enh: Start with new sequence-labeled POA graph implementation #23

lrvdijk · 2024-09-17T21:03:46Z

Tracking PR for a near-total rewrite of POASTA's alignment algorithm.

Implement a POA graph with sequence-labeled nodes instead of character-based nodes
Implement affine, diagonal transition-based, A* algorithm inspired by Gwfa.
Implement linking diagonals across nodes using a superbubble index.
Implement loading a POA graph from a FASTA MSA.
Implement FASTA MSA output for the new POA graph implementation.

The new CLI includes debug output using the `tracing` crate.

This removes the need for a separate bitset.

lrvdijk and others added 21 commits September 17, 2024 16:59

enh: Start with new sequence-labeled POA graph implementation

4031f66

enh: Implement most of the new A* logic

9729ff6

chore: Cargo updates

8f75a52

fix: Add missing edge when processing alignment

dbe9d74

fix: Fix incorrect inclusion of new seq ID in split edge

6f392de

enh: Improve aligned interval edges in DOT output

9d6f911

enh: Organize into cargo workspaces

d00351a

enh: Start implementation of new CLI

59c225e

The new CLI includes debug output using the `tracing` crate.

fix: Fix alignment backtrace

ec6c1b7

fix: Use tracing instead of eprintln

ee8c9bb

chore: Keep info! tracing in release mode

fa9c1d8

enh: Add sequence to be aligned to graph DOT

21f1504

enh: Update plotting script to support new POA graph

8dcbcbe

Remove old build files

9cf165b

enh: Add ability to import a graph from FASTA MSA

f70599b

enh: Re-order generic parameters of PoastaAligner

9191a3d

fix: Fix tracing terminal colors

61dd7d0

enh: Use most significant bit as visited flag

6079696

This removes the need for a separate bitset.

enh: Don't check if visited in update_if_further

05df110

feat: Enable plotting a range

f500464

chore: Cargo.lock update

c7ac60d

lrvdijk force-pushed the main branch from 84873e8 to 36d56f9 Compare November 27, 2024 21:56

Provide feedback