ajelastic

Python - Elasticsearch Integration (for Aasaanjobs Internal Usage)


License
MIT
Install
pip install ajelastic==1.0

Documentation

Aasaanjobs Elasticsearch SDK

Library to handle elasticsearch related transactions in Python (for internal usage in Aasaanjobs)

Requirements

  • Python 3.5 and greater
  • Elasticsearch Python

Installation

Install using pip

pip install ajelastic

Configuration

The AJ Elastic library reads configurations from module path specified in the environment variable AJELASTIC_SETTINGS_MODULE if its a standalone python project, or the DJANGO_SETTINGS_MODULE environment variable if its a Django project.

export AJELASTIC_SETTINGS_MODULE=tests.settings

Example

Below is an example settings module that can be consumed by the library

# Mandatory Settings

# Elasticsearch host and port
ES_HOST = "http://localhost:9200"
# The environment suffix that will be added in the Elasticsearch index names
ES_ENV = "development"              

# Optional Settings

# List of Elasticsearch indices
ES_INDICES = {
    "User": {
        "name": "users",                                # The index name prefix
        "doc_type": "user",                             # Elastic document type
        "data_functions": {                             
            "single": "tests.data:get_user",          # Function to fetch a user; should accept a single ID argument
            "multi": "tests.data:list_users"          # Function to fetch list of users; should accept two arguments; limit and offset
        },
        "mapping_path": "tests/user_mapping.json"    # Path to the JSON mapping of the document structure
    }
}

Commands to Sync Data with Elasticsearch

This library provides the following command line scripts to synchronize/initialize data

aj-es-reindex

This command will reindex the documents from scratch fetching data from the function defined in the data_functions.multi settings of the entity (i.e., withing the ES_INDICES settings).

For eg, the below command will reindex the User entity, fetching data from the list_users function (as specified in the settings) with a batch size of 100, i.e., in a single request 100 users will be index to elasticsearch.

aj-es-reindex User 100