https://github.com/phaniteja5789/automated-summarization-of-class-methods-with-ollama-and-langchain-chains-leveraging-lcel-for-effic

https://github.com/phaniteja5789/automated-summarization-of-class-methods-with-ollama-and-langchain-chains-leveraging-lcel-for-effic

few-shot-learning inspect langchain llm ollama prompt-template python3

Last synced: 5 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/phaniteja5789/automated-summarization-of-class-methods-with-ollama-and-langchain-chains-leveraging-lcel-for-effic
• Owner: phaniteja5789
• Created: 2024-05-11T15:17:09.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-05-11T15:56:25.000Z (over 1 year ago)
• Last Synced: 2025-02-17T00:44:21.075Z (8 months ago)
• Topics: few-shot-learning, inspect, langchain, llm, ollama, prompt-template, python3
• Language: Python
• Homepage:
• Size: 604 KB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: Readme.md

Awesome Lists containing this project

README

          
This repository details out on the usage of Ollama and Langchain

Problem Statement

	Generate the summary of the function definition, group the similar summaries together

sourcecode.py ==> This is the entry point for the application to run, which imports other code files and perform the necessary operations and displays the final output of the application.

This projects mainly has 3 phases

	phase-1

		In the phase-1, identify the list of methods from each class that is present in the directory "/codes"

		

		After completion of phase-1, for each class it will list out list of methods present in the same class

		

		The entire code for the phase-1 is present in the IdentificationOfMethodsFromClass.py

		

		List of Modules used are os,importlib.util,inspect ==> in order to identify the methods present in the class

		

		

	phase-2

	

		In the phase-2, generate the summary for each method from the class.

		

		List of Modules used are langchain

		

		classes used are 

		

			langchain.prompts.prompt import PromptTemplate ==> Using this, created a Prompt that can be sent to the LLM

			langchain.llms.ollama import Ollama ==> Using this, I used the quantized model in my local machine locally.

			

		

		Used the LCEL(Langchain Expression Language), so that I created a chain that with prompt | llm

		

		

		After completion of phase-2, for every method there exists a summary generated using LLM

		

		

		The LLM used is "llama3"

		

		

	phase-3

	

		In the phase-3, group the summaries together based on their functionality

		

		List of Modules used are langchain

		

		classes used are 

		

			langchain.prompts.prompt import PromptTemplate ==> using this, created a prompt that can be sent as input to the LLM

			

			langchain.llms.ollama import Ollama ==> using this, I used the quantized model in my local machine locally

			

			langchain.prompts.few_shot import FewShotPromptTemplate ==> using this, I can create an examples and send to the LLM so that the response generated from LLM will be accurate

			

		Used the LCEL(Langchain Expression Language), so that I created a chain that with prompt | llm

		

		

		After completion of phase-3, there is a group generated which groups the similar functionalities together

		

		

		The LLM used is "llama3"

		

Output of the project is as below and present in the Output.png