NEW: custom FORMATS

[xfra35]

This is a proposal to allow users to define custom formats.

In particular, it would help fixing bcosca/fatfree#970 by providing a means to define plural schemes for languages other than English:

// custom format definition
$f3->set('FORMATS.polish',function($nb,$mod){
  preg_match_all('/(\d)\s*\{\s*(.+?)\s*\}/',$mod,$matches,PREG_SET_ORDER);
  $terms=[];
  foreach ($matches as $m)
    $terms[$m[1]]=$m[2];
  $k=$nb==1?1:(preg_match('/(?<!1)[234]$/',$nb)?2:5);
  return isset($terms[$k]) ? str_replace('#',$nb,$terms[$k]) : $nb;
});

// usage
$line='{0, polish, 1 {# linia}, 2 {# linie}, 5 {# linii}}';
echo $f3->format($line,1); // 1 linia
echo $f3->format($line,2); // 2 linie
echo $f3->format($line,5); // 2 linii
echo $f3->format($line,12); // 12 linii
echo $f3->format($line,21); // 21 linii
echo $f3->format($line,22); // 22 linie

[SylwesterZarebski]

Thanks a lot, but i'm afraid i do not understand how to use in dictionary files and then in views. Could You give some example how to use with templating engine like {{ @phrase }} or in code $f3->get('phrase');?

[xfra35]

Let's say you are running an English-Polish website.

You would declare two dictionary files:

;en.ini
score = Score: {0, plural, zero {0}, one {1 point}, other {# points}}

;pl.ini
score = Wynik: {0, polish, 1 {1 punkt}, 2 {# punkty}, 5 {# punktów}}

Then in your template:

{{ @score, 1202 | format }}

Expected output:

• English: 1202 points
• Polish: 1202 punkty

[SylwesterZarebski]

Thanks, i'll try this soon.

[ikkez]

hm, do you think that it would be possible to use a regex in the plural statement instead of defining all those custom lang format handlers?! I'm thinking about something like:

;en.ini
score = Score: {0, plural, zero {0}, one {1 point}, other {# points}}

;pl.ini
score = Wynik: {0, plural, one {1 punkt}, #(?<!1)[2|3|4]$# {# punkty}, other {# punktów}}

which transforms the regex part between # # into \d*(?<!1)[2|3|4]$

[xfra35]

I'm not sure that's a good idea, for at least 3 reasons:

• the dictionary file is meant to be filled up by a translator, usually not a techie
• the regex would be the same everywhere, so why defining it 100 times? It just increases the probability of mistakes
• there's a probability that some other languages require some even more complex plural forms that can't easily be handled by a regex

Also the idea of this PR is to provide the ability to define any custom format, not only pluralization formats. I don't have any practical use case at the moment, but I'm sure it may arise some time ^^

NB: Nice regex 👍 That makes the code for Polish plural a bit more concise. I'll update my example.

[ikkez]

Good points. alright I'm with you.. I could imagine a format for humanized "time ago" labels could also be made with this. it's a good and small addition 👍

[eazuka]

@ikkez @bcosca would this PR be merged?

[xfra35]

We're waiting for @SylwesterZarebski's feedback on this.

[ikkez]

works great. here is a little sample to format like time ago:

$f3->set('FORMATS.timeago',function($nb,$mod){
        if (is_string($nb))
            $nb=strtotime($nb);
        $nb=time()-$nb;
        preg_match_all('/(\w+)\s*\{\s*(.+?)\s*\}/',$mod,$matches,PREG_SET_ORDER);
        $terms=[];
        $gaps=[
            's'=>0,
            'm'=>60,
            'mm'=>60*2,
            'h'=>60*60,
            'hh'=>60*60*2,
            'd'=>60*60*24,
            'dd'=>60*60*24*2,
            'M'=>60*60*24*30,
            'MM'=>60*60*24*30*2,
            'Y'=>60*60*24*30*12,
            'YY'=>60*60*24*30*12*2,
        ];
        foreach ($matches as $m)
            $terms[$m[1]]=$m[2];
        $k='s';
        foreach ($gaps as $gk=>$v)
            if ($nb>$v)
                $k=$gk;
        if (strlen($k)>1)
            $nb=round($nb/$gaps[$k[0]]);
        return isset($terms[$k]) ? str_replace('#',$nb,$terms[$k]) : $nb;
    });

    $posted='Comment was posted {0, timeago, 
        s {just now}, 
        m {a minute ago}, 
        mm {# minutes ago}, 
        h {an hour ago}, 
        hh {# hours ago}, 
        d {yesterday}
        dd {# days ago}
        M {last month}
        MM {# months ago}
        Y {last year}
        YY {# years ago}
    }';
    echo $f3->format($posted,strtotime('-30 seconds'))."\n"; // just now
    echo $f3->format($posted,strtotime('-65 seconds'))."\n"; // a minute ago
    echo $f3->format($posted,strtotime('-300 seconds'))."\n"; // 5 minutes ago
    echo $f3->format($posted,'25.10.2016')."\n"; // 8 days ago
    echo $f3->format($posted,'10.06.2016')."\n"; // 5 months ago
    echo $f3->format($posted,'02.08.2015')."\n"; // last year

[eazuka]

Nice

[xfra35]

@ikkez great 👍

Maybe we could hold a small repository in F3Community with various useful formats, such as this one and some non-English plurals.

[eazuka]

good idea @xfra35

[ikkez]

@xfra35 sure, feel free to create one. Or maybe it's something for the wiki page?

[SylwesterZarebski]

It works and translation surely needs comprehensive Wiki page - just spend few hours to discover PREFIX variable needs DOT at end (of course it is in docs, but looks like end of the sentence dot).

Sorry for big delay, but multilingual project was abandoned and resurrected just now.