v0.4.0

[0.4.0]

⚠️ IMPORTANT NOTICE - SIGNIFICANT CONFIGURATION CHANGES

This release introduces significant changes to the accelerator configuration for defining document classes and attributes. The configuration format has been migrated to JSON Schema standards, which provides enhanced flexibility and validation capabilities.

While automatic migration is provided for backward compatibility, customers MUST fully test this update in a non-production environment before upgrading production systems. We strongly recommend:

1. Deploy the update to a test/development environment first
2. Verify all document processing workflows function as expected
3. Test with representative samples of your production documents
4. Review the migration guide at docs/json-schema-migration.md
5. Only proceed with production upgrade after thorough validation

Do not upgrade production systems without completing validation testing.

Added

• Agent Companion Chat Experience

  • Added comprehensive interactive AI assistant interface providing real-time conversational support for the IDP Accelerator
  • Session-Based Architecture: Transformed from job-based (single request/response) to session-based (multi-turn conversations) with unified agentic chat experience
  • Persistent Chat Memory: DynamoDB-backed conversation history with automatic loading of last 20 turns, turn-based message grouping, and intelligent context management with sliding window optimization
  • Real-Time Streaming: AppSync GraphQL subscriptions enable incremental response streaming with proper async task cleanup and thinking tag removal for clean display
  • Code Intelligence Agent: New specialized agent for code-related assistance with DeepWiki MCP server integration, security guardrails to prevent sensitive data exposure, and user-controlled opt-in toggle (default: enabled)
  • Rich Chat Interface: Modern UI with CloudScape Design System featuring real-time message streaming, multi-agent support (Analytics, Code Intelligence, Error Analyzer, General), Markdown rendering with syntax highlighting, structured data visualization (charts via Chart.js, sortable tables), expandable tool usage sections, sample prompts, and auto-scroll behavior
  • Privacy & Security: Explicit user consent for Code Intelligence third-party services, session isolation with unique session IDs, error boundary protection, input validation

• JSON Schema Format for Class Definitions - docs/json-schema-migration.md

  • Document class definitions now use industry-standard JSON Schema Draft 2020-12 format for improved flexibility and tooling integration
  • Standards-Based Validation: Leverage standard JSON Schema validators and tooling ecosystem for better configuration validation
  • Enhanced Extensibility: Custom IDP properties use standard JSON Schema extension pattern (x-aws-idp-* prefix) for clean separation of concerns
  • Modern Data Contract: Define document structures using widely-adopted JSON Schema format with robust type system (string, number, boolean, object, array)
  • Nested Structure Support: Natural representation of complex documents with nested objects and arrays using JSON Schema's native properties and items keywords
  • Automatic Migration: Existing legacy configurations automatically migrate to JSON Schema format on first load - completely transparent to users
  • Backward Compatible: Legacy format remains supported through automatic migration - no manual configuration updates required
  • Comprehensive Documentation: New migration guide with format comparison, field mapping table, and best practices

• IDP CLI Single Document Status Support with Programmatic Output

  • Enhanced status command to support checking individual document status via new --document-id option as alternative to --batch-id
  • Added programmatic output capabilities with exit codes (0=success, 1=failure, 2=processing) for scripting and automation
  • JSON format output (--format json) provides structured data for parsing in CI/CD pipelines and scripts
  • Live monitoring support with --wait flag works for both batch and single document status checks
  • Mutual exclusion validation ensures only one of --batch-id or --document-id is specified

• Error Analyzer CloudWatch Tool Enhancements

  • Enhanced CloudWatch log filtering with request ID-based filtering for more targeted error analysis
  • Improved XRay tool tracing and logging capabilities for better diagnostic accuracy
  • Enhanced error context correlation between CloudWatch logs and X-Ray traces
  • Consolidated and renamed tools
  • Provided tools access to agent
  • Updated system prompt

• Error Analyzer CloudWatch Tool Enhancements

  • Enhanced CloudWatch log filtering with request ID-based filtering for more targeted error analysis
  • Improved XRay tool tracing and logging capabilities for better diagnostic accuracy
  • Enhanced error context correlation between CloudWatch logs and X-Ray traces
  • Consolidated and renamed tools
  • Provided tools access to agent
  • Updated system prompt

Fixed

• UI Robustness for Orphaned List Entries - #102
  • Fixed UI error banner "failed to get document details - please try again later" appearing when orphaned list entries exist (list# items without corresponding doc# items in DynamoDB tracking table)
  • Root Cause: When a document had a list entry but no corresponding document record, the error would trigger UI banner and prevent display of all documents in the same time shard
  • Solution: Enhanced error handling to gracefully handle missing documents - now only shows error banner if ALL documents fail to load, not just one
  • Enhanced Debugging: Added detailed console logging with full PK/SK information for both list entries and expected document entries to facilitate cleanup of orphaned records
  • User Impact: All valid documents now display correctly even when orphaned list entries exist; debugging information available in browser console for identifying problematic entries

v0.3.21

[0.3.21]

Added

• Claude Sonnet 4.5 Haiku Model Support

  • Added support for Claude Haiku 4.5
  • Available for configuration across all document processing steps

• X-Ray Integration for Error Analyzer Agent

  • Integrated AWS X-Ray tracing tools to enhance diagnostic capabilities of the error analyzer agent
  • X-Ray context enables better distinction between infrastructure issues and application logic failures
  • Added trace ID persistence in DynamoDB alongside document status for complete traceability
  • Enhanced CloudWatch error log filtering for more targeted error analysis
  • Simplified CloudWatch results structure for improved readability and analysis
  • Updated error analyzer recommendations to leverage X-Ray insights for more accurate root cause identification

• EU Region Support with Automatic Model Mapping

  • Added support for deploying the solution in EU regions (eu-central-1, eu-west-1, etc.)
  • Automatic model endpoint mapping between US and EU regions for seamless deployment
  • Comprehensive model mapping table covering Amazon Nova and Anthropic Claude models
  • Intelligent fallback mappings when direct EU equivalents are unavailable
  • Quick Launch button for eu-central-1 region in README and deployment documentation
  • IDP CLI now supports eu-central-1 deployment with automatic template URL selection
  • Complete technical documentation in docs/eu-region-model-support.md with best practices and troubleshooting

Changed

• Migrated Evaluation from EventBridge Trigger to Step Functions Workflow

  • Moved evaluation processing from external EventBridge-triggered Lambda to integrated Step Functions workflow step
  • Race Condition Eliminated: Evaluation now runs inside state machine before WorkflowTracker marks documents COMPLETE, preventing premature completion status when evaluation is still running
  • Config-Driven Control: Evaluation now controlled by evaluation.enabled configuration setting instead of CloudFormation stack parameter, enabling runtime control without stack redeployment
  • Enhanced Status Tracking: Added EVALUATING status to document processing pipeline for better visibility of evaluation progress
  • UI Improvements: Added support for displaying EVALUATING status in processing flow viewer and "NOT ENABLED" badge when evaluation is disabled in configuration
  • Consistent Pattern: Aligns evaluation with summarization and assessment patterns for unified feature control approach

• Migrated UI Build System from Create React App to Vite

  • Upgraded to Vite 7 for faster build times
  • Updated to React 18, AWS Amplify v6, react-router-dom v6, and Cloudscape Design System
  • Reduced dependencies and node_modules size
  • Implemented strategic code splitting for improved performance
  • Environment variables now use VITE_ prefix instead of REACT_APP_ for local development

Fixed

• IDP CLI Code Cleanup and Portability Improvements - #91, #92
  • Removed dead code from previous refactors in batch_processor.py (51 lines)
  • Replaced hardcoded absolute paths with dynamic path resolution in rerun_processor.py for cross-platform compatibility

v0.3.20

[0.3.20]

Added

• Agentic extraction preview with Strands agents (experimental) introducing intelligent, self-correcting document extraction with improved schema compliance and accuracy improvements over traditional methods.

  • Leverages the Strands Agent framework with iterative validation loops and automatic error correction to deliver schema compliance
  • Provides structured output through Pydantic models with built-in validators, automatic retry handling, and superior handling of complex nested structures and date standardization
  • Includes sample notebooks and configuration assets demonstrating agentic extraction for Pattern-2 lending documents
  • Programmatic access available via structured_output function in lib/idp_common_pkg/idp_common/extraction/agentic_idp.py
  • Currently this is an experimental feature. Future extensibility includes UI-based validation customization, code generation, and Model Context Protocol (MCP) integration for external data enrichment during extraction

• IDP CLI - Command Line Interface for Batch Document Processing

  • Added CLI tool (idp_cli/) for programmatic batch document processing and stack management
  • Key Features: Deploy/update/delete CloudFormation stacks, process and reprocess documents from local directories or S3 URIs, live progress monitoring with rich terminal UI, download processing results locally, validate manifests before processing, generate manifests from directories with automatic baseline matching
  • Selective Reprocessing: New rerun-inference command to reprocess documents from specific pipeline steps (classification or extraction) while leveraging existing OCR data for cost/time optimization
  • Evaluation Framework: Workflow for accuracy testing including initial processing, manual validation, baseline creation, and automated evaluation with detailed metrics
  • Analytics Integration: Query aggregated results via Athena SQL or use Agent Analytics in Web UI for visual analysis
  • Use Cases: Rapid configuration iteration, large-scale batch processing, CI/CD integration, automated accuracy testing, automated environment cleanup, prompt engineering experiments
  • Documentation: README with Quick Start, Commands Reference, Evaluation Workflow, and troubleshooting guides

• Extraction Results Integration in Summarization Service

  • Integrates extraction results from the extraction service into summarization module for context-aware summaries
  • Features: Fully backward compatible (works with or without extraction results), automatic section handling, error resilient with graceful continuation, comprehensive logging
  • Configuration: Enable by adding {EXTRACTION_RESULTS} placeholder to task_prompt in config.yaml
  • Benefits: Context-aware summaries referencing extracted values, improved accuracy and quality, better extraction-summary alignment

Changed

• Containerized Pattern-2 deployment pipeline that builds and pushes all Lambda images via CodeBuild using the new Dockerfile, plus automated ECR cleanup and tests.
  • Lambda docker image deployments have a 10 GB image size limit compared to the 250 MB zip limit of regular deployment. This however doesn't allow for viewing the code in the AWS console.
    The change was introduced to accommodate the increased package size of introducing Strands into the package dependencies.

Fixed

• Discovery function times out when processing large documents.
  • increase lambda discovery processor timeout to 900s
• Corrected baseline directory structure documentation in evaluation.md
  • Fixed incorrect baseline structure showing flat .json files instead of proper directory hierarchy
  • Updated to correct structure: <document-name>/sections/1/result.json
  • Reorganized document for better logical flow and user experience
• GovCloud Template Generation - Removed GraphQLApi References - #82
  • Fixed invalid GovCloud template generation where ProcessChanges AppSync resources were not being removed, causing "Fn::GetAtt references undefined resource GraphQLApi" errors
  • Updated scripts/generate_govcloud_template.py to remove all ProcessChanges-related resources and extend AppSync parameter cleanup to all pattern stacks
  • Fixed InvalidClientTokenId validation error by ensuring CloudFormation client uses the correct region when validating templates (commercial vs GovCloud)
• Enhanced Processing Flow Visualization for Disabled Steps
  • Fixed UX issue where disabled processing steps (when summarization.enabled: false or assessment.enabled: false in configuration) appeared visually identical to active steps in the "View Processing Flow" display
  • Key Benefit: Users can now immediately see which steps are actually processing data vs. steps that execute but skip processing based on configuration settings, preventing confusion about whether summarization or assessment ran
  • Limitation: the new visual indicators are driven from the current config, which may have been altered since the document was processed. We will address this in a later release. See Issue #86.

Known Issues

• GovCloud Deployments fail, due to lack of ARM support for CodeBuild. Fix targeted for next release.

v0.3.19

[0.3.19]

Added

• Error Analyzer (Troubleshooting Tool) for AI-Powered Failure Diagnosis

  • Introduced intelligent AI-powered troubleshooting agent that automatically diagnoses document processing failures using Claude Sonnet 4 with the Strands agent framework
  • Key Capabilities: Natural language query interface, intelligent routing between document-specific and system-wide analysis, multi-source data correlation (CloudWatch Logs, DynamoDB, Step Functions), root cause identification with actionable recommendations, evidence-based analysis with collapsible log details
  • Web UI Integration: Accessible via "Troubleshoot" button on failed documents with real-time job status, progress tracking, automatic job resumption, and formatted results (Root Cause, Recommendations, Evidence sections)
  • Tool Ecosystem: 8 specialized tools including analyze_errors (main router), analyze_document_failure, analyze_recent_system_errors, CloudWatch log search tools, DynamoDB integration tools, and Lambda context retrieval - additional tools will be added as the feature evolves.
  • Configuration: Configurable via Web UI including model selection (Claude Sonnet 4 recommended), system prompt customization, max_log_events (default: 5), and time_range_hours_default (default: 24)
  • Documentation: Comprehensive guide in docs/error-analyzer.md with architecture diagrams, usage examples, best practices, troubleshooting guide.

• Claude Sonnet 4.5 Model Support

  • Added support for Claude Sonnet 4.5 and Claude Sonnet 4.5 - Long Context models
  • Available for configuration across all document processing steps

Fixed

• Problem with setting correctly formatted WAF IPv4 CIDR range - #73

• Duplicate Step Functions Executions on Document Reprocess - GitHub Issue #66

  • Eliminated duplicate workflow executions when reprocessing large documents (>40MB, 500+ pages)
  • Root Cause: S3 copy_object operations were triggering multiple "Object Created" events for large files, causing queue_sender to create duplicate document entries and workflow executions
  • Solution: Refactored reprocess_document_resolver to directly create fresh Document objects and queue to SQS, completely bypassing S3 event notifications
  • Benefits: Eliminates unnecessary S3 copy operations (cost savings)

v0.3.18

Added

• Lambda Function Execution Cost Metering for Complete Cost Visibility
  • Added Lambda execution cost tracking to all core processing functions across all three processing patterns
  • Dual Metrics: Tracks both invocation counts ($0.20 per 1M requests) and GB-seconds duration ($16.67 per 1M GB-seconds) aligned with official AWS Lambda pricing
  • Context-Specific Tracking: Separate cost attribution for each processing step enabling granular cost analysis per document processing context
  • Automatic Integration: Lambda costs automatically integrate with existing cost reporting infrastructure and appear alongside AWS service costs (Textract, Bedrock, SageMaker)
  • Configuration Integration: Added Lambda pricing entries to all 7 configuration files in config_library/ using official US East pricing

Fixed

• Defect in v0.3.17 causing workflow tracker failure to (1) update status of failed workflows, and (2) update reporting database for all workflows #72

v0.3.17

Added

• Edit Sections Feature for Modifying Class/Type and Reprocessing Extraction

  • Added Edit Sections interface for Pattern-2 and Pattern-3 workflows with reprocessing optimization
  • Key Features: Section management (create, update, delete), classification updates, page reassignment with overlap detection, real-time validation
  • Selective Reprocessing: Only modified sections are reprocessed while preserving existing data for unmodified sections
  • Processing Pipeline: All functions (OCR/Classification/Extraction/Assessment) automatically skip redundant operations based on data presence
  • Pattern Compatibility: Full functionality for Pattern-2/Pattern-3, informative modal for Pattern-1 explaining BDA not yet supported

• Analytics Agent Schema Optimization for Improved Performance

  • Embedded Database Overview: Complete table listing and guidance embedded directly in system prompt (no tool call needed)
  • On-Demand Detailed Schemas: get_table_info(['specific_tables']) loads detailed column information only for tables actually needed by the query
  • Significant Performance Gains: Eliminates redundant tool calls on every query while maintaining token efficiency
  • Enhanced SQL Guidance: Comprehensive Athena/Trino function reference with explicit PostgreSQL operator warnings to prevent common query failures like ~ regex operator mistakes
  • Faster Time-to-Query: Agent has immediate access to table overview and can proceed directly to detailed schema loading for relevant tables

Fixed

• Fix missing data in Glue tables when using a document class that contains a dash (-).

v0.3.16

[0.3.16]

Added

• S3 Vectors Support for Cost-Optimized Knowledge Base Storage

  • Added S3 Vectors as alternative vector store option to OpenSearch Serverless for Bedrock Knowledge Base with lower storage costs
  • Custom resource Lambda implementation for S3 vector bucket and index management (using boto3 s3vectors client) with proper IAM permissions and resource cleanup
  • Unified Knowledge Base interface supporting both vector store types with automatic resource provisioning based on user selection

• Page Limit Configuration for Classification Control

  • Added maxPagesForClassification configuration option to control how many pages are used during document classification
  • Default Behavior: "ALL" - uses all pages for classification (existing behavior)
  • Limited Page Classification: Set to numeric value (e.g., "1", "2", "3") to classify only the first N pages
  • Important: When using numeric limit, the classification result from the first N pages is applied to ALL pages in the document, effectively forcing the entire document to be assigned a single class with one section
  • Use Cases: Performance optimization for large documents, cost reduction for documents with consistent classification patterns, simplified processing for homogeneous document types

• CloudFormation Service Role for Delegated Deployment Access

  • Added example CloudFormation service role template that enables non-administrator users to deploy and maintain IDP stacks without requiring ongoing administrator permissions
  • Administrators can provision the service role once with elevated privileges, then delegate deployment capabilities to developer/DevOps teams
  • Includes comprehensive documentation and cross-referenced deployment guides explaining the security model and setup process

Fixed

• Fixed issue where CloudFront policy statements were still appearing in generated GovCloud templates despite CloudFront resources being removed
• Fix duplicate Glue tables are created when using a document class that contains a dash (-). Resolved by replacing dash in section types with underscore character when creating the table, to align with the table name generated later by the Glue crawler - resolves #57.
• Fix occasional UI error 'Failed to get document details - please try again later' - resolves #58
• Fixed UI zipfile creation to exclude .aws-sam directories and .env files from deployment package
• Added security recommendation to set LogLevel parameter to WARN or ERROR (not INFO) for production deployments to prevent logging of sensitive information including PII data, document contents, and S3 presigned URLs
• Hardened several aspects of the new Discovery feature

v0.3.15

[0.3.15]

Added

• Intelligent Document Discovery Module for Automated Configuration Generation

  • Added Discovery module that automatically analyzes document samples to identify structure, field types, and organizational patterns
  • Pattern-Neutral Design: Works across all processing patterns (1, 2, 3) with unified discovery process and pattern-specific implementations
  • Dual Discovery Methods: Discovery without ground truth (exploratory analysis) and with ground truth (optimization using labeled data)
  • Automated Blueprint Creation: Pattern 1 includes zero-touch BDA blueprint generation with intelligent change detection and version management
  • Web UI Integration: Real-time discovery job monitoring, interactive results review, and seamless configuration integration
  • Advanced Features: Multi-model support (Nova, Claude), customizable prompts, configurable parameters, ground truth processing, schema conversion, and lifecycle management
  • Key Benefits: Rapid new document type onboarding, reduced time-to-production, configuration optimization, and automated workflow bootstrapping
  • Use Cases: New document exploration, configuration improvement, rapid prototyping, and document understanding
  • Documentation: Guide in docs/discovery.md with architecture details, best practices, and troubleshooting

• Optional Pattern-2 Regex-Based Classification for Enhanced Performance

  • Added support for optional regex patterns in document class definitions for performance optimization
  • Document Name Regex: Match against document ID/name to classify all pages without LLM processing when all pages should be the same class
  • Document Page Content Regex: Match against page text content during multi-modal page-level classification for fast page classification
  • Key Benefits: Significant performance improvements and cost savings by bypassing LLM calls for pattern-matched documents, deterministic classification results for known document patterns, seamless fallback to existing LLM classification when regex patterns don't match
  • Configuration: Optional document_name_regex and document_page_content_regex fields in class definitions with automatic regex compilation and validation
  • Logging: Comprehensive info-level logging when regex patterns match for observability and debugging
  • CloudFormation Integration: Updated Pattern-2 schema to support regex configuration through the Web UI
  • Demonstration: New step2_classification_with_regex.ipynb notebook showcasing regex configuration and performance comparisons
  • Documentation: Enhanced classification module README and main documentation with regex usage examples and best practices

• Windows WSL Development Environment Setup Guide

  • Added WSL-based development environment setup guide for Windows developers in docs/setup-development-env-WSL.md
  • Key Features: Automated setup script (wsl_setup.sh) for quick installation of Git, Python, Node.js, AWS CLI, and SAM CLI
  • Integrated Workflow: Development setup combining Windows tools (VS Code, browsers) with native Linux environment
  • Target Use Cases: Windows developers needing Linux compatibility without Docker Desktop or VM overhead

Fixed

• Throttling Error Detection and Retry Logic for Assessment Functions - GitHub Issue #45

  • Assessment Function: Enhanced throttling detection to check for throttling errors returned in document.errors field in addition to thrown exceptions, raising ThrottlingException to trigger Step Functions retry when throttling is detected
  • Granular Assessment Task Caching: Fixed caching logic to properly cache successful assessment tasks when there are ANY failed tasks (both exception-based and result-based failures), enabling efficient retry optimization by only reprocessing failed tasks while preserving successful results
  • Impact: Improved resilience for throttling scenarios, reduced redundant processing during retries, and better Step Functions retry behavior

• Security Vulnerability Mitigation - Package Updates

• GovCloud Compatibility - Hardcoded Service Domain References

  • Fixed hardcoded amazonaws.com references in CloudFormation templates that prevented GovCloud deployment
  • Updated all service principals and endpoints to use dynamic ${AWS::URLSuffix} expressions for automatic region-based resolution
  • Templates Updated: template.yaml (main template), patterns/pattern-3/sagemaker_classifier_endpoint.yaml
  • Services Fixed: EventBridge, Cognito, SageMaker, ECR, CloudFront, CodeBuild, AppSync, Lambda, DynamoDB, CloudWatch Logs, Glue
  • Resolves GitHub Issue #50 - templates now deploy correctly in both standard AWS and GovCloud regions

• Bug Fixes and Code Improvements

  • Fixed HITL processing errors in both Pattern-1 (DynamoDB validation with empty strings) and Pattern-2 (string indices error in A2I output processing)
  • Fixed Step Function UI issues including auto-refresh button auto-disable and fetch failures for failed executions with datetime serialization errors
  • Cleaned up unused Step Function subscription infrastructure and removed duplicate code in Pattern-2 HITL function
  • Expanded UI Visual Editor bounding box size with padding for better visibility and user interaction
  • Fixed bug in list of models supporting cache points - previously claude 4 sonnet and opus had been excluded.
  • Validations added at the assessment step for checking valid json response. The validation fails after extraction/assessment is complete if json parsing issues are encountered.

v0.3.14

[0.3.14]

Added

• Support for 1m token context for Claude Sonnet 4
• Video demo of "Chat with Document" in ./docs/web-ui.md
• Human-in-the-Loop (HITL) Support Extended to Pattern-2
  • Added HITL review capabilities for Pattern-2 (Textract + Bedrock processing) using Amazon SageMaker Augmented AI (A2I)
  • Enables human validation and correction when extraction confidence falls below configurable threshold
  • Includes same features as Pattern-1 HITL: automatic triggering, review portal integration, and seamless result updates
  • Documentation and video demo in ./docs/human-review.md

Removed

• Windows development environment guide and setup script removed as it proved insufficiently robust

Fixed

• Fix 1-click Launch URL output from the GovCloud template generation script
• Add Agent Analytics to architecture diagram
• Fix various UX and error reporting issues with the new Python publish script
• Simplify UDOP model path construction and avoid invalid default for regions other than us-east-1 and us-west-2
• Permission regression from previous release affecting "Chat with Document"

v0.3.13

[0.3.13]

Added

• External MCP Agent Integration for Custom Tool Extension

  • Added External MCP (Model Context Protocol) Agent support that enables integration with custom MCP servers to extend IDP capabilities
  • Cross-Account Integration: Host MCP servers in separate AWS accounts or external infrastructure with secure OAuth authentication using AWS Cognito
  • Dynamic Tool Discovery: Automatically discovers and integrates available tools from MCP servers through the IDP web interface
  • Secure Authentication Flow: Uses AWS Cognito User Pools for OAuth bearer token authentication with proper token validation
  • Configuration Management: JSON array configuration in AWS Secrets Manager supporting multiple MCP server connections with optional custom agent names and descriptions
  • Real-time Integration: Tools become immediately available through the IDP web interface after configuration

• AWS GovCloud Support with Automated Template Generation

  • Added GovCloud compatibility through scripts/generate_govcloud_template.py script
  • ARN Partition Compatibility: All templates updated to use arn:${AWS::Partition}: for both commercial and GovCloud regions
  • Headless Operation: Automatically removes UI-related resources (CloudFront, AppSync, Cognito, WAF) for GovCloud deployment
  • Core Functionality Preserved: All 3 processing patterns and complete 6-step pipeline (OCR, Classification, Extraction, Assessment, Summarization, Evaluation) remain fully functional
  • Automated Workflow: Single script orchestrates build + GovCloud template generation + S3 upload with deployment URLs
  • Enterprise Ready: Enables headless document processing for government and enterprise environments requiring GovCloud compliance
  • Documentation: New docs/govcloud-deployment.md with deployment guide, architecture differences, and access methods

• Pattern-2 and Pattern-3 Assessment now generate geometry (bounding boxes) for visualization in UI 'Visual Editor' (parity with Pattern-1)

  • Added comprehensive spatial localization capabilities to both regular and granular assessment services
  • Automatic Processing: When LLM provides bbox coordinates, automatically converts to UI-compatible (Visual Edit) geometry format without any configuration
  • Universal Support: Works with all attribute types - simple attributes, nested group attributes (e.g., CompanyAddress.State), and list attributes
  • Enhanced Prompts: Updated assessment task prompts with spatial-localization-guidelines requesting bbox coordinates in normalized 0-1000 scale
  • Demo Notebooks: Assessment notebooks now showcase automatic bounding box processing

• New Python-Based Publishing System

  • Replaced publish.sh bash script with new publish.py Python script
  • Rich console interface with progress bars, spinners, and colored output using Rich library
  • Multi-threaded artifact building and uploading for significantly improved performance
  • Native support for Linux, macOS, and Windows environments

• Windows Development Environment Setup Guide and Helper Script

  • New scripts/dev_setup.bat (570 lines) for complete Windows development environment configuration

• OCR Service Default Image Sizing for Resource Optimization

  • Implemented automatic default image size limits (951×1268) when no image sizing configuration is provided
  • Key Benefits: Reduction in vision model token consumption, prevents OutOfMemory errors during concurrent processing, improves processing speed and reduces bandwidth usage

Changed

• Reverted to python3.12 runtime to resolve build package dependency problems

Fixed

• Improved Visual Edit bounding box position when using image zoom or pan