[
https://issues.apache.org/jira/browse/TIKA-4303?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17912556#comment-17912556
]
ASF GitHub Bot commented on TIKA-4303:
--------------------------------------
sunluman opened a new pull request, #2098:
URL: https://github.com/apache/tika/pull/2098
Fixes https://issues.apache.org/jira/browse/TIKA-4303
The issue of garbled text is caused by
`OneNotePropertyEnum.CachedTitleString` not being correctly parsed. It should
be parsed using `handleRichEditTextUnicode`.
As for why versions `2.7.0` and earlier did not encounter garbled text, I
believe it was due to a previously erroneous line of code:
```java
if (options.getUtf16PropertiesToPrint().contains(propertyValue.propertyId))
```
This line caused the parsing of OneNote files to never append the parsed
content of `OneNotePropertyEnum.ImageFilename, OneNotePropertyEnum.Author, and
OneNotePropertyEnum.CachedTitleString` to the xhtml.
However, when parsing `OneNotePropertyEnum.RichEditTextUnicode`, the logic
for only parsing the latest version’s content was not added. As a result, the
files appeared to be successfully parsed and without garbled text, but in
reality, CachedTitleString was never parsed.
I only fixed the bug in the issue where the title in the uploaded file was
not parsed. During the testing process, I also discovered the following issues:
- Non-rich text content is not checked for the latest version, so when the
content is TextExtendedAscii, it is still parsed repeatedly.
- Dates are not parsed.
- Chinese (or other non-Ascii characters? i'm not sure) characters in the
content are not parsed.
I am not sure whether to create a new issue before proceeding with these
fixes, so these issues have not been addressed in this PR.
> Unable to extract Chinese content in onenote
> --------------------------------------------
>
> Key: TIKA-4303
> URL: https://issues.apache.org/jira/browse/TIKA-4303
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 2.8.0, 2.9.2
> Reporter: lqangi
> Priority: Major
> Attachments: Chinese-notes.one, tika-parsing-chinese-notes-result.png
>
>
> When I tried to extract the contents of onenote file containing Chinese using
> tika, the Chinese part of the file could not be extracted, only the
> non-Chinese content was extracted.
> In addition, some of the extracted content is duplicate, as described in
> [TIKA-3970|https://issues.apache.org/jira/browse/TIKA-3970], it seems to
> extract the historical version of the data along with the extraction, I don't
> know if this issue (TIKA-3970) has been fixed (I see that the code has been
> committed on github, But it doesn't seem to have completely solved the
> problem yet)
> The software versions I use are as follows:
> Tika: 2.8.0
> Onenote: Microsoft® OneNote® LTSC MSO (16.0.14332.20761)
>
> In order to reproduce this problem, just use the 2.8.0 version of Tika App to
> open the attachment "Chinese-Notes.one" and check whether the Chinese content
> in the file is extracted.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
