LlamaIndex + ChatGPT: Ingest Data and Experience UNBELIEVABLE Query Results! 🌟

Key Takeaways

hi everyone I'm really excited to show you how to inest data into a large language model so when you ask any question in regards to your data a large language model can respond to you accurately that's exactly what we're going to see let's get [Music] started to do this we are using llama index using llama index you should be able to injest data index data and also query data the Llama index can handle on un structured data structured data and also semi-structured data I'm going to take you through a step-by-step guide on how to set this up and run this first create a virtual environment cond create hyphen and llama index python 3.11 click enter then activate the virtual envirment cond activate llama index enter now install llama index package using pip install llama index enter this will install the package required in our application these are the steps which we are going to follow first we are going to convert the data to embeddings and then index the data and save it then query the data so create a file called app.py and then from llama index import simple directory reader using simple directory reader we should be able to retrieve data from a directory so if you see the folder structure I have created another folder called Data inside the data I have got another file called data.txt that's where all your data will be you can create as many txt files in this folder and this application should be able to handle all those so if you go inside the data folder I have document about llama index going back to our code we are going to read from the directory so simple directory reader and the name of the folder is data then load the data next create an index for that we need to import Vector store index Vector store index do from documents and then you mention the documents the third step is we're going to save the index storage context. persist so as we seen in the image first step is to get the data from the folder then Vector store index then you are saving that index the fourth step is querying that index next query engine equals index do as query engine next response equals query engine do query and then what is llama index then print response now we going to run this program in your terminal export the open AI API key which you can generate from open AI website then click enter Then type python app.py llama index is a data framework for llm applications you can see clearly that it's retrieving from the vector store this is a line which saves that storage if you see our folder there'll be a folder called storage inside the storage we'll have different files these are the embeddings which we just created using this function next we want to load these embeddings and query with that if we run this program each time it will automatically create embeddings every single time you run the program so it's best to load the embeddings again from the saved index that's exactly what we're going to do now I'm going to create another file called retrieve dop in this we're going to rebuild the storage context load the index and query from the index so from llama index import storage context and then load index from Storage storage context equals storage context Dot from defaults then you're going to mention in which directory the storage is storage directory equals do/ storage that will rebuild the storage context now we are going to load the index load index from storage and then pause the storage context next again as we did we're going to query by initializing the query engine next response equals query engine. query what is llama index print response that's it first rebuild the storage context load the index and then you can query we going to run this code now in your terminal python retrieve dopy and click enter llama index is a data framework for llm application yeah this is loaded from the saved m ings that's it you can load as much as data you want and then query that using the lodge language model I hope you like this video please do like share and subscribe thanks for watching