Danger Rider is an indexing and retrieval system for Python objects.

The idea behind Dangerrider is that it gives a high-speed in-memory storage mechanism that allows you to store, index, and collect aggregate-type data against any python class or set of classes that implement the same interface.

Danger Rider gives you a rich toolset to explore and analyze your objects.

Dangerrider allows you to define indices in python and, in effect, perform very fast aggregate calculations on your in-memory data.

Dangerrider will have the following features:

• Indexes - explicit indexes, defined in python, that work hand-in-glove with your domain model to add aggregation and analysis capabilities. Indexes are always explicit, so as to allow the rapid insertion of new items.
• Aggregators - explicit aggregators, defined in python, that you can use to calculate sums, averages, counts, and anything else that spans objects or sets of objects. Low latency.
• Utility Aggregators - aggregators that will "just work" out of the box. With these, you can get sums, averages, and counts out of the box. You can also define your own
• Dangershell, a shell that will allow you to access your stored data and any annotations related to your object library.
• DangerQL - A sql-like query language that returns your stored objects instead of "rows."
• Optional Persistence - pluggable persistence mechanisms to ensure that your data stays stored. Sometimes you don't care about persisting data, so this is optional.

TBD

TBD

Utility aggregators have been added. Next, we should add the following:

• Ability to do sorting on results (indexed sorting?)
• Ability to limit results
• Grouping for aggregates
• Bigger dataset tests (to prove that this will really be a helpful thing!)
• Query gateway so developers don't have to structure queries in tuples...
• Unit tests for all..

After that, we will add (gevent based?) danger rider daemon, so we can access the fast ram-persisted datastore through a server.

After that, dangershell to access the daemon.

Finally, DangerQL.