You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG.md
+14Lines changed: 14 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,6 +7,15 @@ SPDX-License-Identifier: MIT-0
7
7
8
8
### Added
9
9
10
+
## [0.3.8]
11
+
12
+
### Added
13
+
14
+
-**Lending Package Configuration Support for Pattern-2**
15
+
- Added new `lending-package-sample` configuration to Pattern-2, providing comprehensive support for lending and financial document processing workflows
16
+
- New default configuration for Pattern-2 stack deployments, optimized for loan applications, mortgage processing, and financial verification documents
17
+
- Previous `rvl-cdip-sample` configuration remains available by selecting `rvl-cdip-package-sample` for the `Pattern2Configuration` parameter when deploying or updating stacks
18
+
10
19
-**Text Confidence View for Document Pages**
11
20
- Added support for displaying OCR text confidence data through new `TextConfidenceUri` field
12
21
- New "Text Confidence View" option in the UI pages panel alongside existing Markdown and Text views
- Configurable through Web UI without requiring code changes or redeployment
21
30
22
31
### Changed
32
+
23
33
-**Converted text confidence data format from JSON to markdown table for improved readability and reduced token usage**
24
34
- Removed unnecessary "page_count" field
25
35
- Changed "text_blocks" array to "text" field containing a markdown table with Text and Confidence columns
26
36
- Reduces prompt size for assessment service while improving UI readability
27
37
- OCR confidence values now rounded to 1 decimal point (e.g., 99.1, 87.3) for cleaner display
28
38
- Markdown table headers now explicitly left-aligned using `|:-----|:-----------|` format for consistent appearance
39
+
29
40
-**Simplified OCR Service Initialization**
30
41
- OCR service now accepts a single `config` dictionary parameter for cleaner, more consistent API
31
42
- Aligned with classification service pattern for better consistency across IDP services
32
43
- Backward compatibility maintained - old parameter pattern still supported with deprecation warning
33
44
- Updated all lambda functions and notebooks to use new simplified pattern
34
45
- Removed fixed image target_height and target_width from default configurations, so images are processed in original resolution by default.
46
+
35
47
-**Updated Default Configuration for Pattern1 and Pattern2**
36
48
- Changed default configuration for new stacks from "default" to "lending-package-sample" for both Pattern1 and Pattern2
37
49
- Maintains backward compatibility for stack updates by keeping the parameter value "default" mapped to the rvl-cdip-sample for pattern-2.
50
+
38
51
-**Reduce assessment step costs**
39
52
- Default model for granular assessment is now `us.amazon.nova-pro-v1:0`
40
53
- Improved placement of <<CACHEPOINT>> tags in assessment prompt to improve utilization of prompt caching
41
54
42
55
### Fixed
56
+
43
57
-**Fixed Image Resizing Behavior for High-Resolution Documents**
44
58
- Fixed issue where empty strings in image configuration were incorrectly resizing images to default 951x1268 pixels instead of preserving original resolution
45
59
- Empty strings (`""`) in `target_width` and `target_height` configuration now preserve original document resolution for maximum processing accuracy
Copy file name to clipboardExpand all lines: config_library/pattern-2/lending-package-sample/README.md
+7-14Lines changed: 7 additions & 14 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -11,9 +11,9 @@ This directory contains the default-lending configuration for the GenAI IDP Acce
11
11
12
12
## Validation Level
13
13
14
-
**Level**: 2 - Comprehensive Testing
14
+
**Level**: 2 - Minimal Testing
15
15
16
-
-**Testing Evidence**: This configuration has been tested with lending sample document including payslips, driver's licenses, bank statements, checks, W2 forms, and insurance applications. It demonstrates robust performance in classifying and extracting detailed financial information from standard lending documents.
16
+
-**Testing Evidence**: This configuration has been lightly tested with lending sample document including payslips, driver's licenses, bank statements, checks, W2 forms, and insurance applications.
17
17
-**Known Limitations**: Performance may vary with non-standard document formats, heavily redacted financial documents, or documents with poor image quality that affect OCR accuracy.
18
18
19
19
## Overview
@@ -44,11 +44,10 @@ The configuration defines 6 specialized lending document classes, each with comp
44
44
### Classification Settings
45
45
46
46
-**Model**: Amazon Nova Pro
47
-
-**Method**: Text-based holistic classification
47
+
-**Method**: Multimodal Page Level Classification
48
48
-**Temperature**: 0 (deterministic outputs)
49
49
-**Top-k**: 5
50
50
-**OCR Backend**: Amazon Textract with LAYOUT, TABLES, and SIGNATURES features
51
-
-**OCR Model**: Amazon Claude 3.7 Sonnet for enhanced text extraction and layout understanding
52
51
53
52
The classification component analyzes document content and structure to accurately identify lending document types and establish proper page boundaries within multi-document packages.
54
53
@@ -68,7 +67,7 @@ The extraction component performs comprehensive attribute extraction tailored to
68
67
69
68
### Assessment Settings
70
69
71
-
-**Model**: Amazon Claude 3.7 Sonnet
70
+
-**Model**: us.amazon.nova-pro-v1:0
72
71
-**Granular Assessment**: Enabled with parallel processing
73
72
-**Default Confidence Threshold**: 0.9
74
73
-**Max Workers**: 20 for improved performance
@@ -77,7 +76,7 @@ Enhanced confidence assessment ensures high accuracy for financial data extracti
77
76
78
77
### Summarization Settings
79
78
80
-
-**Model**: Amazon Claude 3.7 Sonnet
79
+
-**Model**: us.amazon.nova-premier-v1:0
81
80
-**Temperature**: 0 (deterministic outputs)
82
81
-**Top-k**: 5
83
82
@@ -97,20 +96,14 @@ This configuration is optimized for processing lending document packages that ty
97
96
98
97
To use this default-lending configuration:
99
98
100
-
1.**Direct Deployment**: Deploy the GenAI IDP Accelerator with this configuration for lending document processing workflows:
0 commit comments