[Day 81] RAG from scratch - chunking is very important!

 Hello :)
Today is Day 81!


A quick summary of today:
  • Continued with my custom RAG from scratch project - extracting knowledge from a bank's terms and conditions (github repo)


So, yesterday my main struggle was reading tables. I remembered gemini can read pictures so I gave the below to gemini to try to read and give me the text. 
The output, not very good. Gemini could not read this table very well, and such tables were common in yesterday's PDF
Yes, this is an image, but even if a powerful gemini LLM could not read this table and output it as text, for now at least, I gave up on this particular PDF, and looked for one with a bit more straigh-forward text and less tables. 

The newly chosen pdf is here: on github

Now... given an allegedly more simple pdf, I used the code as it was to get outputs. But the results were just... really bad. Most times even though the scores were high, and the top-1 included the exact answer, the output was  ~'the context does not include {query} information'.

I decided to go back to the start, where I preprocess the pdf, and explore more chunking mechanisms. I found this youtube video exploring different chunking methods.

It covered:

  1. Character split (which I was using)
  2. Recursive character split
  3. Document spcific splitting
  4. Semantic splitting
  5. Agentic splitting
The youtuber mentioned that the default method he uses is the 2nd one - recursive character split, so I decided to move from 1. to 2., and explore how recursive splitting improves my process. 

This is the final text_splitter I went with:
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1500, chunk_overlap=0) (tbh, I am writing this blog now and 1500 might be an overkill)

I played around a lot with chunk_size and chunk_overlap - to see the effects, and I don't think I mastered knowing the impact yet. I am sure 1500 chunk_size is even that good. Frankly, in the preprocessing there are other components like splitting, regex-ing characters - maybe there are better ways to do these. It will take trial and error to figure this. 

The query list I used to see results is:

Base prompt is like this

I am yet to explore different prompting strategies. So far only concentrated on the preprocessing step. 

After settling on chunk_size 1500 and chunk_overlap 0, I decided to streamline the process a bit, so I created executable preprocess_pdf.py and rag.py files - can be seen in the repo.

As for results, they are incosistent... 

Example is this 'What is meant by Business Day?' query.
The top-1 context includes the answer, but some times I get the answer, sometimes I do not.

I shared my code on reddit and on some ML discord server, to look for advice. But tomorrow I think I shall concentrate on finding a better chunking method. 



That is all for today!

See you tomorrow :)

Popular posts from this blog

[Day 198] Transactions Data Streaming Pipeline Porject [v1 completed]

Image
 Hello :) Today is Day 198! A quick summary of today: data streaming pipeline project [v1 done] Here is a link to the project's repo. Well ... I did not know I can do it in a day (~14 hours) after yesterday's issues but here we are. Turns out in order to insert the full (~70 variables with nested/list structure), I need the proper pyspark schema. And yesterday I did not have that and that is why when I was reading data in the kafka producer I was getting NULL in the columns - my schema was wrong. Well today I not only fixed the schema for the 4 variables I had yesterday, but included *all* the variables that come from the Stripe API ~ 70 (for completeness).  When I run docker-compose, the data streams and is input into the postgres db (and is still running). Unfortunately, the free Stripe API for creating realistic transactions has a limit of 25, so every 3 seconds, 25 new transactions are sent to the db. It has been running half the day (since I got that set up) and as I am wr
Read more

[Day 107] Transforming natural language to charts

Image
 Hello :) Today is Day 107! A quick summary of today: created text2chart - transforming natural language to charts [ github ] [ webapp ] I am super sick, but in the around 3 hours of my fever not being high, I managed to study a bit. I was again thinking about creating some nlp project but was stuck on the idea of how to host a model. After looking around I found this article where they create a similar app called chat2vis  and they use codellama. I know codellama is an open source model that I can get from huggingface and I wondered, how did they host it. Turns out - they ask the user to input a huggingface api key, which then allows for querying and generating content. (AWESOME! this is applicable to some of the other PDF chat apps that I have, but I will look to make an app of them later).  My fever eased a bit, and I sat on my chair to do some work ^^ Here is a breakdown of the app. I used streamlit. Also, 2 default datasets during dev: Financial Statements of Major Companies(2009-
Read more

[Day 54] I became a backprop ninja! (woohoo)

Image
 Hello :) Today is Day 54! A quick summary of today: Read through cs231n's neural network lectures [1] [2] [3] Spent 8 hours going over Andrej Karpathy's backprop ninja tutorial After yesterday's confidence boost in doing backprop, I came in today with the intention of becoming of, as Andrej Karpathy calls it, a backprop ninja.  The cs231n NN lectures were the usual about how NN became to be, param initialization, activation functions, loss functions, regularization and optimizers. Nothing out of the ordinary, but still was good to reassure myself of that knowledge. But that was definitely not the main even of today. As I mentioned, I wanted to become (or at least attempt to become) a backprop ninja today. I have watched that particular tutorial twice, the 1st time I had no idea what is going on, the 2nd time I could follow, and now it was the 3rd time - and now, I wanted to participate and actually see the why and the how of the things we have to do, when we do manual back
Read more