Homogenize trio vs hic architecture

[skoren]

Currently, trio requires users to build homopolymer-compressed databases and we run one round of consensus after rukki resolution and coloring. In contrast, with Hi-C data, we run consensus, map Hi-C data, then run rukki and another round of consensus. This is similar to the architecture for GFAse as well.

This should be universal, perhaps always use uncompressed post-consensus sequence? This would allow users to build normal meryl databases rather than HPC ones, simplifying confusion. Additionally, output file locations differ between HiC and trio (e.g. trio = 6-layout*/*scfmap vs hic = 8-*/6-layout*/*scfmap). These should be promoted to top level to be the same in both cases.

[skoren]

The scfmap files have been promoted and the final outputs are synchronized. However, we still need to homogenize the rest of the pipeline.

[skoren]

Likely the best solution is to use non-HPC assembly for both trio and Hi-C. If we can switch HPC off in MBG and correction, then it will simplify consensus and remove the extra consensus step in the Hi-C pipeline, allowing it to run alongside 6-rukki as it does with trio now.