Rewrite Rule Replace issues

[derflasher]

I'm trying to use the replace function in order to eliminate sentences (mostly ads). I'm having trouble with umlaut characters (ä, ö, ü) and , . : ; Sometimes I quote , and it works, mostly it fails. What can I do to improve it?

[ghost]

Miniflux just uses ReplaceAllString from the standard go regex library under the hood in rewrite_functions.go which should have support for unicode. The search term is a regular expression so if you want to match a literal period for example, you'll have to escape it.

[derflasher]

Ok. Thank you for your quick response. How do I quote ü ö ä and sentences with ","?

[ghost]

Can you link an example feed/entry and your rewrite rule where this issue happens? It might be that the website uses html escape codes instead of plain unicode characters, not sure how miniflux handles that

[derflasher]

Yes, sure:

Here the link to the RSS feed source: https://www.tagesspiegel.de/contentexport/feed/home For example some news: https://m.tagesspiegel.de/gesellschaft/panorama/zusammenstoss-von-zwei-kleinbahnen-mehr-als-210-verletzte-bei-zugunglueck-in-kuala-lumpur/27218686.html https://m.tagesspiegel.de/sport/wie-realistisch-ist-der-direkte-wiederaufstieg-werder-bremen-und-das-schlechte-timing/27218310.html

I want to replace this sentence: [Wenn Sie aktuelle Nachrichten aus Berlin, Deutschland und der Welt live auf Ihr Handy haben wollen, empfehlen wir Ihnen unsere App, die Sie hier für Apple- und Android-Geräte herunterladen können.]

This is anther feed source: http://feeds2.feedburner.com/stadt-bremerhaven/dqXM Example news: https://stadt-bremerhaven.de/apple-tv-4k-2021-und-sky-q-derzeit-viele-beschwerden/

And I want to replace the sentence: In diesem Artikel sind Amazon-Links enthalten. Durch einen Klick darauf ge­lan­gt ihr direkt zum Anbieter. Solltet ihr euch dort für einen Kauf entscheiden, erhalten wir ei­ne kleine Provision. Für euch ändert sich am Preis nichts.

Hope that helps you and me. Thank you so much for your help!

[ghost]

I made your tagesspiegel example work with this: replace("\[Wenn Sie alle aktuellen Nachrichten live auf Ihr Handy haben wollen. empfehlen wir Ihnen unsere App. die Sie.*herunterladen können\.\]"|"") (I had a slightly different ad text). You'll have to search the entry html for the part you want to replace and then you can use something like https://regex101.com set to Golang mode to test your regex. Commas are indeed a bit weird, there is currently no way to have a comma in your search or replacement string because miniflux thinks it is a delimiter for a new rewrite rule (source). You can ignore them with regex periods for now. This should probably be improved in the future. I think you should be able to do the second example yourself as an exercise :smile:

[derflasher]

Thank you so much!

Exercise done: replace("In diesem Artikel sind.*Preis nichts\."|"")

[Tokariew]

With one of recent updates my rule for catching youtube urls in iframes from https://www.angrymetalguy.com/feed/ broke My rule was as follow:

replace("<iframe.*(https:\/\/.*youtub.*?)".*?<\/iframe><\/p>"|"<a href=$1>$1</a>")

is it possible to escape "

[derflasher]

@Tokariew, I have the same problem since a couple of releases. The tagesspiegel rewrite rule replace("\[.*\]"|"") is not working at all. Does anybody know how to fix it?

[Tokariew]

@derflasher No, Idea, Tried to escape the " with \ but with no effect… In logs i only get <input>:1:9: invalid char escape

Unless someone found solution, or it's fixed, I will stay with pinned miniflux to version 2.0.32

[Tokariew]

@derflasher i was able to rewrite my rules using . to match " in text but it seems that replace don't use from 2.0.32 full regex. Using \ to escape special characters or use to match for example \s is not possible

[derflasher]

Yes, @Tokariew, it still does not replace full regex 😒 I always stick to the latest version, so I'll text you one day 😂