sra_metadata as standalone function that creates `metadata.yaml` sheet

[aguang]

See #41 for some context.

We want to split out sra_metadata.py so that it actually is a standalone python script that will create a metadata.yaml sheet from querying SRA, for qcdb.db_load to parse and load. It will take as input a path to a data folder, and an output path where metadata.yaml will go. The data folder should have either SRR ids as generated from sra-tools/fasts-dump, or SRX_SRS ids as generated from bioflows.

[aguang]

For a sheet that's based on data from SRA, db_id should be filled out as an SRS_SRX id, and it should also have fields sample_id: SRS and experiment_id: SRX.

[j-stat]

All set and pushed to GitHub