Langchain OpenAI Assistants API For Beginners 🦜️🔗 Function Calling made EASY! 🤯 (Full Tutorial)

Key Takeaways

this is amazing what if we can integrate Lang chain with your open AI assistance API to create assistance API we have four different steps but what if we can combine all together into one assistance API with the Lang chain and making it simplified that's exactly what we're going to see today let's get [Music] started hi everyone I'm really excited to show you about Lang chain with assistance AP I'm going to take you through step by step on how to do this but before that I regularly create videos in regards to artificial intelligence in my YouTube channel so do subscribe and stay tuned in Lang chain there are functions which simplifies our work of creating an assistant rather than creating an assistant then a thread then a message and then running it you can do all these things in one function to code that first make sure you install L chain and open AI in your virtual environment and click enter next export your open a API key like this and click enter now in your file from Lang chain. agents. open Assistant import open AI assistant runnable this is the function which can make our work easy now using that open AI assistant runnable I'm going to create an assistant name Lang chain instruction you are a personal math tutor and the tools we are using it code interpreter model is gp4 Turbo next output equals assistant. invoke there you ask your question content what's 10 - 4 raised to the 2.7 now we going to print the output by looping through the output now we're going to run this code in your terminal python assistance Lang chain. Pi that's where my code is and click enter now you got the answer here the result is approximately minus 32.2 Lang chain did all the hard work for us and gave us the answer in an easy way much more than that what if we can add our function calling ability here now we we're going to add function calling to this in your terminal pip install duck ducko hyphen search and click enter this duck. go search package will help us to search the internet so we're going to add browsing the internet facility using custom function calling in our assistant so first from Lang chain tools import Duck Duck Go search Run next we are defining the list of tools that is dugdug go search run then we are passing this tool to this list and we are mentioning that as agent equals true this does all the magic the agent feature in the assistant will be capable of handling all the requires action status if you see the original steps for adding a function calling ability you need to have all these five steps but the agent feature will solve that issue now instead of invoking we are importing from L chain agents import agent executor then agent executor equals agent executor and you're passing assistant as the agent and the list of tools next output equals agent executor do invoke previously we did assistant. invoke here we are doing agent executed. invoke and you are asking a question what's the weather in London today divided by 2.7 finally we are printing out the results now I'm going to run this code python assistance Lang chain. and click enter all the five steps will be happening behind the scenes now we got the answer here's the content what's the weather in London today divided by 2.7 the output the current weather in London is 6° and if it's for Celsius this is the output and if it's for fenit this is the output as simple as that you integrated all the five steps in literally two bits of code one you're creating assistant with as agent and then you're using agent executor as simple as that I'm going to create more videos in regards to this so stay tuned I hope you like this video do like share and subscribe and thanks for watching