Merge branch 'develop' v0.7.6

CHANGELOG.md

@@ -6,9 +6,21 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.7.6] - 2024-02-16
+### Added
+- PCA UI now shows the status of the step function workflow for each call.
+- Added support for the new Transcribe Call Analytics generative call summarization option. 
+
+### Fixed
+- Dependabot updates for PCA
+- #234 Fix exception for files that contain no speech segments. 
+- Fix input bucket trigger to not create a DynamoDB record for metadata files.
+- Updated NodeJS to v16.
+- Fix bug when deploying with Amazon Titan Text Express.
+
 ## [0.7.5] - 2024-01-17
 ### Added
-- Support for larger prompts by storing LLMPromptSummaryTemplate in S3 rather than SSM. By default, the CF templtae will migrate existing SSM prompts to DynamoDB.
+- Support for larger prompts by storing LLMPromptSummaryTemplate in S3 rather than SSM. By default, the CF template will migrate existing SSM prompts to DynamoDB.
 
 ### Fixed
 - #125 Updated the pca-aws-sf-bulk-queue-space.py function to correctly count jobs based on IN_PROGRESS as well as QUEUED
@@ -160,7 +172,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 - Initial release
 
-[Unreleased]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/compare/v0.7.5...develop
+[Unreleased]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/compare/v0.7.6...develop
+[0.7.6]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/releases/tag/v0.7.6
 [0.7.5]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/releases/tag/v0.7.5
 [0.7.4]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/releases/tag/v0.7.4
 [0.7.3]: https://github.com/aws-samples/amazon-transcribe-post-call-analytics/releases/tag/v0.7.3

VERSION

@@ -1 +1 @@
-0.7.5
+0.7.6

docs/generative_ai.md

@@ -3,17 +3,23 @@
 Post-Call Analytics has an optional step in the step function workflow to generate insights with generative AI. 
 PCA supports [Amazon Bedrock](https://aws.amazon.com/bedrock/) (Titan or Anthropic models) and [Anthropic](https://www.anthropic.com/) (3rd party) foundational models (FMs). Customers may also write a Lambda function and provide PCA the ARN, and use any FM of their choice. The prompts below are based on Anthropic's prompt formats. Learn more about prompt design at Anthropic's [Introduction to Prompt Design].(https://docs.anthropic.com/claude/docs/introduction-to-prompt-design). 
 
-For Amazon Bedrock models, you must [request model access](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access.html) for the models selected.
+For Amazon Bedrock models, you must [request model access](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access.html) for the models selected. 
 
 PCA also supports 'Generative AI Queries' - which simply means you can ask questions about a specific call. These queries appear in a chat-like window from within the call details page.
 
 *All the prompts below were tested with Amazon Titan and Anthropic FMs.*
 
 **Note:** If you choose to call Anthropic directly, data will leave your AWS account!  Also, the Anthropic API key will be stored in [AWS Secrets Manager](https://docs.aws.amazon.com/secretsmanager/latest/userguide/intro.html), under the key `{StackName}-ThirdPartyApiKey`, where `{StackName}` is replaced with your PCA CloudFormation stack's name.
 
-## Generative AI Insights
+## How to enable generative AI Summarization and Insights
 
-When enabled, PCA can run one or more FM inferences against Amazon Bedrock or Anthropic APIs. The prompt used to generate the insights is stored in DynamoDB. The name of the table contains the string `LLMPromptConfigure`, and the table partition key is `LLMPromptTemplateId`. There are two items in the table, one with the partition key value of `LLMPromptSummaryTemplate` and the other with the partition key value of `LLMPromptQueryTemplate`.
+To enable generative AI summarization and insights, update the **CallSummarization** CloudFormation parameter with one of the following values: `BEDROCK`, `BEDROCK+TCA`, `ANTHROPIC`, `TCA-ONLY`.
+
+To use Transcribe Call Analytics (TCA) [generative call summarization](https://docs.aws.amazon.com/transcribe/latest/dg/call-analytics-batch.html#tca-summarization-batch), use `BEDROCK+TCA` or `TCA-ONLY` as the value.
+
+**Note:** If you enable TCA, the summarization templates below will skip the 'Summary' prompt for files that are analyzed with TCA. For audio files that are analyzed with Transcribe `standard` mode, such as mono audio files, the 'Summary' prompt will be executed.
+
+When summarization is enabled, PCA can run one or more FM inferences against Amazon Bedrock or Anthropic APIs. The prompt used to generate the insights is stored in DynamoDB. The name of the table contains the string `LLMPromptConfigure`, and the table partition key is `LLMPromptTemplateId`. There are two items in the table, one with the partition key value of `LLMPromptSummaryTemplate` and the other with the partition key value of `LLMPromptQueryTemplate`.
 
 ### Generative AI interactive queries
 

docs/output_json_structure.md

@@ -36,7 +36,8 @@ Contains header-level information around the analytics that have been generated,
   "OutcomesDetected": [ ],
   "Telephony": [ ],
   "SourceInformation": [ ],
-  "Summary": { }
+  "Summary": { },
+  "ContactSummary": { }
 }
 ```
 
@@ -62,6 +63,7 @@ Contains header-level information around the analytics that have been generated,
 | Telephony            | -        | *[Optional]* A list of telephony-specific metadata fields extract from the CTR files (only present if the chosen telephony CTR parser chooses to write this information out) |
 | SourceInformation    | -        | Source-specific details for the conversation.  Contains just one of any of the possible supported sources |
 | Summary    | -        | Key value pairs that define summary topics and values. These will be rendered inside the GenAI Call Summary panel in the user interface. The key will be rendered as the title, and the value is the body. |
+| ContactSummary    | -        | [Generative call summarization](https://docs.aws.amazon.com/transcribe/latest/dg/call-analytics-batch.html#tca-summarization-batch) output from Transcribe Call Analytics. This is a nested structure. See the structure [here](https://docs.aws.amazon.com/transcribe/latest/dg/tca-output-batch.html#tca-output-summarization-batch). |
 
 ###### SpeakerLabels
 

pca-boto3-bedrock/template.yaml

@@ -3,6 +3,11 @@ Description: >
   PCA Bedrock Boto3 Lambda Layer. This will create an S3 bucket, download
   the Boto3 WHL file, and create a Lambda layer for use.
 
+Parameters:
+  Boto3Version:
+    Type: String
+    Default: "1.34.40"
+
 Resources:
 
   BedrockBoto3Bucket:
@@ -56,6 +61,7 @@ Resources:
       Environment:
         Variables:
           BOTO3_BUCKET: !Ref BedrockBoto3Bucket
+          BOTO3_VERSION: !Ref Boto3Version
       Code:
         ZipFile: |
           import os
@@ -70,6 +76,7 @@ Resources:
           from datetime import datetime
           import cfnresponse
           boto3_bucket = os.environ['BOTO3_BUCKET']
+          boto3_version = os.environ['BOTO3_VERSION']
 
           def upload_file_to_s3(file_path, bucket, key):
               s3 = boto3.client('s3')
@@ -112,8 +119,8 @@ Resources:
             try: 
               if event['RequestType'] != 'Delete':
                 os.chdir('/tmp')
-                print(f"running pip install boto3==1.28.57")
-                subprocess.check_call([sys.executable, "-m", "pip", "install", "boto3==1.28.57", "-t", "python" ])
+                print(f"running pip install boto3=={boto3_version}")
+                subprocess.check_call([sys.executable, "-m", "pip", "install", f"boto3=={boto3_version}", "-t", "python" ])
                 boto3_zip_name = make_zip_filename()
                 zipdir("python",boto3_zip_name)
                 print(f"uploading {boto3_zip_name} to s3 bucket {boto3_bucket}")
@@ -144,7 +151,7 @@ Resources:
       ServiceToken: !GetAtt BedrockBoto3ZipFunction.Arn
       # Rerun BedrockBoto3ZipFunction if any of the following parameters change
       BOTO3_BUCKET: !Ref BedrockBoto3Bucket
-      VERSION: 1
+      VERSION: !Ref Boto3Version
 
   BedrockBoto3Layer:
     Type: "AWS::Lambda::LayerVersion"

pca-main-nokendra.template

@@ -1,6 +1,6 @@
 AWSTemplateFormatVersion: "2010-09-09"
 
-Description: Amazon Transcribe Post Call Analytics - PCA (v0.7.5) (uksb-1sn29lk73)
+Description: Amazon Transcribe Post Call Analytics - PCA (v0.7.6) (uksb-1sn29lk73)
 
 Parameters:
 
@@ -376,18 +376,22 @@ Parameters:
     Type: String
     AllowedValues:
       - 'DISABLED'
+      - 'BEDROCK+TCA'
       - 'BEDROCK'
+      - 'TCA-ONLY'
       - 'SAGEMAKER'
       - 'ANTHROPIC'
       - 'LAMBDA'
     Description: >
       Set to enable call summarization by a Large Language Model. 
+      The BEDROCK+TCA will use Transcribe Call Analytics for summarization and Bedrock for other analytics, and is available only for English.
       The BEDROCK option requires you to choose one of the supported model IDs from the provided list (SummarizationBedrockModelId).
       You must also accept access to that model in the Amazon Bedrock > Model Access console.
+      The TCA-ONLY option will not use Bedrock, but will only use Transcribe Call Analytics summarization, and is available only for English.
       The SAGEMAKER option uses a SageMaker endpoint with the pretrained bart-large-cnn-samsum model with a ml.m5.xlarge instance type.  
       The LAMBDA option requires you to provide a function ARN below. 
       The ANTHROPIC option is a third party service, and you must enter your Anthropic API key in the Third Party LLM API Key section. 
-
+  
   SummarizationBedrockModelId:
     Type: String
     Default: anthropic.claude-instant-v1
@@ -537,11 +541,22 @@ Conditions:
   ShouldCreateKendraIndexDeveloperEdition: !Equals [!Ref EnableTranscriptKendraSearch, 'Yes, create new Kendra Index (Developer Edition)']
   ShouldDeployPcaDashboards: !Equals [!Ref EnablePcaDashboards, 'Yes']
   ShouldLoadSampleFiles: !Equals [!Ref loadSampleAudioFiles, 'true']
-  ShouldDeployBedrockBoto3Layer: !Or [!Equals [!Ref CallSummarization, 'BEDROCK'], !Equals [!Ref GenAIQuery, 'BEDROCK'],]
+  ShouldDeployBedrockBoto3Layer: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'], 
+    !Equals [!Ref CallSummarization, 'BEDROCK+TCA'], 
+    !Equals [!Ref CallSummarization, 'TCA-ONLY'], 
+    !Equals [!Ref GenAIQuery, 'BEDROCK'],
+  ]
   ShouldDeployLLMThirdPartyApiKey: !And [!Not [!Equals [!Ref SummarizationLLMThirdPartyApiKey, '']], !Not [!Equals [!Ref SummarizationLLMThirdPartyApiKey, undefined]]]
-  ShouldTestBedrockModelId: !Or [!Equals [!Ref CallSummarization, 'BEDROCK'], !Equals [!Ref GenAIQuery, 'BEDROCK'],]
+  ShouldTestBedrockModelId: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'], 
+    !Equals [!Ref CallSummarization, "BEDROCK+TCA"],
+    !Equals [!Ref GenAIQuery, 'BEDROCK'],]
   ShouldTestGenAIQueryBedrockModelId: !Equals [!Ref GenAIQuery, 'BEDROCK']
-  ShouldTestSummarizationBedrockModelId: !Equals [!Ref CallSummarization, 'BEDROCK']
+  ShouldTestSummarizationBedrockModelId: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'],
+    !Equals [!Ref CallSummarization, "BEDROCK+TCA"],
+  ]
 
 Resources:
   ########################################################

pca-main.template

@@ -1,6 +1,6 @@
 AWSTemplateFormatVersion: "2010-09-09"
 
-Description: Amazon Transcribe Post Call Analytics - PCA (v0.7.5) (uksb-1sn29lk73)
+Description: Amazon Transcribe Post Call Analytics - PCA (v0.7.6) (uksb-1sn29lk73)
 
 Parameters:
 
@@ -378,14 +378,18 @@ Parameters:
     Type: String
     AllowedValues:
       - 'DISABLED'
+      - 'BEDROCK+TCA'
       - 'BEDROCK'
+      - 'TCA-ONLY'
       - 'SAGEMAKER'
       - 'ANTHROPIC'
       - 'LAMBDA'
     Description: >
       Set to enable call summarization by a Large Language Model. 
+      The BEDROCK+TCA will use Transcribe Call Analytics for summarization and Bedrock for other analytics, and is available only for English.
       The BEDROCK option requires you to choose one of the supported model IDs from the provided list (SummarizationBedrockModelId).
       You must also accept access to that model in the Amazon Bedrock > Model Access console.
+      The TCA-ONLY option will not use Bedrock, but will only use Transcribe Call Analytics summarization, and is available only for English.
       The SAGEMAKER option uses a SageMaker endpoint with the pretrained bart-large-cnn-samsum model with a ml.m5.xlarge instance type.  
       The LAMBDA option requires you to provide a function ARN below. 
       The ANTHROPIC option is a third party service, and you must enter your Anthropic API key in the Third Party LLM API Key section. 
@@ -539,12 +543,22 @@ Conditions:
   ShouldCreateKendraIndexDeveloperEdition: !Equals [!Ref EnableTranscriptKendraSearch, 'Yes, create new Kendra Index (Developer Edition)']
   ShouldDeployPcaDashboards: !Equals [!Ref EnablePcaDashboards, 'Yes']
   ShouldLoadSampleFiles: !Equals [!Ref loadSampleAudioFiles, 'true']
-  ShouldDeployBedrockBoto3Layer: !Or [!Equals [!Ref CallSummarization, 'BEDROCK'], !Equals [!Ref GenAIQuery, 'BEDROCK'],]
+  ShouldDeployBedrockBoto3Layer: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'], 
+    !Equals [!Ref CallSummarization, 'BEDROCK+TCA'], 
+    !Equals [!Ref CallSummarization, 'TCA-ONLY'], 
+    !Equals [!Ref GenAIQuery, 'BEDROCK'],
+  ]
   ShouldDeployLLMThirdPartyApiKey: !And [!Not [!Equals [!Ref SummarizationLLMThirdPartyApiKey, '']], !Not [!Equals [!Ref SummarizationLLMThirdPartyApiKey, undefined]]]
-  ShouldTestBedrockModelId: !Or [!Equals [!Ref CallSummarization, 'BEDROCK'], !Equals [!Ref GenAIQuery, 'BEDROCK'],]
+  ShouldTestBedrockModelId: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'], 
+    !Equals [!Ref CallSummarization, "BEDROCK+TCA"],
+    !Equals [!Ref GenAIQuery, 'BEDROCK'],]
   ShouldTestGenAIQueryBedrockModelId: !Equals [!Ref GenAIQuery, 'BEDROCK']
-  ShouldTestSummarizationBedrockModelId: !Equals [!Ref CallSummarization, 'BEDROCK']
-
+  ShouldTestSummarizationBedrockModelId: !Or [
+    !Equals [!Ref CallSummarization, 'BEDROCK'],
+    !Equals [!Ref CallSummarization, "BEDROCK+TCA"],
+  ]
 
 Resources:
   ########################################################
@@ -911,6 +925,7 @@ Resources:
           - !Ref BulkUploadBucketName
         BulkUploadMaxDripRate: !Ref BulkUploadMaxDripRate
         BulkUploadMaxTranscribeJobs: !Ref BulkUploadMaxTranscribeJobs
+        CallSummarization: !Ref CallSummarization
         ComprehendLanguages: !Ref ComprehendLanguages
         ContentRedactionLanguages: !Ref ContentRedactionLanguages
         ConversationLocation: !Ref ConversationLocation
@@ -977,7 +992,7 @@ Resources:
 
   BedrockBoto3Layer:
     Type: AWS::CloudFormation::Stack
-    Condition: ShouldDeployBedrockBoto3Layer
+    # Condition: ShouldDeployBedrockBoto3Layer
     Properties:
       TemplateURL: pca-boto3-bedrock/template.yaml
 
@@ -1041,6 +1056,7 @@ Resources:
         GenAIQueryBedrockModelId: !Ref GenAIQueryBedrockModelId
         FetchTranscriptArn: !GetAtt PCAServer.Outputs.FetchTranscriptArn
         SummarizerArn: !GetAtt PCAServer.Outputs.SummarizerArn
+        StepFunctionName: !Ref StepFunctionName
         LLMThirdPartyApiKey: !If 
           - ShouldDeployLLMThirdPartyApiKey
           - !Ref LLMThirdPartyApiKeySecret

pca-server/cfn/lib/llm.template

@@ -5,6 +5,7 @@ Description: Amazon Transcribe Post Call Analytics - PCA Server - S3 Trigger
 Transform: AWS::Serverless-2016-10-31
 
 Parameters:
+
   LLMPromptSummaryTemplate:
     Type: String
     Description: >-

pca-server/cfn/lib/trigger.template

@@ -116,7 +116,7 @@ Resources:
     Properties:
       Code:  ../../src/trigger
       Handler: index.handler
-      Runtime: nodejs14.x
+      Runtime: nodejs16.x
       Role: !GetAtt ConfigureBucketRole.Arn
       Environment:
         Variables:

pca-server/cfn/pca-server.template

@@ -9,18 +9,25 @@ Parameters:
     Description: URL for ffmpeg binary distribution tar file download - see https://www.johnvansickle.com/ffmpeg/
 
   CallSummarization:
-    Default: 'DISABLED'
+    Default: 'BEDROCK'
     Type: String
     AllowedValues:
       - 'DISABLED'
-      - 'SAGEMAKER'
+      - 'BEDROCK+TCA'
       - 'BEDROCK'
-      - 'LAMBDA'
+      - 'TCA-ONLY'
+      - 'SAGEMAKER'
       - 'ANTHROPIC'
+      - 'LAMBDA'
     Description: >
-      Set to enable call summarization by a Large Language Model. The SAGEMAKER option uses a SageMaker endpoint with
-      the pretrained bart-large-cnn-samsum model with a ml.m5.xlarge instance type.  The LAMBDA option requires you 
-      to provide a function ARN below. The ANTHROPIC option is a third party service, and you must enter your Anthropic API key below.
+      Set to enable call summarization by a Large Language Model. 
+      The BEDROCK+TCA will use Transcribe Call Analytics for summarization and Bedrock for other analytics. 
+      The BEDROCK option requires you to choose one of the supported model IDs from the provided list (SummarizationBedrockModelId).
+      You must also accept access to that model in the Amazon Bedrock > Model Access console.
+      The TCA-ONLY option will not use Bedrock, but will only use Transcribe Call Analytics summarization.
+      The SAGEMAKER option uses a SageMaker endpoint with the pretrained bart-large-cnn-samsum model with a ml.m5.xlarge instance type.  
+      The LAMBDA option requires you to provide a function ARN below. 
+      The ANTHROPIC option is a third party service, and you must enter your Anthropic API key in the Third Party LLM API Key section. 
 
   SummarizationBedrockModelId:
     Type: String
@@ -68,7 +75,10 @@ Parameters:
 
 Conditions:
   ShouldCreateBoto3Layer: !Equals [!Ref Boto3LayerArn, '']
-  ShouldDeployBedrockSummarizer: !Equals [!Ref CallSummarization, "BEDROCK"]
+  ShouldDeployBedrockSummarizer: !Or [
+    !Equals [!Ref CallSummarization, "BEDROCK"],
+    !Equals [!Ref CallSummarization, "BEDROCK+TCA"],
+  ]
   ShouldDeploySageMakerSummarizer: !Equals [!Ref CallSummarization, "SAGEMAKER"]
   ShouldEnableAnthropicSummarizer: !Equals [!Ref CallSummarization, "ANTHROPIC"]
   ShouldEnableEndOfCallLambdaHookFunction: !Equals [!Ref CallSummarization, "LAMBDA"]

pca-server/src/pca/pca-aws-sf-process-turn-by-turn.py

@@ -79,6 +79,12 @@ def __init__(self, min_sentiment_pos, min_sentiment_neg, custom_entity_endpoint)
         self.simpleEntityMatchingUsed = (self.customEntityEndpointARN == "") and \
                                         (cf.appConfig[cf.CONF_ENTITY_FILE] != "")
 
+    def process_tca_summary(self):
+        if self.api_mode == cf.API_ANALYTICS and \
+           self.asr_output["ConversationCharacteristics"] and \
+           "ContactSummary" in self.asr_output["ConversationCharacteristics"]:
+            self.analytics.contact_summary = self.asr_output["ConversationCharacteristics"]["ContactSummary"]
+
     def generate_sentiment_trend(self, speaker, speaker_num):
         """
         Generates an entry for the "SentimentTrends" block for the given speaker, which is the overall speaker
@@ -1149,6 +1155,9 @@ def parse_transcribe_file(self, sf_event):
         # Update our results data structures, generate JSON results and save them to S3
         self.push_turn_by_turn_results()
 
+        # Update summary structures
+        self.process_tca_summary()
+
         # Write out the JSON data back to our interim S3 location
         json_output, output_filename = self.pca_results.write_results_to_s3(bucket=output_bucket,
                                                                             object_key=sf_event["interimResultsFile"])