Details
Bug
Major
Description
Hello,
I am used to the searchPage method of the DocumentSearchController and it generally works fine.
Unfortunately, it fails in the following case:
Searching for "Article 45 of the Constitution" returns 0 match.
I tried using the API (I.e DocumentSearchController.searchPage) and via the Search tab of the viewer.
This term exists in the document and Acrobat Reader found several occurrences.
In order to be sure that it is not an issue with the cmap, I checked the cids and unics of a page containing this pattern (p 239)
and they seem correct:
cid=0x41:unic=A:0x41
cid=0x72:unic=r:0x72
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x63:unic=c:0x63
cid=0x6c:unic=l:0x6c
cid=0x65:unic=e:0x65
cid=0x20:unic= :0x20
cid=0x34:unic=4:0x34
cid=0x35:unic=5:0x35
cid=0x20:unic= :0x20
cid=0x6f:unic=o:0x6f
cid=0x66:unic=f:0x66
cid=0x20:unic= :0x20
cid=0x74:unic=t:0x74
cid=0x68:unic=h:0x68
cid=0x65:unic=e:0x65
cid=0x20:unic= :0x20
cid=0x43:unic=C:0x43
cid=0x6f:unic=o:0x6f
cid=0x6e:unic=n:0x6e
cid=0x73:unic=s:0x73
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x74:unic=t:0x74
cid=0x75:unic=u:0x75
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x6f:unic=o:0x6f
cid=0x6e:unic=n:0x6e
Moreover, this term is not split over multiple lines.
The pdf is accessible on http://www.itu.int/dms_pub/itu-s/oth/02/02/S02020000244501PDFE.pdf
I am really puzzled, thank you very much for your help.
I am used to the searchPage method of the DocumentSearchController and it generally works fine.
Unfortunately, it fails in the following case:
Searching for "Article 45 of the Constitution" returns 0 match.
I tried using the API (I.e DocumentSearchController.searchPage) and via the Search tab of the viewer.
This term exists in the document and Acrobat Reader found several occurrences.
In order to be sure that it is not an issue with the cmap, I checked the cids and unics of a page containing this pattern (p 239)
and they seem correct:
cid=0x41:unic=A:0x41
cid=0x72:unic=r:0x72
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x63:unic=c:0x63
cid=0x6c:unic=l:0x6c
cid=0x65:unic=e:0x65
cid=0x20:unic= :0x20
cid=0x34:unic=4:0x34
cid=0x35:unic=5:0x35
cid=0x20:unic= :0x20
cid=0x6f:unic=o:0x6f
cid=0x66:unic=f:0x66
cid=0x20:unic= :0x20
cid=0x74:unic=t:0x74
cid=0x68:unic=h:0x68
cid=0x65:unic=e:0x65
cid=0x20:unic= :0x20
cid=0x43:unic=C:0x43
cid=0x6f:unic=o:0x6f
cid=0x6e:unic=n:0x6e
cid=0x73:unic=s:0x73
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x74:unic=t:0x74
cid=0x75:unic=u:0x75
cid=0x74:unic=t:0x74
cid=0x69:unic=i:0x69
cid=0x6f:unic=o:0x6f
cid=0x6e:unic=n:0x6e
Moreover, this term is not split over multiple lines.
The pdf is accessible on http://www.itu.int/dms_pub/itu-s/oth/02/02/S02020000244501PDFE.pdf
I am really puzzled, thank you very much for your help.
Activity
- All
- Comments
- History
- Activity
- Remote Attachments
- Subversion
The issue is related to how the DocumentSearchControllerImpl is splitting up the words in the search phrase. A change was made as part of PDF-745 that introduced the issue but there is no reference in the JIRA as to what the new code was for. Removing code for now as it no behaves more like text work detection algorithm used for text extraction.
Show
Patrick Corless
added a comment - The issue is related to how the DocumentSearchControllerImpl is splitting up the words in the search phrase. A change was made as part of PDF-745 that introduced the issue but there is no reference in the JIRA as to what the new code was for. Removing code for now as it no behaves more like text work detection algorithm used for text extraction.
Thank you very much for your quick fix, I am eager to use it.
Show
Olivier Chuzel
added a comment - Thank you very much for your quick fix, I am eager to use it.
"Article 45 of the Constitution" is found with version 5.1.2 but there's a regression since 'No. 5.33' is not found anymore. But searching 'No. 5.' returns occurrences of No. 5.33
Shall I create a new issue?
Show
Olivier Chuzel
added a comment - "Article 45 of the Constitution" is found with version 5.1.2 but there's a regression since 'No. 5.33' is not found anymore. But searching 'No. 5.' returns occurrences of No. 5.33
Shall I create a new issue?
Thanks for posting this issue. We're currently working with a customer on improving our text extraction ordering. In the document in question an extra space is being added to work. A fix for this should soon be available.