Updated save options for ocr_outputs_with_words

Local text redaction now produces ocr results with words json and can make dataframe format

Further updates to line level duplicate identification

Updated packages. Corrected CSV logger headings, can now submit custom log csv names to S3. Started work on identifying and deduplicating at the line level

Minor update to ensure whole page redactions are applied correctly to a document with existing redactions

Updated duplicate pages interface to include subdocuments and review. Updated relevant user guide. Minor package updates

Updated duplicate pages functionality. Improve redaction efficiency a little with concat method. Minor modification to documentation and interface

Added folder with CDK code and app. Updated config py file to be compatible with all temp folders needed for read only file systems

Added possibility of changing model and entity types in config file

Added capability to redact all redactions with the same text based on the selected row. Rerranged buttons on review page a little. Improved page navigation efficiency.

Now xfdf Adobe exports can export redacted text that is searchable in Acrobat

Expanded checks for out of range page cropboxes

Updated gradio version. Minor changes to redactor function sequence. Minor formatting and wording changes.

Added config options for compressing output pdfs, returning output redacted pdfs at all, and for changing the length of time for showing previous Textract jobs

More checks on ocr outputs in redaction functions

Corrected a couple of bugs. Now Textract whole document API call outputs will load also the input PDF into the app

Updated logging format for timestamps to be compatible with AWS. Added load_dynamo_logs.py example file.

Fix for image file redaction

Minor changes for cost codes, package updates. Added pyproject.toml file

Now local OCR outputs can be saved to file and reloaded to save preparation time. Bug fixing in logs and tabular data redaction. Update to documentation

Improved logging format a little. Now possible to save logs to DynamoDB

Improved efficiency of review page navigation, especially for large documents. Updated user guide

Updated Textract logging

Added button to convert Textract API outputs to ocr_output files easily. Corrected Textract job file location

Added compatibility with gradio_image_annotation for passing through id and text properties to annotator. Corrected csv location for Textract api calls. Other minor changes

Minor function documentation changes. Requirements update for new Gradio and version of Gradio annotator that allows for saving preferred redaction format and to include box id

Corrected RUN_AWS_FUNCTIONS environment variable reference when downloading cost codes

Made changes to hopefully resolve issue with downloading cost centre details from S3 to container

Fixed issue where S3 cost codes are defined but not local cost code location

Fixed issue in Docker containers built locally without correct folder permissions. Improved config file. Updated Gradio version to fix issue with selecting filtered rows. Minor bug fixes.

Implemented Textract document API calls and associated output tracking/download. Fixes to config and cost code implementation. General minor bug fixes.

Added workaround to issue with selectdata and dataframes for filtered dataframes. Rearranged some components.

Corrected issue where the cropbox method was being overwritten in the review code

Corrected set_cropbox in redaction function. Reset cost code selection to correct method.

Cost code dataframe should now pass over selected cost code correctly

Modified config entries to not assume allow list or cost codes file exists. Reduced concurrency to 3 and put input and output files in user subfolders by default

Fixed s3 load for allow list and cost codes

Major update. General code revision. Improved config variables. Dataframe based review frame now includes text, items can be searched and excluded. Costs now estimated. Option for adding cost codes added. Option to extract text only.

Fixed manual entry for allow, deny, and full page redaction lists

More config options. Fixed some bugs with removing elements from review page and Adobe export. Some UI rearrangements

Added features to review dataframe to filter and exclude features based on text. Text should now appear consistently in review_df (for boxes not modified). Larger spacy model returned to use. Gradio upgrade.

Now redact on whole PDF mediabox size (larger than viewable size sometimes), then converted back to cropbox size for print and Adobe review. Improved some error raising and app flow

Integrated AWS Comprehend and fuzzy matching functions with tabular data redaction.

Allowed for output files to be saved into user-specific folders. Added deny list capability to xlsx/csv file redaction

Allowed for Textract and Comprehend API calls through AWS keys. File preparation function incorporated into main redaction function to avoid needing user to 'check in' during redaction process

Fixed issues with log file list picking up logs from other file runs. Updated packages.

Added concurrency limit to run options. Trying again to load in zoom/rotate options from gradio_image_annotator fork.

Laid groundwork for passing in AWS API keys. Duplicate pages option should now work for pages with no text.

App now correctly updates custom fuzzy recognisers

Fixed issues with gradio version 5.16. Fixed fuzzy search error with pages with no data.