#55 CQuery: tree iterator

Open
opened 4 years ago by ranides · 3 comments
ranides commented 4 years ago

Create simple abstract class which encapsulates algorithm visiting tree. It will implementing require few method: getChildren, getParent, etc

grammar library will implement ANTLR functionality on top of this.

Create simple abstract class which encapsulates algorithm visiting tree. It will implementing require few method: getChildren, getParent, etc `grammar` library will implement ANTLR functionality on top of this.
ranides commented 4 years ago
Owner

We should implement at least two strategies:

  • BFS
  • DFS

Inner visitor will be just BFS/DFS with skip(1), nothing more.

We should expect at least two accessor interfaces:

  • TreeIterator
    • getNode() => node
    • getParent() => TreeIterator
    • getChildren() => CQuery
    • TreeIterator sequential()
    • TreeIterator paralell()
  • TreeVisitor
    • visit(Consumer)
    • visitParent(Consumer)
    • visitChildren(Consumer)
    • TreeVisitor sequential()
    • TreeVisitor paralell()
We should implement at least two strategies: - BFS - DFS Inner visitor will be just BFS/DFS with skip(1), nothing more. We should expect at least two accessor interfaces: - TreeIterator - getNode() => node - getParent() => TreeIterator - getChildren() => CQuery<TreeIterator> - TreeIterator sequential() - TreeIterator paralell() - TreeVisitor - visit(Consumer<node>) - visitParent(Consumer<TreeVisitor>) - visitChildren(Consumer<TreeVisitor>) - TreeVisitor sequential() - TreeVisitor paralell()
ranides commented 4 years ago
Owner

Both strategies should offer configuration to stop recursive visits without filtering out nodes itself. That means, before creating CQuery, which offers only filtering out leaf results, we should support something like CTreeQuery, with additional methods like: dontRecurseIf (of course name should be better)

Both strategies should offer configuration to stop recursive visits without filtering out nodes itself. That means, before creating CQuery, which offers only filtering out leaf results, we should support something like CTreeQuery, with additional methods like: `dontRecurseIf` (of course name should be better)
ranides commented 4 years ago
Owner

bardzo bardzo pod znakiem zapytania, czy to trzeba generalizować w taki sposób? Po co? ANTLR w praktyce nie skorzysta z 99% funkcjonalności, XMLElement tak samo, innych tree nie będziemy wpinać. W praktyce każde tree ma jednak swoją strukturę i optymalizacje dot. zawartości, nie ma sensu robić ogólnego iteratora.

Na przykład wszystkie kryteria w ANTLR (po typie, po atrybutach) nie będą sprawdzane liniowo warunek po warunku, tylko mapowane in that or another way.

bardzo bardzo pod znakiem zapytania, czy to trzeba generalizować w taki sposób? Po co? ANTLR w praktyce nie skorzysta z 99% funkcjonalności, XMLElement tak samo, innych tree nie będziemy wpinać. W praktyce każde tree ma jednak swoją strukturę i optymalizacje dot. zawartości, nie ma sensu robić ogólnego iteratora. Na przykład wszystkie kryteria w ANTLR (po typie, po atrybutach) nie będą sprawdzane liniowo warunek po warunku, tylko mapowane in that or another way.
Sign in to join this conversation.
No Milestone
No assignee
1 Participants
Loading...
Cancel
Save
There is no content yet.