Google Cloud App with computer vision

Goal To create a google cloud based web application module that will identify document content based on a document template. Once the document is identified, the application must be able to transfer the document to a corresponding document folder. Analysis and Resolution For this web application, all it needs is to classify documents and transfer them to a corresponding document folder once identified. We currently have 54 kinds of source documents and we need to classify and segregate all of them so that each kind will have it's own folder in google cloud. The application will run regularly to check the main repository in google cloud storage and once it detects that a file is available in it, it should compare the file to 54 source document templates and it should transfer the file to it's destination folder after processing. Requirements Any programming language and approach would do as long as it can get the input files and post the processed/identified file to the corresponding file folder. Developers can use google vision api for the code or any vision library as long as the app works in google cloud as a service or api. This cloud web api service is scheduled out of user input. Developer can choose to have scheduling for this. The end goal is that any file available in google cloud main repository must be processed by this module. This module is independent and will only connect to GCS for getting input file and posting output files back to correct GCS sub folders. Assumptions Varal will provide developer access to its google cloud platform console. The developer will deploy the web application as a web service api in services section of Google App Engine of the main project. During development, he can create separate project where he can test to avoid possible conflicts, if proven to be needed. After development and testing, he should be able to migrate the application to the main project and make necessary configurations in the main project if he created another project to develop. The unnecessary project should be disabled to avoid bills. As end product, the application should be running as a backend service in the app engine section. The application’s concern is solely for processing the input files in the main google cloud storage bucket then transferring the identified files to the correct sub folders.