Hi Flowise,

Here is a PR which aims to improve CSV Data loading, th key points are:

• Improve CSV Handling by using a custom loader with Papaparse (auto detect separator, auto detect headers, streaming capability, direct local file load capabilities, remote file load capability, ...) which is more memory friendly and efficient
• Clean some code (create methods to make code more readable, extract some utils functions, ...)
• Cleanup S3 Loaders (File & Directory) & use custom CSVLoader in it.

This also opens some opportunities:

• Improve memory management by directly loading CSV Files into the parser
• Stream responses instead of getting fat raw data
• Load CSV files from a remote URL