ISO-8859-5 (cyrillic) and TextIndexNG2
Hi there, wondering if you might be able to help me? I've got plain text documents that contain iso-8859-5 chracters (not utf-8) and I'd like to get TextIndexNG2 to catalog their contents. I've tried creating an Index with default encoding set to iso-8859-5 but no joy, the cataloger won't catalog the docs! I've also tried setting the Zope locale to russian, but again no good. Any pointers greatly appreciated! Thanks, Mark
--On 10. Januar 2007 17:10:03 +0000 mark hellewell <mark.hellewell@gmail.com> wrote:
Hi there, wondering if you might be able to help me?
I've got plain text documents that contain iso-8859-5 chracters (not utf-8) and I'd like to get TextIndexNG2 to catalog their contents.
We have TXNG V3.1.13! V2 is no longer supported
I've tried creating an Index with default encoding set to iso-8859-5 but no joy, the cataloger won't catalog the docs!
"won't catalog" means what? Provide a detailed description what you are doing and how to reproduce the behavior.
I've also tried setting the Zope locale to russian, but again no good.
Locales don't affect TXNG in any way. -aj
Hi On 1/10/07, Andreas Jung <lists@zopyx.com> wrote:
--On 10. Januar 2007 17:10:03 +0000 mark hellewell <mark.hellewell@gmail.com> wrote:
I've tried creating an Index with default encoding set to iso-8859-5 but no joy, the cataloger won't catalog the docs!
"won't catalog" means what? Provide a detailed description what you are doing and how to reproduce the behavior.
Inside my CMF-based sites portal_catalog I add a TextIndexNG2 index, specifying the Default encoding as iso-8859-5. I then upload my text document containing the characters, and click Reindex through the ZMI. If I then look at the catalog entry for that document it shows no data was gathered for the Index in question. The same process works ok with UTF-8 documents and an Index with a default encoding of UTF-8.
Locales don't affect TXNG in any way.
Ok, thanks :) mark
-aj
Just a quick follow up to my question yesterday: should this work in principle? Cheers, Mark On 1/10/07, mark hellewell <mark.hellewell@gmail.com> wrote:
Hi
On 1/10/07, Andreas Jung <lists@zopyx.com> wrote:
--On 10. Januar 2007 17:10:03 +0000 mark hellewell <mark.hellewell@gmail.com> wrote:
I've tried creating an Index with default encoding set to iso-8859-5 but no joy, the cataloger won't catalog the docs!
"won't catalog" means what? Provide a detailed description what you are doing and how to reproduce the behavior.
Inside my CMF-based sites portal_catalog I add a TextIndexNG2 index, specifying the Default encoding as iso-8859-5.
I then upload my text document containing the characters, and click Reindex through the ZMI.
If I then look at the catalog entry for that document it shows no data was gathered for the Index in question.
The same process works ok with UTF-8 documents and an Index with a default encoding of UTF-8.
Locales don't affect TXNG in any way.
Ok, thanks :)
mark
-aj
--On 11. Januar 2007 14:00:39 +0000 mark hellewell <mark.hellewell@gmail.com> wrote:
Just a quick follow up to my question yesterday: should this work in principle?
Of course it must work in principle. However as I said: V2 is no longer maintainted (and I don't care about V2 related issues). -aj
mark hellewell wrote at 2007-1-10 17:25 +0000:
... Inside my CMF-based sites portal_catalog I add a TextIndexNG2 index, specifying the Default encoding as iso-8859-5.
I then upload my text document containing the characters, and click Reindex through the ZMI.
If I then look at the catalog entry for that document it shows no data was gathered for the Index in question.
In order for an index to determine a given object's value (to be indexed), the object needs to provide some attribute or method (that the index can interprete as the object's value). By default, the index name is used as the attribute/method to look for. If the object does not provide such an attribute/method, then the index will interpret this as "nothing to index for this value". TextIndexNG allows you to specify alternative attribute/method names (yes, it allows more than a single one and will combine them). What you report is almost surely caused by your object not providing the attribute/method the index expects to obtain the object's value. -- Dieter
participants (3)
-
Andreas Jung -
Dieter Maurer -
mark hellewell