Cemba-data

- Modify stampede2 script, so it will copy the miniconda env to /tmp and setup all the path variables

Fix snm3C summary file naming issue

Compress the contact in the end

As required by TACC, each job now clone the conda mapping environment into /tmp dir of the computation node, so the mapping pipeline don't need to I/O /work

These lines added to the sbatch template file

shell
conda create -p /tmp/mapping --clone mapping
conda init do not work in shell script, this worked as an replacement
source $(dirname $(which conda))/../etc/profile.d/conda.sh
conda activate /tmp/mapping

1. `yap sbatch` can recognize failed snakemake commands. It will automatically retry twice failed ones and put the final results in sacct.csv.gz
2. `yap sbatch` will recognize previously successful jobs, and prevent rerun when resubmitting the same command list
3. `yap sbatch` default changed:
python
if mode == 'm3c':
time_str = "6:00:00"
total_mem_mb = 160000
elif mode == 'mc':
time_str = "4:00:00"
total_mem_mb = 192000
elif mode == 'mct':
time_str = "6:00:00"
total_mem_mb = 192000

Support V2-single demultiplex mode. V2-single means each plate is indexed by a single PCR index. The user only need to change the way they make plate info (single multiplex group based on the index), the demultiplex results are the same as before (six multiplex groups based on columns). See example in [documentation](https://hq-1.gitbook.io/mc/prepare/inhouse-prepare-sample-sheet#plateinfo-examples)