Introduction

Overview

pyoseo is an ordering server. It processes requests for spatial data and makes the data available to the requesting user upon completion.

It implements the OGC OSEO standard. OSEO stands for Ordering Services Framework for Earth Observation Products. The standard defines interfaces, bindings and requirements for establishing a workflow for ordering of Earth Observation data between a server and client.

An effective OSEO server needs to implement the following components:

• web server - listens to incoming requests, passes them to the oseo engine for processing. Issues the appropriate response back to the client;
• oseo engine - Parses the requests and build responses according to the OSEO standard;
• database - records every order and associated parameters. Tracks an order’s state as it gets processed as well as associated data;
• queueing system - A process that is continuously running, preparing orders as they are requested. Order processing is asynchronous;
• order processing module - A set of tasks that perform the actual processing of each ordered item;
• authentication framework - Ensures that the client that made the request has the necessary permissions to use the server;
• order delivery module - Makes the ordered items available to the client after they have been processed.

Some of these components are specific to each particular implementation. As such, instead of trying to cater to every possible situation, pyoseo provides hooks into these modules and leaves their concrete implementation to each particular instance. This means that the following components are not provided by pyoseo and must be implemented independently:

• Authentication framework. Check the Hooking up your authentication module page for more details;
• Order processing and delivery modules. Check the Hooking up your order processing modules page
• Order options parsing. Check the Hooking up your custom order options page for more details.

Implementation

Pyoseo is implemented in Python. It essentially glues together some widely used python libraries, harnessing their power to bring the OSEO standard into fruition:

• The web server and database components are essentially a django project
• The queueing system component uses celery
• Parsing, validating and creation of OSEO elements is done with pyxb and lxml

So we basically have a django project running, listening for HTTP POST requests. There is an oseo_endpoint() view that instantiates the OseoServer class. This class then proceeds to:

• Parse the request,
• Authenticate the request by calling the external authentication class
• Instantiate a specialized OSEO operation class
• Receive the response from the OSEO operation class and return it to the django view, who then returns it to the web server

An OSEO operation class is tasked with processing the request. Depending on the request, this class will communicate with the database, creating, updating or querying order records and order items. It will also communicate with the celery workers by issuing processing tasks.

When a processing task gets scheduled for execution, it retrieves the order’s relevant parameters from the database and then calls the external processing class, which is responsible for actually fetching the ordered items, applying any requested options. The processing class will also perform the delivery of the items to the specified endpoint.