This repository contains a CDK project that deploys a fully automated language analysis pipeline.
The application performs a language analysis of the loaded documents divided into the following categories:
This part of the analysis calculates:
- Rate of use of foreign words
- Most frequently occurring foreign words
This one is calculated in base of different parameters:
- TTR: text-to-word ratio: ratio of distinct words in a document to the total number of words in a document.
- MLTD: measurement of textual lexical diversity.
- Amount of words in each document
- Most repeated nouns, adjectives, adverbs and verbs.
The following type of errors are calculated:
- Typographical / punctuation
- Grammatical
- Writing style
- Inconsistency
- Spelling
Depending on your needs and use case, there are two variants of the application that you can deploy: the first one indexes documents in Amazon OpenSearch service to offer text-based search. The second bypasses Amazon OpenSearch service and makes use of Amazon QuickSight for displaying the results.
- Automated language analysis with text search capabilities
- Automated language analysis with QuickSight integration (under development)
Select the variant that best suits your use case to view its documentation and deployment instructions.
The application is deployed to your AWS account via AWS CDK, so you will need to install it in order to continue with the deployment. Execute the following command to install CDK:
npm install -g aws-cdk
You can read more about getting started with CDK by reading the AWS Cloud Development Kit developer guide.
See CONTRIBUTING for more information.
This project is licensed under the Apache-2.0 License.