rer.solrpush 1.3.3

Prodotto per Regione Emilia-Romagna relativo all'indicizzazione dei contenuti con solr

Product that allows SOLR indexing/searching of a Plone website.

SOLR schema configuration

This product works with some assumptions and SOLR schema need to have some particular configuration.

You can see an example in config folder of this product.

By default we mapped all base Plone indexes/metadata into SOLR, plus some additional fields:

<field name="searchwords" type="string" indexed="true" stored="true" required="false" multiValued="true" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="showinsearch" type="boolean" indexed="true" stored="false" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="url" type="string" indexed="false" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="site_name" type="string" indexed="true" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="path_depth" type="pint" indexed="false" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="path_parents" type="string" indexed="true" stored="true" multiValued="true"/>
<field name="view_name" type="string" indexed="true" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="@id" type="string" indexed="false" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="@type" type="string" indexed="false" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>
<field name="title" type="text_it" indexed="false" stored="true" required="false" multiValued="false" termVectors="false" termPositions="false" termOffsets="false"/>

• searchwords, view_name, path_parents, path_depth, site_name are needed for query filter and boost (see below)

• showinsearch is needed to allow/disallow single content indexing

• url is an index where we store frontend url

• @id, @type and title are needed for plone.restapi-like responses

plone.restapi related metadata are not indexed from Plone, but they are copied in SOLR:

<copyField source="Title" dest="title"/>
<copyField source="portal_type" dest="@type"/>
<copyField source="url" dest="@id"/>

Control Panel

• Active: flag to enable/disable SOLR integration

• Solr URL: SOLR core url

• Portal types to index in SOLR

• Public frontend url

Hidden registry fields

There are some “service” registry fields hidden to disallow users to edit them.

• ready: a flag that specifies if the product is ready/initialized. It basically indicates that schema.xml has been loaded.

• index_fields: is the list of SOLR fields loaded from schema.xml file.

schema.xml load

SOLR fields are directly read from schema.xml file exposed by SOLR.

This schema is stored in Plone registry for performance reasons and is always synced when you save solr-controlpanel form or click on Reload schema.xml button.

File indexing

If Tika is configured on SOLR, you can send attachments to it and they will be indexed as SearchableText in the content.

To allow attachments indexing, you need to register an adapter for each content-type that you need to index.

File content-type is already registered, so you can copy from that:

<adapter
  for="plone.app.contenttypes.interfaces.IFile"
  provides="rer.solrpush.interfaces.adapter.IExtractFileFromTika"
  factory=".file.FileExtractor"
  />

from rer.solrpush.interfaces.adapter import IExtractFileFromTika
from zope.interface import implementer


@implementer(IExtractFileFromTika)
class FileExtractor(object):
    def __init__(self, context):
        self.context = context

    def get_file_to_index(self):
        """
        """
        here you need to return the file that need to be indexed

N.B.: SearchableText index should be multivalued.

Search configuration

In solr controlpanel (/@@solrpush-settings) there are some field that allows admins to setup some query parameters.

‘qf’ specifies a list of fields, each of which is assigned a boost factor to increase or decrease that particular field’s relevance in the query.

For example if you want to give more relevance to results that contains searched text into their title than in the text, you could set something like this:

title^1000.0 SearchableText^1.0 description^500.0

You can also elevate by searchwords.

bq specifies an additional, optional, query clause that will be added to the user’s main query to influence the score. For example if we want to boost results that have a specific searchwords term:

searchwords:something^1000

Solr will improve ranking for results that have “something” in their searchwords field.

bf specifies functions (with optional boosts) that will be used to construct FunctionQueries which will be added to the user’s main query as optional clauses that will influence the score. Any function supported natively by Solr can be used, along with a boost value. For example if we want to give less relevance to items deeper in the tree we can set something like this:

recip(path_depth,10,100,1)

path_depth is an index that counts tree level of an object.

Collections

There are two new Collection’s criteria that allows to search on SOLR also in Collections:

• Search with SOLR: if checked, searches will be redirected to SOLR (the default is always on local Plone Site).

• Sites: a list of indexes plone sites on SOLR. The user can select on which sites perform the query. If no sites are set (or this criteria not selected), the default search will be made only in the current site.

There is also a customized querybuilder that perform queries to SOLR or to Plone catalog.

Results from SOLR are wrapped into some brain-like objects to be fully compatible with Collection views.

Development buildout

In the buildout there is a solr configuration (in conf folder) and a recipe that builds a solr instance locally.

To use it, simply run:

> ./bin/solr-foreground

Installation

Add rer.solrpush to buildout:

[buildout]

...

eggs =
    rer.solrpush

and run bin/buildout command.

Contribute

• Issue Tracker: https://github.com/RegioneER/rer.solrpush/issues

• Source Code: https://github.com/RegioneER/rer.solrpush

Compatibility

This product has been tested on Plone 5.1 and 5.2

Credits

Developed with the support of Regione Emilia Romagna;

Regione Emilia Romagna supports the PloneGov initiative.

Authors

This product was developed by RedTurtle Technology team.

Contributors

• RedTurtle, [e-mail]

Changelog

1.3.3 (2023-11-08)

• Fix RSS Feed. [cekk]

1.3.2 (2023-11-08)

• Update translation for content_remove_error. [cekk]

1.3.1 (2023-08-01)

• Update translation. [cekk]

1.3.0 (2022-09-29)

• Add search_enabled flag to temporary disable search on SOLR. [cekk]

1.2.0 (2022-01-20)

• Custom scales view to get images from remote contents (to handle also direction). [cekk]

1.1.0 (2021-12-22)

• Add indexers for path infos. [cekk]

1.0.0 (2021-12-20)

• Fix elevate logic. [cekk]

• Add invariant validation for elevate. [cekk]

0.8.0 (2021-11-18)

• SolrBrains now can return img tags if the original content has an image. [cekk]

0.7.1 (2021-10-14)

• Removed unused view. [cekk]

0.7.0 (2021-10-14)

• Add new criteria: solr_portal_types to select a list of portal_types indexed on SOLR. [cekk]

• Add link to Elevate control panel also in user actions. [cekk]

• Fix remote elevate conditions. [cekk]

0.6.4 (2021-09-27)

• Fix how querybuilder create queries. [cekk]

0.6.3 (2021-09-21)

• Add new feature: if “Query debug” flag is enabled in settings, the SOLR query will be shown to managers. [cekk]

• In example schema.xml files (dev and test), set “searchwords” as lowercase type, to be case insensitive. [cekk]

• Disable facet.limit default value (100) to get all facets. [cekk]

• Use swallow_duplicates in Keywords vocabulary to avoid duplicated tokens by truncated strings by SimpleTerm init. [cekk]

0.6.2 (2021-07-15)

• Do not escape queries in querybuilder because solr_search already manage them. [cekk]

0.6.1 (2021-06-10)

• [fix] now sort_on is not ignored on querybuilder customization. [cekk]

• [fix] remove / from frontend_url when not needed in indexing. [cekk]

0.6.0 (2021-05-20)

• Add criteria for search by Subject stored in SOLR. [cekk]

• Now solr brains also return right content-type icon. [cekk]

0.5.1 (2021-04-29)

• Fix release. [cekk]

0.5.0 (2021-04-20)

• Handle all possible exceptions on search call. [cekk]

• Fix encodings (again) for attachement in POST calls. [cekk]

• Handle multilanguage paths in querybuilder for collections (use navigation root path instead portal path). [cekk]

0.4.1 (2021-03-26)

• Fix encodings for attachement in POST calls. [cekk]

0.4.0 (2021-03-25)

• Handle encodings for attachement POST calls. [cekk]

0.3.4 (2021-03-18)

• Fix logs. [cekk]

0.3.3 (2021-03-15)

• Make immediate commits optional from control panel. [cekk]

0.3.2 (2021-02-15)

• Handle simple datetmie dates. [cekk]

0.3.1 (2021-02-11)

• Fix tika indexing parameters: now modified and created dates are correctly indexed. [cekk]

0.3.0 (2021-02-09)

• Refactor elevate control panel and use collective.z3cform.jsonwidget. [cekk]

• Some improvements in indexing. [cekk]

0.2.4 (2021-01-28)

• Fix logic in maintenance view. [cekk]

0.2.3 (2021-01-27)

• Fix maintenance sync view. [cekk]

0.2.2 (2020-12-14)

• Fix encoding problems in escape_special_characters method for python2. [cekk]

• Remove collective.z3cform.datagrifield dependency and temporary disable elevate control panel. [cekk]

0.2.1 (2020-12-03)

• Fix date indexes in query when they already are in “solr syntax”. [cekk]

0.2.0 (2020-12-03)

• Add styles for elevate widget [nzambello]

• Refactor indexer logic. [mamico]

• Add support for bq and qf in search. [mamico]

• Index files with tika. [cekk]

• Add support for collections. [cekk]

• Mute noisy solr logs in maintenance. [cekk]

0.1.2 (2019-12-12)

• Remove noisy logger for queries. [cekk]

0.1.1 (2019-12-12)

• Add new index: path_depth [cekk]

• Fix unicode errors when there is a site name with accents. [cekk]

0.1.0 (2019-12-05)

• Initial release. [cekk]