Skip to content

Split Conformational Analysis into Explicit Serial Phases #363

Description

@harryswift01

Goal

Refactor the current serial conformational analysis into explicit phases that can later be mapped onto a chunked map-reduce workflow.

Background

The current workflow already has a two-pass structure:

identify peaks:
    collect selected-frame dihedral angles
    build histograms
    identify peaks

assign states:
    inspect selected-frame dihedral angles
    assign states using peaks
    collect state labels and flexible-dihedral counts

This sub-issue should make those phases explicit without changing execution behaviour.

Scope

  • Keep execution serial.

  • Refactor ConformationStateBuilder or related conformational helpers into clearer phases:

    • static dihedral/topology discovery
    • selected-frame angle collection
    • histogram/peak identification
    • state assignment from peaks
    • flexible-dihedral counting
    • final output construction
  • Extract angle collection logic from peak identification.

  • Extract peak construction from collected angle/histogram data.

  • Extract state assignment so it can later operate on chunk-local dihedral results.

  • Preserve the frame-index contract:

    • MDAnalysis Dihedral.run(...) uses active analysis-universe frame indices.
    • Dihedral.results.angles is indexed locally from zero.
    • absolute/source frame indices must not be used directly to index result arrays.
  • Do not add chunking or Dask yet.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

Type

Projects

No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions