[
https://issues.apache.org/jira/browse/PDFBOX-1652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15695430#comment-15695430
]
Ryosuke Fujita commented on PDFBOX-1652:
----------------------------------------
Christian is right.
For example, I want to extract Japanese string "グループ" (which means and sounds
"group")、but extracted string is sometimes "グルプー" (which sounds like "grupuu").
And another case is, I want to retrieve "様々な" (which means "variety of") but,
extracted as "様な々" (became meaningless word)
Christian's patch covers these 2 characters, but other modifier letters like
"〲" or "〻"(which also regareded as diacritic in text extraction), we Japanese
sometimes use.
So, I would like TextPosition#isDiacritic() to be implemented as all
diacritical mark(*) return true, and other characters return false;
* https://en.wikipedia.org/wiki/Combining_character
> TextPosition: Japanese alphabetic characters 30fc and 3005 treated as
> diacritics
> --------------------------------------------------------------------------------
>
> Key: PDFBOX-1652
> URL: https://issues.apache.org/jira/browse/PDFBOX-1652
> Project: PDFBox
> Issue Type: Bug
> Components: Text extraction
> Affects Versions: 1.8.1
> Reporter: Christian Kohlschütter
> Labels: PatchAvailable
> Attachments: PDFBOX-1652.patch
>
>
> For the purpose of determining the position in text, the Japanese characters
> U+30fc (KATAKANA-HIRAGANA PROLONGED SOUND MARK) and U+3005 (IDEOGRAPHIC
> ITERATION MARK) are currently regarded "simple" diacritics. Apparently, they
> are fully-fledged characters in terms of text positioning.
> This can have the effect that when extracting text, some characters get
> actually reversed (particularly ーン can get ンー).
> A patch to fix this is attached.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
