[Task]: Create a helper function fill_in_missing for TFT Criteo tests

### What needs to happen?

This function should accept a rank 2 SparseTensor, and a default value, and return a rank 1 Tensor. It will assume the input is from a VarLenFeature and has dimensions [batch_size, 0] or [batch_size, 1] depending on the max size of the feature over the batch. It's assumed each feature has 0 or 1 values (0 for missing, 1 for present).

It will emit a Tensor which is constructed using the code

feature = tf.sparse_to_dense(
feature.indices, [feature.dense_shape[0], 1], feature.values,
default_value=-1)
feature = tf.squeeze(feature, axis=1)

### Issue Priority

Priority: 3 (nice-to-have improvement)

### Issue Components

- [X] Component: Python SDK
- [ ] Component: Java SDK
- [ ] Component: Go SDK
- [ ] Component: Typescript SDK
- [ ] Component: IO connector
- [ ] Component: Beam examples
- [ ] Component: Beam playground
- [ ] Component: Beam katas
- [ ] Component: Website
- [ ] Component: Spark Runner
- [ ] Component: Flink Runner
- [ ] Component: Samza Runner
- [ ] Component: Twister2 Runner
- [ ] Component: Hazelcast Jet Runner
- [ ] Component: Google Cloud Dataflow Runner