SpringSaLaD: auto-generated _tot observables land in the wrong structure, and there is no way to move one

オープン
  • 何か問題がありますか？
#2,071 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
3/5
見積もり時間
半日
初心者へのやさしさ
72/100
issue の種類
バグ
明瞭さ
明確に書かれている
活発さ
活発
技術スタック
java
領域
desktop

調査の方向性

Start at Model.createObservable in vcell-core/src/main/java/bit/vcell/model/Model.java:1524-1550 (actual path vcell-core/src/main/java/bit/vcell/... per payload: vcell-core/src/main/java/bit/vcell/model/Model.java) and see how the getStructure(0) fallback interacts with a molecule's anchor restriction. Then inspect ObservableTableModel:195 for the Structure column and addObservable's empty SpeciesPattern seed. Done looks like: creating an anchored molecule yields an observable in a valid structure with no Problems-tab warning, verified by the three-compartment SpringSaLaD reproduction.

索引モデルが issue の本文から書いたものです。

説明

Found while writing a SpringSaLaD tutorial script. In a three-compartment SpringSaLaD
model, every observable VCell creates for you lands in Extracellular, and for an
anchored molecule that is a standing warning with no remedy but deletion.

What happens

Creating a molecule creates an observable counting it. Model.createObservable
(vcell-core/src/main/java/cbit/vcell/model/Model.java:1524-1550) falls back to
getStructure(0) when no structure is given, and in a SpringSaLaD model structure 0 is
Extracellular.

For a molecule anchored to the membrane, that is immediately wrong, and the Problems tab
says so:

Molecule RK cannot be present in the structure due to anchoring.

The Structure column is not editable — ObservableTableModel:195 returns false for it
unconditionally — so the observable cannot be moved to where it belongs.

And the replacement route leaves its own complaint

The obvious workaround is to delete it and create a new one through
New > In Membrane Membrane, which does ask for a compartment. That produces an observable
in the right structure with the right pattern, and a different warning:

Molecule of Species Pattern is empty.

addObservable(o) seeds the new observable with o.addSpeciesPattern(new SpeciesPattern()),
and setting the BioNetGen definition afterwards does not consume that empty pattern. There
is no table-reachable way to remove it.

So neither path reaches a clean model, and the only way to have no warnings is to have no
observable for the anchored molecule at all.

Reproduction

  1. Three structures named Extracellular / Membrane / Intracellular.
  2. A molecule RK(B~State0~State1,Anchor~Anchor,K~Off~On), anchored to Membrane via
    Object Properties > "Only these:" > Membrane.
  3. A seed species for it, created In Membrane Membrane.
  4. Look at Observables, and at the Problems tab.

Suggested fixes

  • The auto path could use the molecule's anchor. When a molecule is restricted to one
    structure, that structure is the only sensible home for an observable counting it.
  • Failing that, make the Structure column editable, restricted to the structures the
    molecule is allowed in.
  • Separately, the empty seed species pattern should be consumed when a definition is set,
    or not created until there is something to put in it.

Note

Nothing is actually lost for a SpringSaLaD run: LangevinLngvWriter does not read
observables at all, and SpringSaLaD results come from the solver's own molecule, state,
bond, site and cluster counters. The cost is a model that cannot be made warning-free,
which matters most to a new user who does not yet know which warnings to ignore.

主要言語
Java
スター
140
フォーク
35
平均マージ
17時間 39分
マージ済み PR（30日）
19

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

virtualcell/vcell のほかの issue

virtualcell/vcell の issue をすべて見る

似ている issue

Java の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。