Skip to content

Latest commit

 

History

History
121 lines (81 loc) · 4.54 KB

CHANGELOG.md

File metadata and controls

121 lines (81 loc) · 4.54 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

0.2.2 - 2024-03-24

Added

  • New exercises in exercises/2024/ChatGPT.
  • ChatGPT template in notes.
  • Script to extract exercises from ChatGPT responses in scripts/generation/chatgpt_extract.py.

0.2.1 - 2023-01-29

Updated

  • Refactored scripts/export/export_dozeloc.py to use new Python API and allow selecting exercises by description language.

0.2.0 - 2023-01-28

Added

  • Import script that allows importing exercises into the repo without cloning it or running any scripts.
  • Second variant of suggest_tags that uses a list of predefined tag candidates.
  • Script to automatically translate exercise text into another (natural) language.
  • Basic Python API tracked as Poetry project.

0.1.0 - 2022-01-06

Added

  • Script to check that each exercise description file includes a YAML header with a UUID that is truly unique in the database.
  • GitHub Actions script that runs UUID-check after each push.
  • Several new exercises that were not created for a specific course but just for fun (covering integer overflows, floating point operations, and bitwise operators)
  • General file to collect exercise ideas

Changed

  • Versioning scheme: Added scripts and changes to general structure now increase MINOR version and updates that only affect exercise content increase PATCH version.

Fixed

  • add_uuid script did not recognize existing UUIDs properly (missing M flag for regex)
  • AuD exercises did not have a UUID

0.0.4 - 2021-12-29

Added

  • Type aliases for more complex types such as Index and IndexVector
  • Function explain_similarity, which returns the terms that had the most weight in calculating cosine similarity.
  • Parameter explain in find_similar, which allows printing explanations for which terms determined the similarity.

Changed

  • Term frequency is now divided by total number of terms in order to make index independent of document length.
  • Uses log ((n+1) / count) instead of log ((n+1) / count) in idf to avoid weights becoming zero for terms that appear in each document.
  • Index is now a dictionary instead of a list of tuples.
  • Use custom tokenization algorithm, which separates by spaces and then trims non-word characters from both ends of the token.
  • Only apply stemming to words, which only consist of word characters and -.
  • Removes keep_default from dict_reduce and introduces dict_filter instead.

Fixed

  • Applies make_hashable to tags to ensure that same tag names are used on both sides of comparison.
  • Forgot to apply IDF weights to query vector.

0.0.3 - 2021-12-27

Added

  • this changelog
  • Python script that can generate recommendations which tags to add or remove from an exercise based on existing exercise tags

Changed

  • Switched to keyword-style tags for all existing exercises.
  • Updated import scripts to provide keyword-style tags and add UUIDs.
  • Added explanation of UUID use in readme.

0.0.2 - 2021-12-21

Added

  • script to add UUID to exercises
  • all exercises can now be identified via a UUID
  • import script for exercises from "Algorithmen und Datenstrukturen"
  • exercise files for Algorithmen und Datenstrukturen 2019
  • requirements.txt file for scripts

Changed

  • Tag scheme in readme now uses - category: value as syntax instead of - tagname.

0.0.1 - 2021-12-15

Added

  • README
  • exercises from "Grundlagen der Informatik (BI Master)"
  • import script for dozeloc exercises from "Grundlagen der Informatik (BI Master)"
  • first exercises from other courses
    • Brückenkurs 2019
    • Objektorientierte Programmierung 2014
    • Objektorientierte Programmierung (BI Master) 2015
    • Girl's Day 2019 (full)