gotreesitter

gotreesitter command

A command-line tool for the gotreesitter parser. It is pure Go, so it needs no C compiler.

Install

go install github.com/odvcencio/gotreesitter/cmd/gotreesitter@latest

This works from the first release that contains cmd/gotreesitter. The binary is about 34 MB because it embeds all grammars. Build with -tags 'grammar_subset grammar_subset_go' (or another grammar tag set) to embed fewer grammars and shrink it; binsize_test.go guards this.

Commands

gotreesitter parse main.go                      # syntax tree as an S-expression
gotreesitter parse -json config.json            # same, as JSON
gotreesitter query -q '(function_declaration name: (identifier) @name)' main.go
gotreesitter highlight main.go                  # byte range, capture, text
gotreesitter tags main.go                       # symbol definitions and references
gotreesitter outline main.go                    # nested symbol outline
gotreesitter grep 'fmt.Println($X)' main.go     # match a code pattern
gotreesitter grep -r 'log.Print($X)' 'fmt.Println($X)' main.go   # rewrite (add -w to write)
gotreesitter test --root ~/gts-corpora/grammar_parity --langs json,go
gotreesitter languages

• The language comes from the file name or shebang. Pass -l <name> to choose one. Use - as the file to read standard input (then -l is required).
• Put flags before the file name.
• Positions are 1-based line:column.
• tags and outline need a tags query. About 84 of the 206 languages have one. Other languages exit with code 2 and say so.
• test runs upstream tree-sitter corpus files and compares each tree. It replaces the old corpuscheck command. It exits 2 if no language ran or a language named in --langs was skipped.

Exit codes