This is used in a feature-level ml-service. RankerExample proto contains a feature map from name to value, and ExamplePreprocessor can vectorize it to a float list with an ExamplePreprocessorConfig. The vectorized list can be used as tensor and fed to tflite models.
Files in this directory are copied from chromium repo: //components/assist_ranker/.
ranker_example.proto and example_preprocessor.proto in ../proto are copied from chromium repo: //components/assist_ranker/proto/