Skip processing empty dataset at any stage

[bhavitvyamalik]

In the pipeline, same categories.json is copied for every next step followed by copying all datasets from previous step to current step. However there might be edge cases where:

• Downloaded dataset is empty (affects clean(next) step)
• Dataset after cleaning becomes empty (affects decontaminate (next) step)
• Dataset after decontamination might become empty (affects gather (next) step) --very rare but still a possibility

I think it will be prudent to have a sanity check so that we don't copy empty datasets from previous step to next step. We can remove if dataset is empty check in _cmd_exit_str after this.