I believe you can give InvoicesIndia package a try, but since your fields are too specific, you can do a labelling session in Data Manager for your documents and retrain this model for your documents. If you would like to build fields and train model only on your documents, you can go for the DocumentUnderstanding package via the same Data Manager labeling approach.
Since there is no boolean based datatype in data manager fields, you can skip adding and labelling the signature fields here. When you configure extractors, you can go for form extractor for signature field and your ML extractor for other fields.