Knowledge Base
Your source of data for ASAP
Create a new knowledge base (Documents and files)
Go to ‘Knowledge Base’.
Click ‘Add.
Name your knowledge base and choose ‘Document and Files’ as the source of the knowledge base.
Pick your languages. The language support setting will determine factors such as document translation settings and semantic search settings.
Click ‘Create’.
Add your desired files for processing. The documents can be used after approximately 5-10 minutes. To check your file status, scroll down and click on 'View Document'.
Miscellaneous Knowledge, Limitations, and Constraints
Each Knowledge Base has no limits on the maximum number of files you can add.
CSVs are not chunked automatically; it will respect the contents of the row.
CSV headers have a max limit of 50 characters.
In the header, there are some words which are restricted. Please do not use these words:
ID
Data
Docs and PDFs are automatically chunked by our system and the chunking happens at around 900-1000 tokens. Approximately 4-5 English characters = 1 token. 1 Thai character (including vowels) = 1 token. To identify approximately how many tokens a piece of text is, you can use OpenAI's tokenizer: https://platform.openai.com/tokenizer
The languages you pick currently determine 2 things:
Documents you upload will be translated from the source language to the languages you pick. If you choose both English and Thai, all documents uploaded to the system will be saved in the original language and the translation to the other language. For example, an English document will have both Thai and English translations available.
ASAP's semantic search will attempt to match the user input question to the document with the appropriate language. For example, if I send a question in Thai, ASAP will match the Thai question to the Thai translation of the document.
Last updated