repeat_formatter.go

package main

import (
	"bufio"
	"fmt"
	"log"
	"os"
	"strings"
)

type Bounds struct {
	lower string
	upper string
}

func main() {

	if len(os.Args) != 2 {
		log.Fatal("   Invalid command-line args. Correct Usage:\n./repeat_formatter /path/to/file > outputfile.txt")
	}

	filename := os.Args[1]

	seqs := getData(filename)
	outputData(seqs)
}

func getData(filename string) map[string][]Bounds {
	file, err := os.Open(filename)
	if err != nil {
		log.Fatal(err)
	}
	defer file.Close()

	seqs := make(map[string][]Bounds)

	scanner := bufio.NewScanner(file)

	currentName := ""

	for scanner.Scan() {
		line := scanner.Text()
		if strings.HasPrefix(line, "@") {
			fields := strings.Split(line, " ")
			name := fields[0][1:]
			seqs[name] = make([]Bounds, 0)
			currentName = name
		} else {
			fields := strings.Split(line, " ")
			if len(fields) < 2 {
				continue
			}
			seqs[currentName] = append(seqs[currentName], Bounds{fields[0], fields[1]})
		}
	}

	return seqs
}

func isBase(str string) bool {
	bases := []string{"A", "C", "G", "T"}
	for _, b := range bases {
		if strings.HasPrefix(str, b) {
			return false
		}
	}
	return true
}

func outputData(seqs map[string][]Bounds) {
	out := ""

	for key, val := range seqs {
		for _, b := range val {
			out += key + "\t" + b.lower + "\t" + b.upper + "\n"
		}
	}

	fmt.Println(out)
}