collective.dexteritytextindexer 2.3.0

Dynamic SearchableText index for dexterity content types

Introduction

collective.dexteritytextindexer provides a dynamic SearchableText indexer for dexterity content types. It makes it possible to index fields of multiple behaviors as SearchableText.

Usage

For enabling the indexer just add the behavior to the list of behaviors of your content types.

In your profiles/default/types/YOURTYPE.xml add the behavior:

<?xml version="1.0"?>
<object name="example.conference.presenter" meta_type="Dexterity FTI"
   xmlns:i18n="http://xml.zope.org/namespaces/i18n"
   i18n:domain="example.conference">

 <!-- enabled behaviors -->
 <property name="behaviors">
     <element value="collective.dexteritytextindexer.behavior.IDexterityTextIndexer" />
 </property>

</object>

Now you need to mark the fields you want to have in your SearchableText. This is done with directives:

from collective import dexteritytextindexer
from plone.autoform.interfaces import IFormFieldProvider
from plone.supermodel.model import Schema
from zope import schema
from zope.interface import alsoProvides

class IMyBehavior(Schema):

    dexteritytextindexer.searchable('specialfield')
    specialfield = schema.Text(title=u'Special field')

alsoProvides(IMyBehavior, IFormFieldProvider)

If you want to mark fields of an existing 3rd party behavior, it can be done using this utility function:

from plone.app.dexterity.behaviors.metadata import ICategorization
from collective.dexteritytextindexer.utils import searchable

searchable(ICategorization, 'categorization')

The title and description on plone.app.dexterity’s IBasic behavior are marked as searchable by default. For marking them as no longer searchable, there is a utility function:

from plone.app.dexterity.behaviors.metadata import IBasic
from collective.dexteritytextindexer.utils import no_longer_searchable

no_longer_searchable(IBasic, 'title')

Alternatively, if you specified your model as a plone.supermodel XML model, you can mark the field searchable that way:

<model xmlns="http://namespaces.plone.org/supermodel/schema"
       xmlns:indexer="http://namespaces.plone.org/supermodel/indexer">
  <schema based-on="plone.supermodel.model.Schema">

      <field name="specialfield" type="zope.schema.TextLine"
             indexer:searchable="true">
        <title>Special field</title>
      </field>

  </schema>
</model>

Your SearchableText indexer includes now your custom field on your behavior, as soon you enable it in your content type, where IDexterityTextIndexer behavior is enabled too.

Registering a custom field converter

By default, a field is converted to a searchable text by rendering the widget in display mode and transforming the result to text/plain. However, if you need to convert your custom field in a different way, you only have to provide a more specific converter multi-adapter.

Convert multi-adapter specification:

Interface:

collective.dexteritytextindexer.IDexterityTextIndexFieldConverter

Discriminators:

context, field, widget

Example:

from collective.dexteritytextindexer.converters import DefaultDexterityTextIndexFieldConverter
from collective.dexteritytextindexer.interfaces import IDexterityTextIndexFieldConverter
from my.package.interfaces import IMyFancyField
from plone.dexterity.interfaces import IDexterityContent
from z3c.form.interfaces import IWidget
from zope.component import adapts
from zope.interface import implements

class CustomFieldConverter(DefaultDexterityTextIndexFieldConverter):
    implements(IDexterityTextIndexFieldConverter)
    adapts(IDexterityContent, IMyFancyField, IWidget)

    def convert(self):
         # implement your custom converter
         # which returns a string at the end
         return ''

ZCML:

<configure xmlns="http://namespaces.zope.org/zope">

    <adapter factory=".converters.CustomFieldConverter" />

</configure>

There is already an adapter for converting NamedFiles properly. It’s registered only if plone.namedfile is installed.

Extending indexed data

Sometimes you need to extend the SearchableText with additional data which is not stored in a field. It’s possible to register a named adapter which provides additional data:

from collective import dexteritytextindexer
from zope.component import adapts
from zope.interface import implements

class MySearchableTextExtender(object):
    adapts(IMyBehavior)
    implements(dexteritytextindexer.IDynamicTextIndexExtender)

    def __init__(self, context):
        self.context = context

    def __call__(self):
        """Extend the searchable text with a custom string"""
        return 'some more searchable words'

ZCML:

<configure xmlns="http://namespaces.zope.org/zope">

    <adapter factory=".indexer.MySearchableTextExtender"
             name="IMyBehavior"
             />

</configure>

This is a named adapter! This makes it possible to register multiple extenders for the same object on different behavior interfaces. The name of the adapter does not matter, but it’s recommended to use the name of the behavior (this may reduce conflicts).

If your behavior has a defined factory (which is not attribute storage), then you need to define a marker interface and register the adapter on this marker interface (dexterity objects do not provide behavior interfaces of behaviors, which are not using attribute storage).

Contributors

(In order of appearance)

• Jonas Baumann

• Philippe Gross

• Lukas Graf

• Izhar Firdaus

• Sune Broendum Woeller

• Nejc Zupan

Changelog

2.3.0 (2019-09-27)

• A few updates to make build pass on travis again, i.e. add index/allow-hosts to buildout configs, set version.cfg reference to ‘latest’ for 5.0.x, bump setuptools and zc.buildout versions [staeff]

• Update the Add-On to also run in Plone 5.2 with Python 3. Fixes https://github.com/collective/collective.dexteritytextindexer/issues/31 [staeff]

2.2.1 (2017-07-26)

• Fix a bug in the RichText field indexer when the value of RichText field is None. The bug was introduced in 2.2.0. [mbaechtold]

2.2.0 (2017-06-23)

• Added TupleFieldConverter, to enable indexing of portal catalog KeywordIndex indices. [jone, mtrebron, maurits]

• Improved code quality. [maurits]

• Add the ability to index a RichText field. Fixes #23. [smcmahon]

2.1.1 (2016-11-03)

• Remove traces of plone.directives.form (which implicitly added grok as a dependency). [gforcada]

• Silent a plone.behavior warning. [gforcada]

• Specify compatibility in setup.py for versions which are tested and remove unittest2 dependency [tomgross]

2.1.0 (2016-04-14)

• Add schemaeditor support [datakurre]

2.0.2 (2016-04-07)

• Handle indexing from scripts. Fixes https://github.com/collective/collective.dexteritytextindexer/issues/12 [gforcada]

2.0.1 (2014-01-02)

• Fix encoding error when transforming HTML to text. [jone]

2.0 (2013-03-16)

• Drop Plone 4.1 support. [jone]

• Mark title and description of p.a.dexterity’s IBasic as searchable by default. [jone]

• Add a no_longer_searchable utility function. [jone]

• Plone 4.3 support. [jone]

• Eliminate grok / martian dependencies in favor of the new plone.supermodel directives. #5 [jone]

1.5.1 (2013-02-20)

• Do not try to convert text/plain in files, just index it as it is. [zupo]

1.5 (2012-08-16)

• Fix missing field bug. #3

  • Log an error when indexing an object and one of its schemas defines a missing field as searchable.

  • Make sure that indexing other existing fields of the same schema works.

  • searchable utils function: raise AttributeError when field is missing.

  [jone]

• Added support for marking fields searchable in plone.supermodel XML models. This is done by implementing a IFieldMetadataHandler that is capable of serializing/de-serializing the corresponding taggedValue to/from XML. [lgraf]

• Add utils.searchable method for marking fields of third party schemas as searchable. [kagesenshi]

1.4.1 (2011-11-17)

• ignore the request in the get_field_widget method, to avoid problems with request variables wich have the same name than the field. [phgross]

• Added test-buildout for plone-4.1.x [lgraf]

1.4 (2011-08-24)

• Added IntFieldConverter, wich return the plain value instead of the render method (600000 –> 600,000) [phgross]

1.3

• Fixed querying of tagged values: use helper function mergedTaggedValueList - which also looks up tagged values on superclasses. [jbaumann]

• Fixed html to text transform call: added source mimetype. [jbaumann]

1.2

• Fixed data transforms in NamedFileConverter [lgraf]

1.1

• Made reindexer more robust, since sometimes the field values may be wrong. [jbaumann]

• Do not traverse to “view” in indexer, this could cause security issues especially in tests. Using now a fake-view for enabling z2 mode. [jbaumann]

1.0

• Fixed assertion bug when using a IDynamicTextIndexExtender adapter. [jbaumann]

1.0b3

• Moved IDynamicTextIndexExtender to interfaces module. [jbaumann]

• The plone.namedfile is now optional. The new namedfile converting adapter is only registered if its present [jbaumann]

• Re-implemented converting of field data with an newly introduced adapter. The default converter just converts the widget in display mode and transforms it to text/plain. [jbaumann]

• Fixed tests for compaitbility with plone.app.testing 4.0a3: Use TEST_USER_ID instead of TEST_USER_NAME [jbaumann]

• fixed Bug UnicodeError: while indexing lists or dicts with special chars (Non-Ascii characters) [phgross]

1.0b2

• Fixed MANIFEST.in [jbaumann]

1.0b1

• Initial release