Python Friday

Our tests should be independent from each other, but how do we know that they really are?

Python 3.15 did an additional round of bug fixing what took it one extra week compared to the annual release cycle.

When we build complex projects or maintain custom bug fixes for third-party libraries like scrapetube, we often need to install a package from a local directory directly into our active virtual environment (.venv).

Last week we created a RAG on top of our Obsidian vault that allows us to search the stories of Sherlock Holmes.

With our Sherlock Holmes books nicely organised in an Obsidian vault, we can start building a RAG agent that uses this data store to ask questions about the many adventures of the famous detective.

Obsidian is a popular note-taking application that uses plain Markdown files, Frontmatter metadata and the WikiLinks syntax ([[]]) to connect notes.

My goal is to create an Obsidian Vault filled with all the books on Sherlock Holmes.

Last week we used EbookLib to parse EPUB books.

An EPUB file looks like a single book, but under the hood it is a ZIP archive full of XHTML files, images and a manifest that ties them together.

Obsidian 1.12 introduced an official CLI that talks to the running app.

When we design a group of related classes, we often want to guarantee that every member provides the same set of methods.

Last week we made our first steps with Deep Agents of LangChain.

With Deep Agents we get a relatively new component from LangChain/LangGraph that allows us to build high-level agents.

If we dive into Python scripting or need to share a small utility with a colleague, we quickly run into a familiar problem: Incompatible or missing package versions.

Frontmatter is the metadata format for Markdown and YAML that we can use in MkDocs and many other tools.

A multi-agent system is an architecture style where we split a larger task across several specialised agents instead of relying on one LLM call to do everything.

Last week we saw how FastMCP can help us to create a MCP server.

An MCP (Model Context Protocol) server is an open-standard integration that acts as a bridge between Large Language Models and external data sources or tools.

A common way to test the "quality" of an AI solution is to ask for how many r’s are in the word raspberry.

The more complex our applications get, the harder it is to follow along our graph.

When we switch from our hand-written long-term memory solution in last week's post to a pre-build SqliteSaver, we not only need less code, but we gain new options.

Last week we added short-term memory to our LangGraph application.

A while back we added a chat history to our LLM client so that we did not need to constantly repeat ourselves.

Last week we created a basic LangGraph example for the human-in-the-loop pattern.

Last week we learned how to build our own tools to use in LangGraph.

LangGraph gets interesting as soon as we start to integrate it with our tasks.