H5MD implementation

[RubenChM]

To solve mmb-irb/MDDB-REST-API#4, H5MD could be used as main file to store the trajectory information of every project.
This would be a big change, so the benefits should justify the new for a system to update the data already stored in the database:

1. Adding a system to update the database would enable adding new fields for efficiency (mdcount, unique resname/atom names) and syncronize it with changes in the workflow (now PDB references are saved from the workflow instead of fetched).
2. Having a reference format would enable converting it to other with software like Chemfiles)
3. Saving storage and transfer bandwidth. H5MD frames can be access while compressed, allowing faster upload/download and need less disk.

Links:

• https://www.nongnu.org/h5md/
• Paper H5MD
• GMX H5MD
  • Issues
  • Merge request

TODO:

• Upload compression algorithm on a separate conda