Getting Started with the Python OJAI Client

This section describes the software required to run the Python OJAI client, client/server security, and how to specify your connection string. It also provides links to documentation that shows you how to write Python OJAI applications.

The Python OJAI client is available starting with the EEP 6.0 release.

Software Requirements

You must have the following software installed to run the client:

Client Software Installation Notes
Python Use Python 2.7 or later
pip See https://pip.pypa.io/en/stable/installing/ for instructions specific to your environment.
Python OJAI client Install the client by using the following command: 
pip install maprdb-python-client

You also must have access to the following software:

To run a Python OJAI application, you simply need to install and configure the Data Access Gateway:

Python OJAI Client Security

The client supports username/password authentication. The initial connection (and token renewal) use these credentials. Subsequent communication uses JWT.

When connecting to a secure cluster, the client uses:

  • X.509 certificates to authenticate with the Data Access Gateway
  • TLS v1.2 to encrypt communication between the client and the Data Access Gateway

Python OJAI Client Connection String

The string you use to connect your OJAI client to the cluster must have the following format:

"[ojai:mapr:thin:v1@]<hostname>[:<port>][?<option_name>=<option_value>;...]"

The prefix ojai:mapr:thin:v1@ is optional.

<hostname>
Name of the Data Access Gateway host. For information about the host name, see Configuring SSL for OJAI Clients.
<port>
Port number (see Ports Used by HPE Ezmeral Data Fabric Software) that gRPC clients use to connect to the Data Access Gateway

Default: 5678

auth=<scheme_name>

The authentication scheme for the current connection; currently, only basic

user=<username>

The user name for basic authentication

password=<password>

The password for basic authentication

ssl=true|false

Whether to establish a secure connection using SSL/TLS

An error is returned if there is a mismatch between your client and Data Access Gateway security settings. The default for this option is true, which is the required setting if connecting to a secure Data Access Gateway. If connecting to a non-secure Data Access Gateway, set it to false.

If set to false, the other SSL parameters are ignored.

Note that the grpc.service.ssl.enabled property controls the SSL setting for the Data Access Gateway. For more information, see Administering the Data Access Gateway.

sslCA=<path to PEM file containing CA certificate>

Path to a local file containing Certificate Authority (CA) signed certificates in PEM format. For information about the PEM format, see Configuring SSL for OJAI Clients.

Required when the ssl option is set to true.

sslTargetNameOverride=<CA certificate common name>

Fully qualified domain name specified in the SSL server certificate, which is different from the <hostname> in the connection string.

For example, imagine that you are using the following:

  • Public network host name is ec2-203-0-113-25.compute-1.amazonaws.com.
  • Internal DNS is node1.mydomain.com.
  • CA signed certificate is issued to node1.mydomain.com.

Using these names, you must specify the following connection string:

"ec2-203-0-113-25.compute-1.amazonaws.com:5678?ssl=true;sslCA=/opt/app/conf/rootca.pem;sslTargetNameOverride=node1.mydomain.com"

Other examples of connection strings are the following:

"ojai:mapr:thin:v1@localhost:5678?auth=basic;user=fred;password=george;sslCA=/opt/app/conf/rootca.pem"
"localhost:5678?ssl=false;auth=basic;user=fred;password=george"

Python OJAI Connection Retry Options

If your OJAI client cannot connect to your cluster, it waits 10 ms. After 10 ms, it makes a second connection attempt. If that fails, it continues the attempts up to a configurable number of retries. The following parameters control the number of retries and the wait time between attempts:

Connection Option Parameter Description Default Value
ojai.mapr.rpc.wait-multiplier Multiplier that determines the wait time for subsequent attempts after the initial 10 ms wait. The previous wait time is multiplied by this parameter. 1000
ojai.mapr.rpc.wait-max-attempt Maximum wait time between attempts regardless of the multiplier parameter 18000 ms
ojai.mapr.rpc.max-retries Maximum number of retry attempts 7

The following examples demonstrate how these parameters work, including the default case:

Attempt # Wait Time (in ms) for each Retry Attempt

Default Parameters:

{
    'ojai.mapr.rpc.wait-multiplier': 1000,
    'ojai.mapr.rpc.wait-max-attempt': 18000,
    'ojai.mapr.rpc.max-retries': 7
}
 
{
    'ojai.mapr.rpc.wait-multiplier': 2,
    'ojai.mapr.rpc.wait-max-attempt': 90,
    'ojai.mapr.rpc.max-retries': 5
}
1 10 10
2 10*1000 = 10000 10*2 = 20
3 18000

10000*1000 = 10,000,000, which exceeds 18000		 20*2 = 40
4 18000 40*2 = 80
5 18000 90

80*2 = 160, which exceeds 90
6 18000 Error
7 18000 N/A
8 Error N/A

To set these retry options, you must pass them in the ConnectionFactory.get_connection call:

connection_str = 'localhost:5678?auth=basic;user=mapr;password=mapr;' \
    'ssl=true;' \
    'sslCA=/opt/mapr/conf/ssl_truststore.pem;' \
    'sslTargetNameOverride=node1.mapr.com'
options = {
    'ojai.mapr.rpc.wait-multiplier': 5,
    'ojai.mapr.rpc.wait-max-attempt': 50,
    'ojai.mapr.rpc.max-retries': 3
}
connection = ConnectionFactory.get_connection(connection_str=connection_str,options=options)

Writing a Python OJAI Application

For information about writing a Python OJAI application, see the Python sections in the following topics:

Querying in OJAI Applications

Provides an introduction to the basic flow of an OJAI application that queries a HPE Ezmeral Data Fabric Database JSON table

Examples: Querying JSON Documents

Contains code samples of OJAI applications that query HPE Ezmeral Data Fabric Database JSON tables

Managing JSON Documents

Describes how to perform CRUD (create, query, update, and delete) operations on JSON documents in HPE Ezmeral Data Fabric Database JSON tables