Recent Python Library Changes

Some notable changes in the Python library:

• r184 touches up the mapreduce sample application, so you should find it to be much easier to use now. See the main project page for screenshots and a link to the video showing how to use it.
• r171 adds mapreduce support for python with user-space shuffler.
• r159 integrates mapper library with Files API.
• r136 uses scatter index to split shards. This should generate more equal shards and works much faster.
• r131 enables starting mapper jobs as part of transaction.
• r129 introduces NamespaceInputReader which iterates over available namespaces.
• r111 supports iterating over datasets with namespaces.
• r76 makes job starting process more bulletproof by splitting it into two phases, and moving seconds phase to taskqueue
• r74 adds support for key-only input reader
• r73 adds support for BlobstoreZipLineInputReader
• r66 restores all code comments in source files
• r42 allows you to implement your own mutation pools
• r40 adds control api to programmatically start jobs