Skip to content

Latest commit

 

History

History
130 lines (89 loc) · 2.62 KB

devnotes.md

File metadata and controls

130 lines (89 loc) · 2.62 KB
Raw

dev notes

Notes for development.

MVP checklist

  • visualization of the graph itself from the nodes and edges (no haplotypes)
    • layout
      • use tubemap as the layout engine
    • nodes
    • text for the sequence
    • edges (curves)
      • improved curve calculation
  • visualization of haplotypes on the z-axis (paths)
    • should be simple to do if we can visualize the graph
  • controls
    • better camera positioning
    • move graph with left, right arrow keys (simulate window move)
    • select haplotypes with up, down arrow keys
  • interactivity
    • navigation buttons for mobile
    • display number of nodes, edges, paths, selected path
    • hover node should show coverage (% of paths that walks over this node)
  • demo graphs
    • find/generate better graphs to display different SVs
    • take screenshots for paper

Backend

  • server (likely out of scope)
    • hook into tube-maps server
    • move window with arrow keys (actually dynamically download graph)

vg notes

Running a vg container

docker run -it --name pgv-vg \
    -v "$(pwd)":/vg/pgv \
    quay.io/vgteam/vg:latest \
    bash
docker start pgv-vg -i

Data flow

# Construct vg graph from a FASTA and VCF file (reference + variants).
vg construct -r z.fa -v z.vcf.gz > z.vg


# Get some basic stats on the graph.
vg stats -z z.xg
vg stats -N z.xg


# Index the graph.
vg index z.vg -x z.xg


# Construct a subgraph given a region.
vg chunk -x z.xg -g -c 20 -r 200 -T -b chunk -E regions.tsv > z.chunk.vg

# Index and generate a DOT and JSON file for the subgraph.
vg index z.chunk.vg -x z.chunk.xg 
vg view -dnp z.chunk.xg | dot -Tsvg -o z.chunk.svg

vg view -j z.chunk.xg > z.chunk.json

Cactus example

vg chunk -x cactus.xg -c 20 -r 24:27 | vg view -j - | jq . > cactus_r24\:27_c20.json

Docker image

Build the image

# Build project
yarn core:build
yarn web:build

# Build image (with latest tag)
docker build -t pgv .

Run the image locally

docker run -it --name pgv --rm -v "$(pwd)/examples":/pgv/ui/examples -p 8000:8000 pgv:latest

Publish to quay.io

# Login first
docker login quay.io

# Tag image
docker image tag pgv quay.io/wlgao/pgv:latest

# Push to quay.io
docker image push quay.io/wlgao/pgv

Build and publish to Docker Hub (for multiple platforms)

# Login first
docker login docker.io

# Create a builder instance
docker buildx create --use

# Build and push
docker buildx build --platform linux/amd64,linux/arm64 -t wlgao/pgv --push .