When you parse a file with Tree-sitter, it produces a data structure called a concrete syntax tree (or CST), which is like an abstract syntax tree, except that it retains information about the location of every token in the document. For implementing these commands, we don't need any information other than the concrete syntax tree itself, because we don't care about specific types of syntax nodes, all we need to know is which ranges in the document correspond to nodes in the syntax tree. Tree queries let you write syntactic patterns in a simple, declarative language, and efficiently iterate over all of the groups of nodes in a syntax tree that match those patterns.