{"version":"https:\/\/jsonfeed.org\/version\/1","title":"mathspp.com feed","home_page_url":"https:\/\/mathspp.com\/blog\/tags\/programming","feed_url":"https:\/\/mathspp.com\/blog\/tags\/programming.json","description":"Stay up-to-date with the articles on mathematics and programming that get published to mathspp.com.","author":{"name":"Rodrigo Gir\u00e3o Serr\u00e3o"},"items":[{"title":"itertools cheatsheet","date_published":"2026-07-13T16:28:00+02:00","id":"https:\/\/mathspp.com\/blog\/itertools-cheatsheet","url":"https:\/\/mathspp.com\/blog\/itertools-cheatsheet","content_html":"

Cheatsheet with visual diagrams that explain how the iterables from itertools<\/code> work.<\/p>\n\n

This cheatsheet contains diagrams that explain how the iterables from the module itertools<\/code> work in a visual way.<\/p>\n

Download this cheatsheet<\/a><\/p>\n

\"A4<\/p>\n

\"A4<\/p>\n

Download this cheatsheet<\/a><\/p>\n

\n

Reshaping<\/a><\/h2>\n

batched<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

islice<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

pairwise<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

chain<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

groupby<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

Filtering<\/a><\/h2>\n

filterfalse<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

\ntakewhile<\/code> & dropwhile<\/code>\n<\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

compress<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

Infinite<\/a><\/h2>\n

repeat<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

cycle<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

count<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

Complementary<\/a><\/h2>\n

zip_longest<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

starmap<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

accumulate<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

By default, accumulate<\/code> uses addition as its operation.\nThe magic in using functools.reduce<\/code> and itertools.accumulate<\/code> well is in realising the function that combines elements can be an arbitrary function of two arguments.\nFor example, you can imagine you're a painter and you're going to mix the colours in your palette:<\/p>\n

\"Visual\n\"Visual<\/p>\n

Combinatorial<\/a><\/h2>\n

product<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

When you want to compute the product of an iterable with itself two or more times, you can also use the argument repeat<\/code>:<\/p>\n

\"Visual\n\"Visual<\/p>\n

permutations<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

combinations<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

combinations_with_replacement<\/code><\/a><\/h3>\n

\"Visual\n\"Visual<\/p>\n

\n

Download this cheatsheet<\/a><\/p>\n

\"A4<\/p>\n

\"A4<\/p>\n

Download this cheatsheet<\/a><\/p>","summary":"Cheatsheet with visual diagrams that explain how the iterables from itertools work.","date_modified":"2026-07-22T18:06:55+02:00","tags":["python","programming","itertools","modules"],"image":"\/user\/pages\/02.blog\/itertools-cheatsheet\/thumbnail.webp"},{"title":"Write a coding agent from first principles: better tools","date_published":"2026-07-06T15:00:00+02:00","id":"https:\/\/mathspp.com\/blog\/write-a-coding-agent-from-first-principles-better-tools","url":"https:\/\/mathspp.com\/blog\/write-a-coding-agent-from-first-principles-better-tools","content_html":"

Improve the capabilities of your agent by providing it with better tools.<\/p>\n\n

Introduction<\/a><\/h2>\n

This tutorial builds on the coding agent you implemented in the tutorial “Write a coding agent from first principles”<\/a>.\nIn this tutorial, you'll take your agent and improve its capabilities by implementing the text edit and bash command tools that Anthropic provides.<\/p>\n

Why use Anthropic's tools?<\/a><\/h2>\n

In the previous tutorial you implemented a coding agent that has a few tools that it can use to read, write, and execute, code.\nThe tool \"bash\"<\/code> can be used to execute arbitrary commands and the tools \"read\"<\/code>, \"write\"<\/code>, \"replace\"<\/code>, and \"insert\"<\/code>, can be used to edit files.<\/p>\n

As it turns out, these tools are so universally useful that Anthropic trained its models on specific schema definitions for these tools.\nThe tools still run on the client side, so you'll still get tool use blocks in the API responses, but you don't have to define the schema for the tool.\nYou just specify the tools by their Anthropic types and names<\/a>, and the LLMs will happily request tool uses.<\/p>\n

Anthropic trains their models on a number of useful tools but you'll focus your attention on two tools that emulate the functionality you already have:<\/p>\n

  1. Text editor tool<\/strong>: this tool replaces the four tools you defined to read, write, replace, and insert, text in text files<\/li>\n
  2. Bash tool<\/strong>: this tool provides a persistent bash session that can run bash commands<\/li>\n<\/ol>

    By replacing your tools with Anthropic's, the agent will be able to make better tool calls consistently, since Anthropic trains their models on their specific tool schemas.<\/p>\n

    The native text editor tool<\/a><\/h2>\n

    To define support for Anthropic's text editor tool you need to add it to your list of tools.\nThe name of the tool is \"str_replace_based_edit_tool\"<\/code> and its type is \"text_editor_20250728\"<\/code>.\n(The type carries a versioning suffix<\/a> that may influence the tool's behaviour, so make sure you use the right date suffix.)<\/p>\n

    Since you'll be using Anthropic's text editor tool, you can delete the functions read<\/code>, write<\/code>, replace<\/code>, and insert<\/code>, and the corresponding dictionaries that go in the list TOOLS<\/code>.\nInstead, add the dictionary that specifies the Anthropic tool:<\/p>\n

    # ...\nTOOLS = [\n    {\n        \"type\": \"text_editor_20250728\",\n        \"name\": \"str_replace_based_edit_tool\",\n    }\n]\n\n# Bash tool defined and added later.<\/code><\/pre>\n
    For organisation purposes, you'll define the text editor tool and the bash tool in their own submodules, so create the folder tools<\/code> and then create the file tools\/str_replace_based_edit_tool.py<\/code> under src\/agent<\/code>.\nIn there, you'll define the code to handle the tool call.<\/p>\n
    The text editor tool<\/a> is a 4-in-1 tool that allows you to view<\/em>, replace<\/em>, create<\/em>, and insert<\/em>, text.\nTo disambiguate the action you want to do, the tool use request includes a command<\/em>:<\/p>\n
    # Example tool use dictionary:\n{\n  \"type\": \"tool_use\",\n  \"id\": \"toolu_01A09q90qw90lq917835lq9\",\n  \"name\": \"str_replace_based_edit_tool\",\n  \"input\": {\n    \"command\": \"view\",  # <--\n    # ...\n  }\n}<\/code><\/pre>\n
    You'll use the key \"command\"<\/code> from the...<\/p>","summary":"Improve the capabilities of your agent by providing it with better tools.","date_modified":"2026-07-07T09:58:48+02:00","tags":["llms","programming","python"],"image":"\/user\/pages\/02.blog\/write-a-coding-agent-from-first-principles-better-tools\/thumbnail.webp"},{"title":"Write a coding agent from first principles","date_published":"2026-06-22T14:32:00+02:00","id":"https:\/\/mathspp.com\/blog\/write-a-coding-agent-from-first-principles","url":"https:\/\/mathspp.com\/blog\/write-a-coding-agent-from-first-principles","content_html":"
    Learn how to write a coding agent in this Python tutorial that teaches how to interact with an LLM through an API, how to manage the conversation context, and how to do tool calling.<\/p>\n\n
    Introduction<\/a><\/h2>\n
    This tutorial will show you how to create your own coding agent<\/strong> from first principles.\nBy doing so, you'll understand how coding agents work under the hood.<\/p>\n
    Prerequisites<\/a><\/h2>\n
    To be able to follow this tutorial, you'll need<\/p>\n