[SOLVED] Recoll fails to index application/vnd.oasis.opendocument.presentation

Turbocapitalist · 04-29-2018, 11:00 AM

Recoll fails to index OpenDocument Format presentations, and might not be indexing any OpenDocument Format (ODF) files for that matter. The Recoll documentation mentions OpenOffice files, but I suspect that refers to the old format and not ODF. OpenOffice moved to OpenDocument Format ages and ages ago.

Recoll indexes and retrieves any corresponding PDF files produced from the ODF files, just not the original ODF files. So I know that the right directories are being indexed.

What do I need to set to get it to index the normal productivity suite files?

Code:

$ apt-cache policy recoll | head -n 3
recoll:
  Installed: 1.22.4-1
  Candidate: 1.22.4-1

Edit: I should add that what should be prerequisites, unzip and xsltproc, are present.

Code:

$ apt-cache policy xsltproc | head -n 3
xsltproc:
  Installed: 1.1.29-2.1
  Candidate: 1.1.29-2.1

$ apt-cache policy unzip | head -n 3
unzip:
  Installed: 6.0-21
  Candidate: 6.0-21

medoc · 05-02-2018, 07:22 AM

Hi,

Recoll is supposed to index libre/openoffice docs in their current format, but only recoll 1.23 supports the newer 'flat' pure XML format of openoffice files. 1.22 only supports the older 'zipped' format.

If your files are pure xml (just take a look at them), you can retrofit the new handler, or it might be easier to just move to 1.23

If your files are of the zipped format, we need to look at why they are not indexed, please answer to this thread or contact me at jfd@recoll.org

jf

Turbocapitalist · 05-02-2018, 08:00 AM

Thanks.

I checked the ascii-backports repository and it does not have recoll. But it is not indexing the old multi-file OpenDocument Format files either.

These documents are nearly all in the old multi-file zipped format. For example, here's a small one:

Code:

$ unzip -l foo.odt
Archive:  foo.odt
  Length      Date    Time    Name
---------  ---------- -----   ----
       39  2007-12-26 13:52   mimetype
        0  2007-12-26 13:52   Configurations2/statusbar/
        0  2007-12-26 13:52   Configurations2/accelerator/current.xml
        0  2007-12-26 13:52   Configurations2/floater/
        0  2007-12-26 13:52   Configurations2/popupmenu/
        0  2007-12-26 13:52   Configurations2/progressbar/
        0  2007-12-26 13:52   Configurations2/menubar/
        0  2007-12-26 13:52   Configurations2/toolbar/
        0  2007-12-26 13:52   Configurations2/images/Bitmaps/
     6746  2007-12-26 13:52   content.xml
    10321  2007-12-26 13:52   styles.xml
     1024  2007-12-26 13:52   meta.xml
    12826  2007-12-26 13:52   Thumbnails/thumbnail.png
     7592  2007-12-26 13:52   settings.xml
     1866  2007-12-26 13:52   META-INF/manifest.xml
---------                     -------
    40414                     15 files

I know the directories are indexed because some of these were exported to PDF which are in the same directories. The PDF files turn up in recoll's search results.

I tried making a test file in the the new flat format and it was not indexed either when I update the index or rebuild the index from scratch. I also tried exporting as a regular multi-file document but that does not show up either even after rebuilding the index. If I export any files to PDF and update the recoll index, they show up in the searches.

I have LibreOffice 5.2.6 at the moment.

Code:

$  apt-cache policy libreoffice | head -n 3
libreoffice:
  Installed: 1:5.2.7-1+deb9u4
  Candidate: 1:5.2.7-1+deb9u4

LibreOffice 1:6.0.4~rc1-4~bpo9+2 is available in ascii-backports but I doubt that affects this.

medoc · 05-02-2018, 09:39 AM

Please try:

/usr/share/recoll/filters/rclsoff.py /path/to/my/presentation.odp

Turbocapitalist · 05-02-2018, 12:01 PM

It looks like maybe a dependency is not available:

Code:

$ /usr/share/recoll/filters/rclsoff.py foo.odp
RECFILTERROR HELPERNOTFOUND python:libxml2/python:libxslt1

medoc · 05-02-2018, 12:14 PM

Yes, it does

Actually, this was probably shown in the 'missing helpers' menu from the GUI.

Turbocapitalist · 05-02-2018, 12:27 PM

I just checked Recoll menu Help -> Missing Helpers and it pops up with this in the dialog box:

External applications/commands needed for your file types and not found, as stored by the last indexing pass in /home/tc/.recoll/missing:

No helpers found missing

Indeed, ~/.recoll/missing is empty.

In Preferences -> Index Configuration there are no ODF-related file names, mime types, or ignored endings.

Turbocapitalist · 05-02-2018, 12:33 PM

Manually adding python-libxslt1 solved the problem:

Code:

$ sudo apt-get install python-libxslt1
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following additional packages will be installed:
  python-libxml2
The following NEW packages will be installed:
  python-libxml2 python-libxslt1

Now, with the index updated, the ODF files are found.

Code:

$ lsb_release -rd
Description:    Devuan GNU/Linux testing/unstable
Release:        testing/unstable

Thanks!

medoc · 05-03-2018, 12:27 AM

Quote:

Originally Posted by Turbocapitalist

I just checked Recoll menu Help -> Missing Helpers and it pops up with this in the dialog box:

External applications/commands needed for your file types and not found, as stored by the last indexing pass in /home/tc/.recoll/missing:

No helpers found missing

Indeed, ~/.recoll/missing is empty.

In Preferences -> Index Configuration there are no ODF-related file names, mime types, or ignored endings.

This definitely looks like a bug, I'm not too sure what happened ! Anyway, all is well that ends well