Hi!
This is a prototype (German) Langchain agent webapp particularly useful for situations when customers inquire about the location of a specific product or its stock status (especially when the vending shelf appears empty).
A great use case for this would be to connect the LLM-agent to a SAP inventory management system SQL database and display the app on a work phone. A prime application of this tool would be its integration with a SAP inventory management system's SQL database, allowing the app to be displayed on a work phone for easy access. To avoid the need for creating a specific SQL database for this demonstration—and since SQL agents operate in a manner very similar to CSV agents—we've opted to design an imaginary supermarket inventory using a CSV file for this project. (Important: the file is UTF-8 encoded and uses commas for separation.)
The following libraries are needed for the project:
- os (for setting the OpenAI API-key as an environment variable)
- langchain
- streamlit (for creating prototype webapps)
As always, the code is thoroughly commented.
Have fun!
![webapp_screenshot](https://private-user-images.githubusercontent.com/140447879/320312722-90e11fd0-fa52-4e3c-8b71-e8cf56d9ff57.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3MjU0MTQ5ODYsIm5iZiI6MTcyNTQxNDY4NiwicGF0aCI6Ii8xNDA0NDc4NzkvMzIwMzEyNzIyLTkwZTExZmQwLWZhNTItNGUzYy04YjcxLWU4Y2Y1NmQ5ZmY1Ny5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjQwOTA0JTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI0MDkwNFQwMTUxMjZaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT1mNTVmMzk0NjhiNTNhZmQ5MzQ0NWQxM2FiNDY3ZDQ3ZWIxYTJhNWNlNTcxNTBiODMwMzEyMWU3MTdjMjZmOTEyJlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCZhY3Rvcl9pZD0wJmtleV9pZD0wJnJlcG9faWQ9MCJ9.ANuhkzP8SDFGyf7AvvI3z9vfQj8RKEE04qgh52mJQiY)