collective.elastic.ingest 2.0.0b1

Installation

We recommended to use a Python virtual environment, create one with python3 -m venv venv, and activate it in the current terminal session with source venv/bin/activate.

Install collective.elastic.ingest ready to use with redis and opensearch:

pip install collective.elastic.ingest[redis,opensearch]

Depending on the queue server and index server used, the extra requirements vary:

• index server: opensearch, elasticsearch.

• queue server: redis or rabbitmq.

Configuration

Configuration is done via environment variables and JSON files.

Environment variables are:

INDEX_SERVER

The URL of the ElasticSearch or OpenSearch server.

Default: localhost:9200

INDEX_USE_SSL

Whether to use a secure connection or not.

Default: 0

INDEX_OPENSEARCH

Whether to use OpenSearch or ElasticSearch.

Default: 1

INDEX_LOGIN

Username for the ElasticSearch 8+ or OpenSearch server.

Default: admin

INDEX_PASSWORD

Password for the ElasticSearch 8+ or OpenSearch server.

Default: admin

CELERY_BROKER

The broker URL for Celery. See docs.celeryq.dev for details.

Default: redis://localhost:6379/0

PLONE_SERVICE

Base URL of the Plone Server

Default: http://localhost:8080

PLONE_PATH

Path to the site to index at the Plone Server

Default: Plone

PLONE_USER

Username for the Plone Server, needs to have at least Site Administrator role.

Default: admin

PLONE_PASSWORD

Password for the Plone Server.

Default: admin

MAPPINGS_FILE

Absolute path to the mappings configuration file. Configures field mappings from Plone schema to ElasticSearch.

No default, must be given.

PREPROCESSINGS_FILE

Configures preprocessing of field values before indexing.

Default: Uses a defaults file of this package.

ANALYSIS_FILE

(optional) Absolute path to the analysis configuration file.

SENTRY_DSN

(optional) Sentry DSN for error reporting.

Default: disabled

Starting

Run celery worker:

celery -A collective.elastic.ingest.celery.app worker -c 1 -l info

Or with debug information:

celery -A collective.elastic.ingest.celery.app worker -c 1 -l debug

The number is the concurrency of the worker. For production use, it should be set to the number of Plone backends available for indexing load.

OCI Image

For use in Docker, Podman, Kubernetes, …, an OCI image is provided at …

The environment variables above are used as configuration.

Additional the following environment variables are used:

CELERY_CONCURENCY

The number of concurrent tasks to run.

Default: 1

CELERY_LOGLEVEL

The log level for celery.

Default: info

Examples

Example configuration files are provided in the /examples directory.

Text Analysis

Test analysis is optional. Skip this on a first installation.

Search results can be enhanced with a tailored text analysis. The simple fuzzy search, which can be used without any analysis configuration, has its limits. This is even more true in complex languages like German.

This is an advanced topic.

You can find detailed information about text analysis in the ElasticSearch documentation. We provide an example analysis configuration for a better search for German compounded words.

Example: A document with the string ‘Lehrstellenbörse’ can be found by querying ‘Lehrstelle’. It shall be found too by querying ‘Börse’ using a decompounder with a word list ‘Lehrstelle, Börse’ and an additional stemmer. The example analyzer configuration applies a stemmer, which can handle inflections of words. This is an important enhancement for better search results.

The analysis configuration is a configuration of analyzers. The example provided here uses two of them: german_analyzer and german_exact.

The first decompounds words according the word list in lexicon.txt. A stemmer is added.

The second one is to allow also exact queries with a quoted search string.

These two analyzers are to be applied to fields. You can apply them in your mapping.

Example:

"behaviors/plone.basic/title": {
    "type": "text",
    "analyzer": "german_analyzer",
    "fields": {
        "exact": {
            "type": "text",
            "analyzer": "german_exact_analyzer"
        }
    }
},

Check your configured analysis with:

POST {{elasticsearchserver}}/_analyze

{
    "text": "Lehrstellenbörse",
    "tokenizer": "standard",
    "filter": [
        "lowercase",
        "custom_dictionary_decompounder",
        "light_german_stemmer",
        "unique"
    ]
}

The response delivers the tokens for the analyzed text ‘Lehrstellenbörse’.

Note: The file elasticsearch-lexicon.txt with the word list used by the decompounder of the sample analysis configuration in analysis.json.example has to be located in the configuration directory of your elasticsearch server.

Source Code

The sources are in a GIT DVCS with its main branches at github. There you can report issues too.

We’d be happy to see many forks and pull-requests to make this addon even better.

Maintainers are Jens Klein, Katja Suess and the BlueDynamics Alliance developer team. We appreciate any contribution and if a release is needed to be done on PyPI, please just contact one of us. We also offer commercial support if any training, coaching, integration or adaptions are needed.

Installation for development

• clone source code repository,

• enter repository directory

• recommended: create a Virtualenv python -mvenv env

• development install ./bin/env/pip install -e .[test,redis,opensearch]

• load environment configuration source examples/.env.

License

The project is licensed under the GPLv2.