Finalizing Dawg 2.0

[reedacartwright]

A possible list of tasks that need to be completed to polish up the Dawg 2.0 code base in preparation for a paper.

• Replace CMake build system with Meson #68
  • Update code to C++17 standards
• Change internal API to be more developer friendly. #72
  • Replace trick file format with Yaml. Use rapidyaml library for parsing.
  • Support linking libdawg with R and Python
• Reduce memory consumption with large trees #66
• Optimize alignment construction algorithm. #69
• Use CLI11 for program options.
• Update the random number library to the latest algorithms used in the Cartwright Lab. #70
• Explore alternative strategies for simulation. Figure out in what parameter ranges which algorithms are most efficient. #71
• Create testing framework, both unit tests and functional tests.
• Auto-format the code.
• Lint the code.
• Validate models #67
• Optimize code

[reedacartwright]

Here's the beginnings of a proposal for replacing trick with yaml. At the top level, there are three+ sections. parts defines partitions of the root sequence and the root sequence itself. rules defines evolutionary models. regions defines sections of tree space with known topology and branch length, tied to a partition, and has an evolutionary model.

reps: 100 # can be overridden on the command line
seed: [1,2,3] # can be overridden on the command line

parts:
- name: exon_1
  length: 1000 # ignored because seq is specified
  seq: "ACGT"
  type: DNA

rules:
-
  name: default
  subst:
    model: k2p
    alpha: 1
    freqs: [0,2,3,4]
  ins:
    rate: 0.01
    model: powerlaw
    alpha: 1.5
    max_len: 100
  del:
    rate: 0.01
    type: powerlaw
    alpha: 1.5
    max_len: 100

regions:
-
  tree: "(A:0.1,B:0.1).R;"
  rule: default
  part: exon_1