How to create a General Document Processor project
Create a data program
- Go to https://super.ai/store/templates/general-document-processor and create a new data project by clicking "try it out".
- Choose your workers for human-in-the-loop (HITL) task for both quality assurance and to train the AI models.
Upon initial setup, always choose "me, or people that I invite" to process the data.
- Upload a file via an option that suits you best.
Upon initial setup, only 1 file can be uploaded to the system - after the setup, multiple documents are allowed to upload.
Choose Document Type
Choose desired Document Type for the list. In case your document type is not on the list, you want to choose “custom” and click "add document" to define fields to extract and their data types.
"Fields" refer to keys in key-value pairs, so the category of information that you want to extract.
You can edit pre-defined fields and define new ones.
By default, Name and Data Type are required to define a field. Here, the Name is free to choose and will create a unique ID automatically.
The Data Types are:
|String||A sequence of alphanumeric text or other symbols that can also contain spaces and numbers|
|Integer||Whole numbers, negative values are allowed|
|Float||A number that has a decimal place|
|Boolean||Represents logical value, can contain only two possible values: True or false.|
|Date||Values of a calendar date in the format DD-MM-YYYY|
|Table||A systematic arrangement of data in rows and columns (once a table is defined, all tables get detected automatically)|
|Single Choice||Choose exactly one item from a list of several predefined alternatives that you define|
|Multiple Choice||Choose items from a fixed number of alternatives that you define. Here, you define the number of maximum selections|
And you can mark the fields as required fields, and whether they require an annotation (by default, they do).
If you want to change existing fields, you can do so using the edit option. You can further delete or deactive fields.
Click "Next" to start processing the first document based on your settings.
You can either wait or "Skip to the dashboard" to proceed.
Updated 5 months ago