ggozad/collective.classification

Content classification/clustering through language processing

collective.classification's Issues

Improvement: for a folder, opt to classify a (non-folder) default item therein

= Use case =

Text at http://www.brighton.ac.uk/centrim/research/projects/mental-models/folder_contents is minimal.

That folder defaults to view a page http://www.brighton.ac.uk/centrim/research/projects/mental-models/overview in which text is substantial.

= Feature (improvement) =

http://www.brighton.ac.uk/centrim/research/projects/mental-models/@@subjectsuggest should:

• detect that there is a default view for the folder
• offer an option, to make suggestions based on that item (not based upon the folder).

A more exotic additional option would be to make suggestions based upon /both/ the folder /and/ its default item.

Confusing status message after reparsing

After you have pressed "Reparse all documents", the status info show "Term extractor trained and NP storage updated. You will need to re-train the classifier as well."

The first sentece is left-over from previous release, and isn't that usefriendly. Maybe it should just say: "Documents reparsed. You will need to re-train the classifier as well."

Feature request: Limit the number of documents to parse and classify

On really huge sites it is not doable to parse all documents.
It would be nice if we had a Int field for limiting the query.

prefer @@suggest-categories to @@subjectsuggest

= Motivations =

1. Plone speak for developers may be 'subjects', but to everyday users of Plone 3.x, the expression is:

    categories

(Historically: the expression was 'keywords', but let's ignore that.)

1. Order the words in a more spoken way, e.g. speaking 'suggest categories' /sounds/ better than speaking 'categories suggest'.
2. Hyphenate, e.g. @@manage-portlets and @@manage-viewlets so for consistency,

    @@suggest-categories

Error when selecting N-gram but no category

As a default no category is selected so when you chose N-gram and press save, you will get the error:

Time 2010/03/02 11:17:05.890 GMT+1
User Name (User Id) soren (soren)
Request URL http://localhost:8080/nordic/@@classifier-settings-controlpanel
Exception Type ValueError
Exception Value concat() expects at least one object!
Traceback (innermost last):

Module plone.postpublicationhook.hook, line 74, in publish
Module ZPublisher.mapply, line 88, in mapply
Module ZPublisher.Publish, line 42, in call_object
Module zope.formlib.form, line 769, in call
Module Products.Five.formlib.formbase, line 55, in update
Module zope.formlib.form, line 750, in update
Module zope.formlib.form, line 594, in success
Module collective.classification.browser.controlpanel, line 153, in save_action
Module nltk.corpus.reader.tagged, line 211, in tagged_sents
Module nltk.corpus.reader.tagged, line 148, in tagged_sents
Module nltk.corpus.reader.util, line 412, in concat
ValueError: concat() expects at least one object!