jsl-elasticsearch 0.1.dev0

Generate ElasticSearch templates from JSL

jsl_elasticsearch generates ElasticSearch mappings from JSL definitions.

It requires Python 3.4 or later (mainly for functools.singledispatch).

It is specifically aimed at use cases where ElasticSearch is being used as a time series database for JSON data with schemas defined using the jsl Python library.

The main API is jsl_elasticsearch.render_es_template:

def render_es_template(document, title, role, doc_type="content"):
    """Render an ElasticSearch time series template for given JSL document

    Template name is generated from the given *title* and *role*
    Document variables are resolved using the given *role*
    *doc_type* specifies the ElasticSearch mapping name (default: "content")
    """

The @timestamp field expected by Kibana is added automatically, and string fields are flagged as not_analyzed by default (so ElasticSearch treats them as opaque tokens, rather than as plain text fields to be analyzed for full text search)

The following JSL field types are currently supported:

• jsl.StringField
• jsl.NumberField
• jsl.IntField
• jsl.ArrayField
• jsl.DictField
• jsl.DocumentField

An additional field type is also defined:

• jsl_elasticsearch.TextField

With string fields being flagged as opaque tokens by default, TextField is a new StringField subclass that flags the field for full text search in the ElasticSearch mapping, but is otherwise handled exactly like StringField by JSL.