Adding New Fields
Follow the interactive tutorial below and learn how to add new fields for extraction with super.AI Intelligent Document Processing (IDP).
Summary: How to add a new field
Preparation
- Access Project Settings:
- Navigate to the dashboard and select the project to add a new field to.
- Click
Project Settings
>Add new field
.
Field Settings
- Name:
- Provide a name. An ID is generated automatically.
- Data Type:
- Choose the data type from the dropdown. E.g.,
String
. - 📌 Data type options are defined below this tutorial.
- Choose the data type from the dropdown. E.g.,
- Annotation Requirement:
- Toggle: Decide if the field needs annotation.
- Type: If toggled on, select the annotation type:
BoundingBox
orTable
.BoundingBox
: Draws a box around the field.Table
: Uses table editing for document review.
- 🚫 Annotations may affect data extraction accuracy.
- Redaction Requirement:
- Toggle: Decide if the field needs redaction.
- Type: If toggled on, select redaction type:
Blacken
orBlur
.Blacken
: Draws black marks over the new field when it is detected in documents. Useful for redacting text.Blur
: Automatically blurs the new field in documents. Useful for faces, objects, etc.
- Description:
- Briefly describe the field for reference in
Job Settings
.
- Briefly describe the field for reference in
Advanced Settings (Optional)
- Status:
- Toggle new field on/off for data processing.
- Aliases:
- Provide alternate names to enhance AI recognition.
- 💡 The platform uses Large Language Models (LLM) for recognition. Aliases can improve accuracy, especially with varied naming.
- Instructions:
- Clearly describe how to extract the field.
- Additional Advice:
- Offer optional tips for improved field extraction.
- Examples:
- Provide valid examples of the extracted outputs.
- Anti Examples:
- Show what the extracted output should NOT look like.
- Validation Rules:
- Constrain valid output types.
Finalization
- Add Field:
- Click
Add field
to add the new field to the project. After doing so, you will be redirected to theJob Settings
page.
- Reorder Fields (Optional):
- Drag to rearrange.
- Save:
- Click
Save changes
for the new field.
- Pop-up Selections (Optional):
Unprocessed
: Search unprocessed documents for the new field.In Progress
: Extract the new field from ongoing processes.Completed
: Reprocess all completed documents to extract the new field.- 🚫 This will require reprocessing all completed documents.
- Confirm Changes:
- Click
Save changes
and await the confirmation message.
Data Type Definitions
Data Type | Description |
---|---|
Single Choice | Select one from predefined options. |
Multiple Choice | Select multiple from given options with set limit. |
String | Sequence of characters, including spaces and numbers. |
Integer | Whole numbers, including negatives. |
Float | Number with a decimal place |
Boolean | Logical value: True or False. |
Date | Calendar date in DD-MM-YYYY format. |
Entity | Distinct data object representing a real-world item. |
Table | Data arranged in rows and columns. |
Calculated | Match values with databases (e.g., Excel). |
Stamp | Data marker or label for verification or categorization. |
Updated about 1 year ago