package main

import (
	"crypto/sha256"
	"flag"
	"fmt"
	"io"
	"log"
	"math"
	"os"
	"path/filepath"
	"regexp"
	"runtime"
	"strings"
)

/*
func writeCache(fn string) error {
}

func readCache(fn string) error {
}
*/

// sha256Sum Takes an io.Reader and computes the checksum returning it as a
// formatted string and an error if any
func sha256Sum(rdr io.Reader) (string, error) {
	h := sha256.New()
	if _, err := io.Copy(h, rdr); err != nil {
		return "", err
	}

	return fmt.Sprintf("%X", h.Sum(nil)), nil
}

// Takes a filepath and returns the sha256sum and an error if any
func getChecksum(fileName string) (string, error) {
	fh, err := os.Open(fileName)
	if err != nil {
		return "", err
	}
	defer fh.Close()

	return sha256Sum(fh)
}

type ChecksummerResult struct {
	Path     string
	Checksum string
	Error    error
	Info     os.FileInfo
}

type ChecksummerInput struct {
	Path string
	Info os.FileInfo
}

type ChecksumSlice []*ChecksummerResult

func (c ChecksumSlice) BuildRemoveList(reg *regexp.Regexp, matchPath bool) ChecksumSlice {
	removing := ChecksumSlice{}
	for _, result := range c {
		s := ""
		if !matchPath {
			s = result.Info.Name()
		} else {
			s = result.Path
		}

		if reg.MatchString(s) {
			removing = append(removing, result)
		}

	}
	return removing
}

// Worker that takes in a filepath off of a channel and outputs the checksum,
// and any errors
func Checksummer(done chan<- bool, input <-chan *ChecksummerInput, results chan<- *ChecksummerResult) {
	for i := range input {
		sum, err := getChecksum(i.Path)
		results <- &ChecksummerResult{
			Path:     i.Path,
			Checksum: sum,
			Error:    err,
			Info:     i.Info,
		}
	}
	done <- true
}

// Scans a path, dumps the results into the paths channel
func Scanner(done chan<- bool, all bool, path string, paths chan<- *ChecksummerInput) {
	err := filepath.Walk(path, func(path string, info os.FileInfo, err error) error {
		if err != nil {
			return err
		}

		if !all {
			if strings.Contains(path, "/.") {
				return nil
			}
		}

		if info.IsDir() {
			return nil
		}

		// paths <- path
		paths <- &ChecksummerInput{Path: path, Info: info}

		return nil
	})
	close(paths)
	if err != nil {
		fmt.Fprintln(os.Stderr, err)
	}
	done <- true
}

func formatBytes(b int64) string {
	if b < 1024 {
		return fmt.Sprintf("%d b", b)
	}

	s := ""

	pfxs := "kmgt"
	for i := 0; i < len(pfxs); i++ {
		pow := math.Pow(float64(1024), float64(i+1))
		// This one is too big, return the previous string
		if b < int64(pow) {
			return s
		}
		s = fmt.Sprintf("%.2f %cb",
			float64(b)/(pow),
			pfxs[i])
	}

	return s
}

func main() {
	fl := flag.NewFlagSet("deduplicator", flag.ExitOnError)

	path := fl.String("path", ".", "Path to deduplicate files in")
	all := fl.Bool("a", false, "Scan hidden files as well")
	script := fl.Bool("s", false, "Output format sutiable for scripts")
	procs := fl.Int("n", runtime.NumCPU(), "Number of processes to run at once")
	remove := fl.Bool("rm", false, "Remove duplicate files? ( Note requires a regex, and will not remove all versions, defaults to a dry run")
	removeRegexStr := fl.String("regex", "", "Regular expression to match duplicated files")
	removeYes := fl.Bool("yes-i-want-my-data-gone", false, "Actually remove the files")
	matchPath := fl.Bool("match-path", false, "match on the path, rather than the filename")

	_ = fl.Parse(os.Args[1:])

	if *remove && *removeRegexStr == "" {
		log.Fatal("A regular expression (-regex '<expr>') is required")
	}

	removeRegex, err := regexp.Compile(*removeRegexStr)
	if err != nil {
		log.Fatalf("Error compiling provided regular expression: %s", err)
	}

	checksums := make(map[string]ChecksumSlice)
	pths := make(chan *ChecksummerInput)
	results := make(chan *ChecksummerResult)
	done := make(chan bool)

	for j := 0; j < *procs; j++ {
		go Checksummer(done, pths, results)
	}

	go Scanner(done, *all, *path, pths)

	finished := 0
wait:
	for {
		select {
		case result := <-results:
			if result.Error != nil {
				fmt.Fprintln(os.Stderr, result.Error)
			}
			sum := result.Checksum

			if _, ok := checksums[sum]; !ok {
				checksums[sum] = []*ChecksummerResult{result}
			} else {
				checksums[sum] = append(
					checksums[sum],
					result)
			}
		case <-done:
			// fmt.Printf("reading from done... finished: %d\n", finished)
			finished++
			if finished >= *procs+1 {
				break wait
			}
		}

	}

	// fmt.Println("Actually removing: ", *removeYes)

	for sum, list := range checksums {
		if len(list) <= 1 {
			continue
		}

		switch {
		case *script && *remove:
			removing := list.BuildRemoveList(removeRegex, *matchPath)

			if len(removing) >= len(list) || len(removing) == 0 {
				for _, result := range list {
					fmt.Printf("%s::%d::%s::%s\n",
						sum,
						result.Info.Size(),
						result.Path,
						"not removing")
				}
				continue
			}

			if !*removeYes {
				for _, f := range removing {
					fmt.Printf("%s::%d::%s::%s\n",
						sum,
						f.Info.Size(),
						f.Path,
						"would remove")
				}
			} else {
				for _, f := range removing {
					fmt.Printf("%s::%d::%s::%s\n",
						sum,
						f.Info.Size(),
						f.Path,
						"removing")
					err := os.Remove(f.Path)
					if err != nil {
						fmt.Fprintln(os.Stderr, err)
					}
				}
			}
		case *remove:
			removing := list.BuildRemoveList(removeRegex, *matchPath)

			if len(removing) >= len(list) {
				fmt.Printf("%s: Not removing, matches all files\n", sum)
				continue
			}

			if len(removing) == 0 {
				fmt.Printf("%s: Not removing, no matches\n", sum)
				continue
			}

			if !*removeYes {
				for _, f := range removing {
					fmt.Printf("Would remove: %s\n", f.Path)
				}
			} else {
				for _, f := range removing {
					fmt.Printf("Removing: %s\n", f.Path)
					err := os.Remove(f.Path)
					if err != nil {
						fmt.Fprintln(os.Stderr, err)
					}
				}
			}
		case *script:
			for _, result := range list {
				fmt.Printf("%s::%d::%s\n", sum, result.Info.Size(), result.Path)
			}
		default:
			fmt.Println(sum)
			for i, result := range list {
				if i == 0 {
					fmt.Println(formatBytes(result.Info.Size()))
				}
				fmt.Println("\t" + result.Path)
			}
			fmt.Println("")
		}
	}

}