11. OpenAI and Llama Index - Financial News Analysis
Вставка
- Опубліковано 28 чер 2024
- In this video, we use OpenAI and Llama Index to analyze financial news. Like the video? Support my content by checking out Interactive Brokers using the link below:
www.interactivebrokers.com/mk...
Source Code:
github.com/hackingthemarkets/...
Note: I have updated the source code to reflect recent changes to Llama Index 0.6 that happened after I recorded this video at the end of April. I have made notes on the changes in Github and how they differ from a few lines in the video.
0:00 What is Llama Index?
1:06 What We’re Building - News Fetching and Indexing
2:26 News Querying and Reporting Interface
4:39 Code Walkthrough
6:13 Fetching News with Interactive Brokers API
10:23 Vector Index with Llama Index
16:45 Querying the Index
19:58 Wrapping Up
---
This content is provided by a paid Influencer of Interactive Brokers. Influencer is not employed by, partnered with, or otherwise affiliated with Interactive Brokers in any additional fashion. This content represents the opinions of Influencer, which are not necessarily shared by Interactive Brokers. The experiences of the Influencer may not be representative of other customers, and nothing within this content is a guarantee of future performance or success. None of the information contained herein constitutes a recommendation, promotion, offer, or solicitation of an offer by Interactive Brokers to buy, sell or hold any security, financial product or instrument or to engage in any specific investment strategy. Investment involves risks. Investors should obtain their own independent financial advice and understand the risks associated with investment products and services before making investment decisions. Risk disclosure statements can be found on the Interactive Brokers website.
Interactive Brokers is a FINRA registered broker and SIPC member, as well as a National Futures Association registered Futures Commission Merchant. Interactive Brokers provides execution and clearing services to its customers. For more information regarding Interactive Brokers or any Interactive Brokers products or services referred to in this video, please visit www.interactivebrokers.com.
Like the video? Support my content by checking out Interactive Brokers: www.interactivebrokers.com/mkt/?src=ptlPY1&url=%2Fen%2Findex.php%3Ff%3D1338
Source Code: github.com/hackingthemarkets/financial-news-llama-index
Note: I have updated the source code to reflect recent changes to Llama Index 0.6 that happened after I recorded this video at the end of April. I have made notes on the changes in Github and how they differ from a few lines in the video.
Timestamps
0:00 What is Llama Index?
1:06 What We’re Building - News Fetching and Indexing
2:26 News Querying and Reporting Interface
4:39 Code Walkthrough
6:13 Fetching News with Interactive Brokers API
10:23 Vector Index with Llama Index
16:45 Querying the Index
19:58 Wrapping Up
Who’s ready for a next round of chants? “Larry Larry Larry Larry” 💪🏻
Mind blowing demos delivered casually. Great content.
I had heard about llama index but haven't looked into it yet. It's mind blowing how it effectively increases the context size to... whatever you want? Magic!
This is wild! Amazing potential for heaps of entertaining projects!
Great stuff as always! ❤
Awesome works Larry! Exactly what I need. Building one personal financial analyst. Currently work on the Twitter side. Plus this one is perfect!
Thank you great video!!!
Great job. My respect.
Thanks man you are the best
tnx man great videos
Good stuff Larry
lol, I like how one of the risks is "reckless short sellers" - baggy alert!
Boom.
Nice work. I am wondering if we could make it for return graph (e.g., bar chart or line chart) if we feed in data like historical revenue and earnings, etc similar to finchat
Fantastic content as always, thanks Larry.
I'm actually a C# developer - does anybody know of a youtube channel, blog, or ??
that has the same (or close to) the high quality of Larry's content?
Thanks and keep up the great work!!
Thomas
awesome work! and hope you didn't sell your $NVDA lol
Awesome. Is there a way to dump an actual prompt that is sent to OpenAI API behind the scene for debug purpose?
What would be a sufficient temperature for the openai model to work properly with this? I feel if your closer to one it will start making stuff up but closer to zero will give you nothing back. What are you using here?
My man!!! Exactly what i was looking into
Great vid, Larry. The OpenAI API key: it charges based on the number of tokens and queries made to the LLM, right? How much does it cost to build and use this financial news analysis app (with average usage)?
They give you enough that the cost is very minor if at all. Of course this depends on how many queries you make but its very reasonable (ie, its probably in the pennies for most users)
Great video! How does Llama Index differ from Open AI Embeddings?
Glad you shaved 🪒
Is LlamaIndex preferred over LangChain?
Dude where did you go? Miss your tuts
apache flink has a version for python can you make a tutorial how setting up?
you should really integrate the db's (vector db's.. maybe nuclia or milvus or something.. one of the open-source ones?) I'm actually incorporating all this stuff into deephaven and it's working nicely.. while doing all of id, having an instance of the ChatCompletion/MarketsHacking buddy to answer questions, see/hear the inputs.. etc..
hey Larry, did you mention in one of your videos about connecting tot Slack and loading messages into a model? I binged watched a bunch of yours and now i can't remember where you mentioned it. What's your recollection?
Thanks
I thought I mentioned it in this one? Would need to look in a bit. I mentioned llama index has various data connectors for Google Docs, discord, etc
@@parttimelarry found it and the widget! thanks! BTW - way back when you did the WhisperAi video with the Fed speech, you remember a ball park cost for the api portion? I want to do a live streaming version.
Hi Larry, thanks a lot for the sharing. I tried and changed to use GPTVectorStoreIndex as LLama-Index has probably changed things around and I kept getting this error: AttributeError: 'VectorStoreIndex' object has no attribute 'save_to_disk'. Can you pls help?
Larry, when is the next one to complete this one ?
A little later today need to go somewhere for a bit but will be back
Why not use free models for embedding like all-mpnet-base-v2 ?
This video is on the OpenAI for Finance playlist, so I have been using OpenAI tools. Will probably do a different series on free open source models in the future.
Larry can you use in jupyter notebook?
Yes, you can use Llama Index in Jupyter. The method I used for programmatically fetching real-time financial news wouldn't work in Jupyter, but I provided the text of the articles I used in the Github repository so you could use that text to try out. Or you can just save a bunch of articles or text to the articles directory.
@@parttimelarry Hey man, Articles provided were html would that works too or has to be json ?
Can you suggest a good free alterantive for news feeds to Interactive Brokers?
I believe Polygon has a free tier, but they limit the amount of API requests
@@parttimelarry Thank you !
You have access to GPT-4-32K Token?
Which IDE u r using to execute Python code? Any idea
VS Code
1st this time 😂
Why LlamaIndex instead of LangChain?
I have another one on LangChain coming, I thought this one was a little simpler for what I was doing here. I will do a project that interacts with additional tools that uses LangChain.
man the code is so much harder to read and understand (so white and zoomed in) :(
this series has been hard to follow along with due to that.. probably just me I guess..
Amazing! Just looking to do exactly this! Great timing!🥳🦾😎