[greenstone-users] Re: Hierarchy classifier - sorting

From William Mann
DateMon Nov 22 06:23:13 2010
Subject [greenstone-users] Re: Hierarchy classifier - sorting
In-Reply-To (2884-125-239-211-134-1290275368-squirrel-webmail-scms-waikato-ac-nz)
Katherine,

I tried the hfile technique but I could never get it to show my
documents. I could get the structure working perfectly, but drilling
down to the document level only gave me a leaf that I couldn't click.
I'm thinking that the problem is that my documents are being handled by
the PagedImage plugin. I'll have another go at it though to see if I can
get it working: that would definitely solve my problems (unless I have
to indicate all 40000+ documents in the hfile!). I'll let you know how
that works out.

--
William

On 11/20/2010 06:49 PM, Katherine Don wrote:
> Hi William
>
> Did you not get hfile working for hierarchy classifier?
> If you don't use an hfile, and just use the | separator, then it will do
> alphabetical sorting. It can't do custom sorting as how would it know what
> order you want?
>
> You can do custom sorting using the hfile.
>
> eg hfile like:
>
> pets 1 Pets
> dog 1.1 Dogs
> cat 1.2 Cats
> wild 2 Wild Animals
> tiger 2.1 Tigers
> aardvark 2.2 Aardvarks
>
> Would give a structure like
>
> Pets
> - Dogs
> - Cats
> Wild Animals
> - Tigers
> - Aardvarks
>
> You need to give your documents metadata values like 'pets' or 'dogs' etc.
>
> The first entry in each line of the hfile is the descriptor, which is the
> value that needs to be set as metadata in the documents. Depending on how
> you want to add metadata, you can change what you use for the descriptor.
>
> eg could do
>
> 1 1 Pets
> 1.1 1.1 Dogs
> 1.2 1.2 Cats
>
> Then you will be assigning metadata 1, 1.1, 1.2 etc
>
> Alternatively, you can encode the whole structure in the metadata.
>
> Pets 1 Pets
> Pets|Dogs 1.1 Dogs
> Pets|Cats 1.2 Cats
>
> Then you will be assigning metadata like Pets, Pets|Dogs etc.
> This may be what you already have? So create a hfile that gives the
> correct ordering and you should be fine.
>
> I hope this helps,
> Katherine
>
>> Hi Katherine,
>>
>> Thanks for all the help you've given me. I've started refreshing my Perl
>> and C++ so I can take a look at the source code and maybe start
>> contributing to the project.
>>
>> That said I have one last question (at least for now): where would I
>> need to intervene to make my hierarchy classifier output it's values in
>> an order specified by me (keep in mind that I'm using this with the
>> PagedImage plugin). Let's say I've set up my .item file like this:
>>
>> <Title>Whatever
>> <Subject and Keywords>01_test|02_about|03_zesty
>> 1:file.jpg
>> 2:file2.jpg
>> ....
>>
>> and
>>
>> <Title>Try Again
>> <Subject and Keywords>02_about|01_test|02_testy
>> 1:file.jpg
>> 2:file2.jpg
>> ...
>>
>> Now when I look at the classifier on the site I get this:
>>
>> 01_test
>> 02_about
>>
>> which is the order that I need. The problem is that this corresponds to
>> the directory structure and is not very user friendly. However if I put
>> things the way I need them:
>>
>> <Title>Whatever
>> <Subject and Keywords>Test|About|Zesty
>> 1:file.jpg
>> 2:file2.jpg
>> ....
>>
>> and
>>
>> <Title>Try Again
>> <Subject and Keywords>About|Test|Testy
>> 1:file.jpg
>> 2:file2.jpg
>> ...
>>
>> Now with the classifier I get this:
>>
>> About
>> This
>>
>> which is not the order I want. What I'd like is the same order as before
>> (like my directory structure) and that is:
>>
>> This
>> About
>>
>> It seems like the hierarchy classifier puts everything in alphabetical
>> order no matter what I do. Is it possible for me to change some code or
>> something to keep the order the same as the structure I have on disk?
>>
>> Thanks again for all the help and have a nice weekend.
>>
>> --
>> William
>>
>>
>>
>