Investigate patristic distance algorithms

[glstott]

Current method is storage-expensive, requires some configuring, and slower than preferable. Swapping to a separate script for patristic distance calculations and inputting those directly into the graph would be faster, but more memory intensive. Maybe look at both approaches and provide them as options for users.

[glstott]

Two new methods used.

• First one, computed outside PMeND on the entire tree. Viable, and fast. I didn't yet optimize, so it creates large files as a result which need to be transferred. Also requires a good amount of memory to run.
• Second option, compute in PMeND using APOC and graph data science procedures. This was also fast, but pretty easy to do even in a local environment with just 8 GB RAM. I think this is the route I'd like to pursue.
• I will comment and post code for each of these in the next couple of weeks.

[glstott]

I ended up using the latter option. It performs better, and doesn't have a significant memory cost. The first alternative listed above also works great, but would add an additional dependency that doesn't seem necessary. In the future, if we want to come back to it, loaded in a file generated from a patristic distance script in Python (so I wouldn't need as large of a file each time). Alternatively, PATRISTIC could have done something similar.