Simple JSON to Cassandra Python client.
Pull request Compare This branch is 3 commits behind joaquincasares:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
CassandraJsonMapper
tests
README.md
setup.py

README.md

CassandraJsonMapper

CassandraJsonMapper is an Apache Cassandra Python driver developed for direct functionality to store JSON-style objects mapped into Cassandra composite columns.

This is done via the main methods: save(key), get(key), delete(key).

History

I needed something small, light-weight, and fast enough on reads and writes for short bursts of data. I was handling JSON returns from REST API calls that I would later do light-weight analytics over. This fit my use case efficiently, so I decided to share it.

Setup

pip install CassandraJsonMapper

Initialize Database

Since CassandraJsonMapper uses composite columns heavily to do it's nesting, the schema for the column family that CassandraJsonMapper will use must look similar to what is provided.

NOTE: Keep in mind that the number of composite columns created must be at least as deep as the deepest JSON document that will be saved.

create keyspace json
  with placement_strategy = 'SimpleStrategy'
  and strategy_options = {replication_factor : 1};

use json;

create column family json
  with column_type = 'Standard'
  and comparator = 'CompositeType(
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType,
    org.apache.cassandra.db.marshal.BytesType,org.apache.cassandra.db.marshal.BytesType)'
  and default_validation_class = 'BytesType'
  and key_validation_class = 'BytesType';

Example Code

import CassandraJsonMapper

db = CassandraJsonMapper.db(keyspace='json', column_family='json')

document = {
    'key2': {
        'test': 'okay'
    }
}
db.save(document)
assert db.get('key2') == document['key2']

db.delete('key2')
assert db.get('key2') == {}

Multi-threaded Example

db.mt_save(document)
db.mt_finish()

assert db.get('key2') == document['key2']

Limitations

In hopes of keeping this client simple, the following is not supported:

  • Cannot append to lists.
  • Cannot delete portions of a document.
  • Reads may not be returned in full, but paging is supported.

Future Implementations

  • Deleting portions of documents.
  • Better list support.

Methods

This client has a few main methods:

def __init__(self, keyspace, column_family, server_list=['localhost'],
             write_consistency=ConsistencyLevel.ONE,
             read_consistency=ConsistencyLevel.ONE,
             request_size=100, batch_size=6000, thread_count=20)
    """Create the Cassandra connection pool and CF connection."""

def save(self, dictionary_payload, write_consistency=None, batch_size=None)
    """Convert and save dictionary into Cassandra."""

def get(self, key, read_consistency=None,
        return_last_row=False, column_start=None, request_size=None)
    """Read and convert Cassandra response into dictionary."""

def delete(self, key, write_consistency=None)
    """Delete dictionary from Cassandra."""

def mt_save(self, dictionary_payload)
    """Save dictionary asynchronously."""

def mt_finish(self)
    """Wait until all pending inserts are performed."""