consolidated go and scala repos into this repo as well

go/README.md

@@ -0,0 +1,5 @@
+# go-rosalind
+
+go-rosalind is a command-line utility for running solutions to [Rosalind.info](rosalind.info) problems that were written in go.
+
+This is mostly just used by me to consolidate some repeated code and practice writing stuff in Go.

go/cmd/cons.go

@@ -0,0 +1,19 @@
+package cmd
+
+import (
+	"github.com/spf13/cobra"
+	"gitlab.com/fizzizist/go-rosalind/pkg/solutions"
+)
+
+var consCmd = &cobra.Command{
+	Use:   "cons",
+	Short: "Finds a Motif in input DNA",
+	Args:  cobra.MinimumNArgs(1),
+	Run: func(cmd *cobra.Command, args []string) {
+		solutions.Cons(args[0])
+	},
+}
+
+func init() {
+	rootCmd.AddCommand(consCmd)
+}

go/cmd/root.go

@@ -0,0 +1,12 @@
+package cmd
+
+import "github.com/spf13/cobra"
+
+var rootCmd = &cobra.Command{
+	Use:   "go-rosalind",
+	Short: "An app for running Rosalind solutions",
+}
+
+func Execute() error {
+	return rootCmd.Execute()
+}

go/cmd/subs.go

@@ -0,0 +1,19 @@
+package cmd
+
+import (
+	"github.com/spf13/cobra"
+	"gitlab.com/fizzizist/go-rosalind/pkg/solutions"
+)
+
+var dnaMotifsCmd = &cobra.Command{
+	Use:   "subs",
+	Short: "Finds a Motif in input DNA",
+	Args:  cobra.MinimumNArgs(1),
+	Run: func(cmd *cobra.Command, args []string) {
+		solutions.FindMotifs(args[0])
+	},
+}
+
+func init() {
+	rootCmd.AddCommand(dnaMotifsCmd)
+}

go/main.go

@@ -0,0 +1,15 @@
+package main
+
+import (
+	"fmt"
+	"os"
+
+	"gitlab.com/fizzizist/go-rosalind/cmd"
+)
+
+func main() {
+	if err := cmd.Execute(); err != nil {
+		fmt.Println(err)
+		os.Exit(1)
+	}
+}

go/pkg/io/file_tools.go

@@ -0,0 +1,86 @@
+package io
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"strings"
+)
+
+func FileToStringArray(filename string) ([]string, error) {
+	file, err := os.Open(filename)
+	if err != nil {
+		return nil, err
+	}
+	defer file.Close()
+
+	scanner := bufio.NewScanner(file)
+	scanner.Split(bufio.ScanLines)
+
+	var fileLines []string
+
+	for scanner.Scan() {
+		fileLines = append(fileLines, scanner.Text())
+	}
+	if err := scanner.Err(); err != nil {
+		return nil, err
+	}
+
+	file.Close()
+
+	return fileLines, nil
+}
+
+func WriteStringsToFile(content []string, filepath string) error {
+	file, err := os.Create(filepath)
+	if err != nil {
+		return err
+	}
+
+	w := bufio.NewWriter(file)
+
+	for _, line := range content {
+		_, err = fmt.Fprintln(w, line)
+		if err != nil {
+			return err
+		}
+	}
+
+	if err = w.Flush(); err != nil {
+		return err
+	}
+	if err = file.Close(); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+// ParseFasta parses a .fasta formatted file returning a list of labels and a list of
+// content in separate string arrays.
+func ParseFasta(filename string) ([]string, []string, error) {
+	lines, err := FileToStringArray(filename)
+	if err != nil {
+		return nil, nil, err
+	}
+
+	var labels []string
+	var content []string
+	var currString []string
+	for _, line := range lines {
+		if line[0] == '>' {
+			labels = append(labels, line)
+			if len(currString) > 0 {
+				content = append(content, strings.Join(currString, ""))
+				currString = []string{}
+			}
+			continue
+		}
+		currString = append(currString, line)
+	}
+	if len(currString) > 0 {
+		content = append(content, strings.Join(currString, ""))
+	}
+
+	return labels, content, nil
+}

go/pkg/solutions/cons.go

@@ -0,0 +1,76 @@
+package solutions
+
+import (
+	"fmt"
+	"log"
+	"strconv"
+	"strings"
+
+	"gitlab.com/fizzizist/go-rosalind/pkg/io"
+)
+
+func CompileOutput(consensus string, profile [][]int) []string {
+	output := []string{consensus}
+	idxMap := []string{"A", "C", "G", "T"}
+
+	for i, val := range profile {
+		var strProf []string
+		for _, inVal := range val {
+			strProf = append(strProf, strconv.Itoa(inVal))
+		}
+		output = append(output, fmt.Sprintf("%s: %s", idxMap[i], strings.Join(strProf, " ")))
+	}
+
+	return output
+}
+
+func Cons(filename string) {
+	_, lines, err := io.ParseFasta(filename)
+	if err != nil {
+		log.Fatalf("Failed to read file: %s", err)
+	}
+
+	profile := make([][]int, 4) // 1 for each codon
+	for i := range profile {
+		profile[i] = make([]int, len(lines[1]))
+	}
+
+	for _, value := range lines {
+		for j, s := range value {
+			switch {
+			case s == 'A':
+				profile[0][j]++
+			case s == 'C':
+				profile[1][j]++
+			case s == 'G':
+				profile[2][j]++
+			case s == 'T':
+				profile[3][j]++
+			}
+		}
+	}
+
+	var consensus []string
+	idxMap := []string{"A", "C", "G", "T"}
+	for i := range profile[0] {
+		maxCons := "A"
+		maxVal := 0
+		for j, s := range idxMap {
+			if profile[j][i] > maxVal {
+				maxCons = s
+				maxVal = profile[j][i]
+			}
+		}
+		consensus = append(consensus, maxCons)
+	}
+
+	fmt.Println("Writing output file...")
+	err = io.WriteStringsToFile(
+		CompileOutput(strings.Join(consensus, ""), profile),
+		"results/cons.txt",
+	)
+	if err != nil {
+		log.Fatalf("Failed writing output file: %s", err)
+	}
+	fmt.Println("Done.")
+}

go/pkg/solutions/subs.go

@@ -0,0 +1,33 @@
+package solutions
+
+import (
+	"fmt"
+	"log"
+	"strconv"
+	"strings"
+
+	"gitlab.com/fizzizist/go-rosalind/pkg/io"
+)
+
+func FindMotifs(filename string) {
+	lines, err := io.FileToStringArray(filename)
+	if err != nil {
+		log.Fatalf("Failed to read file: %s", err)
+	}
+	if len(lines) != 2 {
+		log.Fatalf("Input file should have exactly 2 lines, a DNA string and a motif")
+	}
+
+	dna := lines[0]
+	motif := lines[1]
+	motLen := len(motif)
+	var positions []string
+
+	for i := 0; i <= len(dna)-motLen; i++ {
+		if motif == dna[i:i+motLen] {
+			positions = append(positions, strconv.Itoa(i+1))
+		}
+	}
+	fmt.Println(strings.Join(positions, " "))
+
+}

go/testing/cons.fasta

@@ -0,0 +1,14 @@
+>Rosalind_1
+ATCCAGCT
+>Rosalind_2
+GGGCAACT
+>Rosalind_3
+ATGGATCT
+>Rosalind_4
+AAGCAACC
+>Rosalind_5
+TTGGAACT
+>Rosalind_6
+ATGCCATT
+>Rosalind_7
+ATGGCACT

go/testing/subs.txt

@@ -0,0 +1,2 @@
+GATATATGCATATACTT
+ATAT