Anime "Duration" issue

[sugar]

Hello, wanna report another bug.

Some anime title causes a Notice: Undefined offset: 0 in \src\Atarashii\APIBundle\Parser\AnimeParser.php

Occurred when the anime duration is only hours. i.e; https://myanimelist.net/anime/30346/Doukyuusei_Movie (1 hr.) https://myanimelist.net/anime/18617/Girls_und_Panzer_der_Film (2 hr.)

Anime with duration in hr. min. (i.e. 2 hr. 10 min.) or min. (i.e. 45 min.) are not affected.

Thanks.

Note: i'm using develop branch & API 2.1

[Ratan Dhawtal]

Fix duration parsing issue because of boolean condition

Fixes #119

→ <<cset 98d4389b0daf>>

[Ratan Dhawtal]

Fix duration parsing issue because of boolean condition

Fixes #119

→ <<cset 61477c80710b>>

[sugar]

error occurred (undefined offset 0) on "Unknown" duration.

• https://myanimelist.net/anime/32801/Dungeon_ni_Deai_wo_Motomeru_no_wa_Machigatteiru_Darou_ka_OVA
• https://myanimelist.net/anime/31988/Hibike_Euphonium_2nd_Season

works properly;

• hr.
• hr. min.
• min.

[sugar]

@ratan12 any update ?

[Ratan Dhawtal]

I will take care of it soon.

[sugar]

then, im trying to fix this issue myself.

# Duration:
$extracted = $leftcolumn->filterXPath('//span[text()="Duration:"]');
if ($extracted->count() > 0) {
    $duration = trim(str_replace($extracted->text(), '', $extracted->parents()->text()));
    # match: # hr. # min
    if (strpos($duration, 'hr.') !== false && strpos($duration, 'min.') !== false) {
        preg_match('/(\d+) hr. (\d+) min/', $duration, $matches);
        $animerecord->setDuration((int) $matches[0] * 60 + $matches[2]);
    } 
    # match: # hr.
    else if (strpos($duration, 'hr.') !== false) {
        preg_match('/(\d+) hr./', $duration, $matches);
        $animerecord->setDuration((int) $matches[0] * 60);
    } 
    # match: # min
    else if (strpos($duration, 'min') !== false) {
        preg_match('/(\d+) min/', $duration, $matches);
        $animerecord->setDuration((int) $matches[0]);
    } 
    # fallback (i.e. unknown)
    else {
        $animerecord->setDuration(0);
    }
}

using !== false due strpos() has mixed return values and return false (boolean) if the needle was not found.

[Ratan Dhawtal]

Your code does not match the codes in our repo. Did you modified it? In the code we do not have !==false.

[sugar]

yes.

• 1st, i change the comparasion operator to not identical to prevent strpos() return value issue (0 treated as false, 1 as true)
• 2nd, adding strpos($duration, 'min') to match duration in min
• 3rd, adding fallback by using else and set the duration with 0 (zero, integer)

[Michael Johnson]

Re-opening so we can make a better fix.

[Michael Johnson]

If we don't know the length, we shouldn't be returning a value. Returning 0 means that the length is 0, not "unknown" or something we couldn't parse.

I'm going to be committing a better fix that will handle hr, hr+min, and min situations.

[sugar]

okay.

i'm using fallback for easier data handling.

also, please add sec. https://myanimelist.net/anime/29347/20-dai_no_Heya-hen = 15 sec.

[Michael Johnson]

Please add the seconds request as a new issue. The current API defines duration as an integer in minutes, so the API would need to be changed, which means an incompatible (breaking) change. For now, we won't return a value if the duration is under one minute.

You are welcome to modify the API for your needs. That's why the source is open. We won't normally accept a change where we return data when there isn't an existing value, however. The only exception is when we're already doing that because of legacy behavior.

[sugar]

new issue created (marked as proposal).

[Michael Johnson]

Fix Duration Parsing

Adjust the duration parser to handle all major styles including hour/minute, hour, and minute durations. Don’t break on “unknown” and unrecognized formats.

Fixes #119

→ <<cset c45ee61c1676>>

[Michael Johnson]

Fix Duration Parsing

Adjust the duration parser to handle all major styles including hour/minute, hour, and minute durations. Don’t break on “unknown” and unrecognized formats.

Fixes #119

→ <<cset 13aed2576400>>