Fill in encode, cluster, decode, optimize after all models are completely done

[jz393]

Fill in those method placeholders in summarize.py so we can run the full pipeline and get our results