Conversational apps with LangChain and Python for beginners

These notebooks were created as part of a workshop for the Minds Mastering Machines on April 23rd in Cologne.

Overview and requirements

The following examples were tested using Ollama on-premises and Azure OpenAI. Where the use of Ollama was not possible, only Azure OpenAI was used.

What are LLMs, tokens, context window, embeddings, vector databases

What is LangChain?

LangChain is a framework for developing applications powered by large language models.

Overview of LangChain components

Langchain repository

Here is the LangChain repository for Python.

Project setup

• Bring your own large language model. LangChain supports the followings
• Install LangChain
• Test the installation and run this notebook.

First LangChain application

hello_langchain.ipynb

Use different LLMs

llms.ipynb

Caching and debugging

• caching.ipynb
• debugging.ipynb

Prompting

prompting.ipynb

Prompt Evaluation

promptfoo

Application types

RAG - Retrieval-Augmented Generation

• rag.ipynb
• rag2.ipynb

RAG Evaluation

rag-eval.ipynb

Talk-with-your-database

sql_db_qa.ipynb

Agents

• research.ipynb
• generative-agents.ipynb

UIs for Langchain

• gradio-ollama-chat.py
• chainlit-app.py
• streamlit-ollama-chat.py

Low-code development with Langchain

Here are examples of Langflow and Flowise.

Bringing Langchain applications into production

Simple GRadio Chat App in Hugging Face

Sources

https://www.udemy.com/course/langchain/learn/lecture/37507516#overview https://learn.deeplearning.ai/courses/chatgpt-prompt-eng/lesson/1/introduction https://github.com/CoderPush/chatlit https://www.gradio.app/ https://github.com/langchain-ai/langchain https://flowiseai.com/ https://langchaindart.com/#/ https://python.langchain.com/cookbook/ https://www.langflow.org/ https://docs.smith.langchain.com https://api.python.langchain.com/en/latest/vectorstores/langchain_community.vectorstores.sqlitevss.SQLiteVSS.html