 Keyword Filter Tip: Solution: 100 comments
 @burgatshow @ivory I wasn't sure if modifiers were supported here, since the GUI is a bit misleading. But if so, better use /i @ivory One of my favourite sites is RegExper, which visualises regular expressions. Useful if they are more complex. @ivory which flavor should we pick in the left column to match how Ivory will interpret the regex? @swizzlevixen For basic pattern matching, I didn't change the flavor. I am not a regex expert. Just wanted to provide a basic solution to a very common problem people have with keyword filters. 🙂 If you want to know more about what we use, you can go here: https://unicode-org.github.io/icu/userguide/strings/regexp.html @ivory oh ffs you couldn’t just wrap the keywords in ‘\b…\b’ like everyone else does and all your users expect from decades of Google and other full-text search engines @ivory \byouknowwho\b gets initial Youknowwho and when punctuation follows. @ivory As someone who has been a software developer for over 30 years, I can definitely state that regexes are the work of satan @ivory this looks great, although couldn’t a shortcut option be made available for people who don’t know what regex is? (E.g. match whole word only option?) @mporcheron We plan to revamp our keyword filter creation tools at some point. But for now this works well. 🙂 @ivory What an strangely specific example. 🤣 Keep up the great work, guys. Love what you do.  @ivory OK please tell me you're going to support full PCRE so I can get disproportionately excited by a mobile app  @ivory I love Ivory, but requiring regex knowledge to force matching the exact thing I enter for search is TERRIBLE for new users. You can give users the *option* to use regex, but the *default* should only match exactly what is requested. The “solution” offered is making users fix unexpected geeky behaviour by the app back-end. This is NOT newbie-friendly. @dharrison @ivory The default is for it to match letter-strings as part of larger words. That’s not what regular folks will expect. @michaelgemar Yes we plan to revamp our keyword tools in the future, but this is what we have now. 🙂 We get people asking how to solve this problem on occasion and we wanted to share the current solution. @ivory That’s cool — I know things are happening very fast. I just wanted to flag that the solution offered was less than ideal, and a more friendly long-term fix is desirable. @ivory If the regex should also match when there is no Space before the searchstring (i.e. it is at the beginning of the test), use a regex like this (not tested in ivory):\belon\b @ivory The trouble you get in with a regex like that, though, is that it doesn't block out "Elon." or "#IdiotElon" constructs. Would something like \b(elon)\b Also work to match word boundaries?
[DATA EXPUNGED]
@ivory I guess they are always case insensitive. What regular expressions are supported? @jgrg https://unicode-org.github.io/icu/userguide/strings/regexp.html We intentionally hid the toggle because too many people would probably turn it on without knowing what it is for. @ivory Maybe have an "Advanced" option to reveal it? I'm just worried that I would never have discovered it, and I think it is such a great feature. @ivory oh my, this is exactly what I needed. (With this exact example) 😅 @ivory feature request: Shared Filters/Filter Lists, akin to the Drafts ecosystem. I’m sure the Space Enron filter set would be very popular. @bryanculver Not sure if it's what you are looking for, but you can hold down on a filter in the list and post it: A solution only an opensource hacker could love and only a 13 yr old computer hack would stumble across. "Oh look! If you put a slash and a space in there this other thing shows up and if I do this to the other thing then it does what I wanted in the first place.... VOILAH!!!" Now I'm a power user! Its like intuitively knowing that if you make Mario jump and punch the mushroom, coins will fly out.🤣 @ivory I love the little bit of shade. But in seriousness, thank you for making an awesome app that makes this new place feel more comfortable. @ivory I was today years old when I discovered Ivory's filter interface. I kept looking for it in the top-left menu, didn't find it, and thought it was absent because Ivory is still in development. 😭 @todd_smith That means you found a few more sections as well then? 🙃 We plan on revamping our navigation system soon which will make it easier to find all the hidden sections. @ivory Yup! Another one was "Oh, they DO have lists!" There have been some high-volume people from Twitter that I haven't been following (via bird.makeup) because I was trying to do everything from the Home timeline. I am much happier now. 😊 @rye For sure! We plan to update our whole filter building system at some point. 🙂  @ivory If you add a keyword filter for “Elon”, it will also filter posts with the word "felony," which is entirely correct. 🤭  @ivory have you considered importing the filters from Tweetbot? Right now it's... somewhat difficult to copy them over.  @ivory @egghat a classic 😄 most users wouldn’t expect that to happen (filtering words by parts of it). Advanced users would expect they could do it with a wildcard like „Elon*“, „*elon*“ and so on RegEx always looks like an over engineered thing … and as always: one gives a simple example/solution with three extra chars - it didn’t take long to see a fullblown, respect all edge cases, two side long expression 🤣 (I’m exaggerating a bit, but you may get the point)    @ivory I love this, but maybe something for the non-regex-fluent among us? like a “word" checkbox that adds word boundary stuff to the original version? |
Gregory's Server
Regular Expressions allow you to filter your posts in powerful ways. A good place to test new expressions to make sure they work properly is at https://regex101.com
Have fun!