+ (* TODO: Make a nice(r) abstraction to re-assemble pieces in the pipeline:
+ *
+ * from input to files_by_size
+ * from files_by_size to files_by_sample
+ * from files_by_sample to files_by_digest
+ * from files_by_digest to output
+ *
+ * input |> files_by_size |> files_by_sample |> files_by_digest |> output
+ *)