Advertising delivery engine based on SF1R.

• Sponsored Search. Automatic bid optimization is provided with the semantic of broad match. Currently, the broad match is delivered through succinct self index, which is the fastest top-K union index in the world. However, it's an overkill to use succinct self index for such purpose, we plan to use Roaring Bitmap in future which is the fastest compressed bitmap delivered by Daniel Lemire.

• Online engine for TARE. TARE refers to Targeted Advertising and Recommender Engine, which aims to build an unified engine for both targeted advertising and recommendation. Currently, TARE is tailored for products only.

The dependencies are similar with that of SF1R. Besides, these third party repositores are required:

• avro: The C++ serialization for AVRO.

• libcuchbase: The C++ client for Couchbase.

The SF1R project is published under the Apache License, Version 2.0: http://www.apache.org/licenses/LICENSE-2.0