😏 Hard to follow the news since it is all in Dutch, yet you still want to learn Dutch? What to quickly skim through the latest news, then decide what to read further? Ochtend Flits (Morning Flash) is for you.

😏 “Scratch your own itch,” has to be one of most heard terms in the world of business and startups. However, many times those small annoyances, we face everyday, go unnoticed. To train myself to pay attention to them, I keep a list of pain points on my phone. But I wanted an easier way to keep them and also share them with others to get their feedback I took it offline because I did not see enough interest from others to keep using it.

🗣 Translate something into Egyptian Arabic with a vanilla LLM and you often get output that drifts toward MSA or an exaggerated dialect. Targem uses few-shot learning: it retrieves relevant bilingual examples first, then uses them to steer the model toward the natural, educated Egyptian Arabic you actually want. 💡 For example, instead of the stiff "المتجر الهولندي الذي يبيع كل شيء...", Targem can produce "المحل الهولندي اللي بيبيع كل حاجة وبيحسسك إنك في بيتك".

🎓 I started writing this library as part of my Information Retrieval and Natural Language Processing (IR and NLP) module at the University of East Anglia. It was mainly meant to detect Review Spam (Machine Learning - Classification). However, It has more functionalities such as Vector Space Model (VSM), as well as some other IR functions such as tokenizing, n-grams, stemming and PoS (Part of Speech) tagging.