Building a good knowledge base and managing access to it helps you code faster and more easily when using Lingma.
Build a high-quality knowledge base
A knowledge base helps the model better learn specific company information. It's important to create good quality data, and manage who can access it. You can upload documents to turn them into retrievable knowledge data. Lingma will use this knowledge base for conversational search, code refactoring and improvements, and code generation. This is useful for enterprise standard verification, technical support, and many other scenarios.
Keep the following tips in mind:
Supported formats: PDF, CSV, DOCX, TXT, and Markdown formats are supported. Markdown is recommended.
Naming: Use concise and clear names. Avoid ambiguous naming.
Structure: Use headings to organize content. Don't stack too much information.
Content: Avoid pronouns and provide specific descriptions.
Additional notes: Include explanations for tables and images. Avoid extra blank lines and special characters.
For more information, see Knowledge base as context for chat.
Manage access control
Effective management of access control is essential for safeguarding sensitive information and ensuring that team members can easily access the knowledge they need. Below are some tips to keep in mind:
Knowledge base categories:
General knowledge bases for the whole company
Specific knowledge bases for certain teams or departments
Creating a knowledge base:
Select the AI Chat scenario when you create a knowledge base.
Set it to private.
Managing access control:
Make sure team members only have access to the knowledge bases they need. This protects data privacy and helps them find information faster.
For more information, see Build and manage knowledge bases as context for chat.