verba 0.1.4

A collection of modules used for language processing and modeling

ABOUT

verba is a framework for working with LLMs and performing NLP tasks. It is designed to offer an alternative to the overly-abstracted methods found in commercialized packages such as langchain.

----------------------------

LICENSE

GPL-3 Summary:

You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions. In other words, any derivative work of this software shall be released under the same GPL license as the original software, meaning the modified code must be exactly as free and open-source as the original.

----------------------------

INSTALL

verba is installed using pip:

pip install verba

NOTE: In order to utilize LLM functionality you need Ollama and the Ollama Python package installed on your machine.

SEE:

• https://ollama.com/download
• https://github.com/ollama/ollama-python

----------------------------

Ragby: Retrieval-Augmented Generation (RAG) By Yourself

• A collection of RAG-related methods

-----------

STEP 1) Import & Initialize:

from verba.OllamaHelper import Ragby

ragby = Ragby(chat_model = "llama3", embedding_model = "llama3")

STEP 2) Create Chunks:

• First place your input data file (.txt or .pdf) in a directory named data
• Run:

# for .txt files:
chunks_obj, chunks_path = ragby.make_chunks("my-txt-file.txt")

# for .pdf files:
chunks_obj, chunks_path = ragby.make_chunks_pdf("my-pdf-file.pdf")

• A directory named chunks will be created, containing the two returned objects

STEP 3) Create Embeddings:

• Run:

embeddings_path = ragby.make_embeddings(chunks_obj, "my-txt-file.txt")

• A directory named embeddings will be created, containing the returned object

STEP 4) Create a System Prompt that is relevant to your output:

SYSTEM_PROMPT = "You are a helpful assistant who answers questions using only the provided CONTEXT. Be as concise as possible. If you are unsure, just say 'I don't know'.\n\nCONTEXT:\n"

STEP 5) Run:

CHAT = ragby.chat(
    user_prompt = "What is the summary of this article?", 
    system_prompt = SYSTEM_PROMPT, 
    text_chunks_path = chunks_path, 
    text_embeddings_path = embeddings_path, 
)

print(CHAT)