Jakub Dubec’s Post

As translators, we often encounter the option for Regex in our CAT tools. But what does it mean, and how can it help us? Let’s look at a quick example I was using a few minutes ago: ^\d. In the image that accompanies this post, the checkbox labeled "Regex" is selected, and ^\d is entered in the input field. This might seem complex, but it’s actually quite straightforward. Here’s the breakdown: ^: Indicates the start of a line. \d: Matches any single digit (0-9). Together, ^\d means: 👉 "Find all segments that start with a number." Why is this useful? ✨ Filtering segments: Quickly filter segments starting with numbers, perfect for technical documents, for example. ✨ Finding specific patterns: Quickly find parts of your text that match a pattern, not a specific word, phrase or string. Regex can initially seem daunting, but it's a powerful tool for enhancing our efficiency in CAT tools. Understanding simple expressions like ^\d can significantly improve our text handling capabilities. Do you use regex? #regex #CATtools

A publications status update and announcement of my upcoming course on #memoQ #Regex Assistant libraries. This will be the fourth run of that unless you count the BP WTF sessions, which would make it the sixth. This time around I want to look at more options for regex sharing for other #translation workspaces. https://lnkd.in/d7GdxQdT #xl8 #l10n #localization

𝗛𝗮𝘀 𝗔𝗻𝘁𝗵𝗿𝗼𝗽𝗶𝗰 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘀𝗼𝗹𝘃𝗲𝗱 the problem of constructing prompts by providing the Prompt Generator? Yes and no :-) The value of a good prompt doesn't come solely from its structure (here indeed, especially for less advanced users, the Prompt Builder can make a big difference though), but also from the amount of information contained in the prompt. We can't assume that if the information can be compressed into 2 sentences (those that we provide as the input), it will magically become more powerful if we expand it automatically to 30 sentences. Nevertheless, the Prompt Generator offers two important features: • It prepares a plan needed to perform the assigned task (and this plan becomes part of the prompt), • In this plan, it introduces placeholders to be filled by us, essentially forcing the completion of missing information that we forgot in the original prompt, which enriches the context As a result, we have an additional step that resembles tool-less agentic behavior (or multi-step prompting) with elements of Chain of Thought. For those who deal with actual agents or prompt engineers, this isn't anything new as we do it on a daily basis, but for "classic users" of large language models - it's definitely worth exploring.

A little quality of life upgrade in @ahrefs. A lot of the reports can now match regex. This opens a lot of possibilities. For example here I can quickly get all of our pages for other languages matching the URL pattern, which in our case is a 2 letter language code.

How to Efficiently Detect English Words in Text with Regex?   #TextAnalysis #StringMatching #Regex #DictionarySearch Do you find detecting English words inside a text using a dictionary too slow when the word isn't an exact match? 🤔 Let's brainstorm some solutions together! Here are my thoughts: - Split the text into words and check for exact matches in the English dictionary for O(1) operations. - Consider using an algorithm for string similarity to tac... Source: https://lnkd.in/ghnymzPF #mymetric360

How to Efficiently Detect English Words in Text with Regex?   #TextAnalysis #StringMatching #Regex #DictionarySearch Do you find detecting English words inside a text using a dictionary too slow when the word isn't an exact match? 🤔 Let's brainstorm some solutions together! Here are my thoughts: - Split the text into words and check for exact matches in the English dictionary for O(1) operations. - Consider using an algorithm for string similarity to tac... Source: https://lnkd.in/gfYjgzeb #mymetric360

A little quality of life upgrade in Ahrefs. A lot of the reports can now match regex. This opens a lot of possibilities. For example here I can quickly get all of our pages for other languages matching the URL pattern, which in our case is a 2 letter language code.

How to Efficiently Detect English Words in Text with Regex?   #TextAnalysis #StringMatching #Regex #DictionarySearch Do you find detecting English words inside a text using a dictionary too slow when the word isn't an exact match? 🤔 Let's brainstorm some solutions together! Here are my thoughts: - Split the text into words and check for exact matches in the English dictionary for O(1) operations. - Consider using an algorithm for string similarity to tac... Source: https://lnkd.in/gc_ngDit #mymetric360

How to Efficiently Detect English Words in Text with Regex?   #TextAnalysis #StringMatching #Regex #DictionarySearch Do you find detecting English words inside a text using a dictionary too slow when the word isn't an exact match? 🤔 Let's brainstorm some solutions together! Here are my thoughts: - Split the text into words and check for exact matches in the English dictionary for O(1) operations. - Consider using an algorithm for string similarity to tac... Source: https://lnkd.in/gea5t5mz #mymetric360

I just discovered this powerful regex online resource, you can build, test and debug regex expressions before use them and bonus in a lot of languages: https://regex101.com/