a comma separated list of two letter country codes (uppercase ISO 3166-1) that are allowed to access the web user interface via CloudFront. For example: US,CA. Leave empty if you do not want geo restrictions to be applied

Changed

  • The CloudFormation AllowedSignUpEmailDomain parameter is now optional. If left empty, signup via the web UI is disabled and users will have to be created using Cognito. If you configure a domain, only email addresses from that domain will be allowed to signup and signin via the web UI
  • The CloudFront distribution now defaults to no geographic restrictions. There's a new parameter named CloudFrontAllowedGeos that allows you to add geographic restrictions. If you leave this parameter empty, the previous geographic restriction will be removed on an update to this version. The previous version had a hardcoded value that set the restriction to US only. Set the CloudFrontAllowedGeos to US if you want to preserve the previous default configuration after updating to this version

Fixed

  • Reverted kvs stream parser library version workaround
  • Asterisk server will wait for voice connector to get a phone number before initializing

0.1.0 - 2021-12-16

Added

  • Initial release