From ffd4745b1a129f2bdd5a8737ec79ec3137c50758 Mon Sep 17 00:00:00 2001
From: erius <egor_kapralov2003@mail.ru>
Date: Fri, 7 Jun 2024 02:21:38 +0300
Subject: [PATCH] Implemented externalsort

---
 externalsort/sort.go | 128 +++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 124 insertions(+), 4 deletions(-)

diff --git a/externalsort/sort.go b/externalsort/sort.go
index b9c05b3..b2a5cce 100644
--- a/externalsort/sort.go
+++ b/externalsort/sort.go
@@ -3,21 +3,141 @@
 package externalsort
 
 import (
+	"bufio"
+	"container/heap"
 	"io"
+	"os"
+	"sort"
+	"strings"
 )
 
+// lineReader is simply a wrapper around bufffered reader
+type lineReader struct {
+	br bufio.Reader
+}
+
+func (lr *lineReader) ReadLine() (string, error) {
+	s, err := lr.br.ReadString('\n')
+	if err != nil {
+		return s, err
+	}
+	s = s[:len(s)-1]
+	return s, err
+}
+
+type lineWriter struct {
+	w io.Writer
+}
+
+func (lw *lineWriter) Write(l string) error {
+	_, err := lw.w.Write(append([]byte(l), '\n'))
+	return err
+}
+
 func NewReader(r io.Reader) LineReader {
-	panic("implement me")
+	return &lineReader{*bufio.NewReader(r)}
 }
 
 func NewWriter(w io.Writer) LineWriter {
-	panic("implement me")
+	return &lineWriter{w}
+}
+
+// item for the reader heap
+// stores reader to get the next line from,
+// the line from reader which is used for sorting the heap
+// lastLine bool which determines whether the item has no more lines to read
+// and it can be safely removed from the heap
+type readerHeapItem struct {
+	reader   *LineReader
+	line     string
+	lastLine bool
+}
+
+type readerHeap []readerHeapItem
+
+func (h readerHeap) Len() int {
+	return len(h)
+}
+
+func (h readerHeap) Less(i, j int) bool {
+	return h[i].line < h[j].line
+}
+
+func (h readerHeap) Swap(i, j int) {
+	h[i], h[j] = h[j], h[i]
+}
+
+func (h *readerHeap) Push(x any) {
+	*h = append(*h, x.(readerHeapItem))
+}
+
+func (h *readerHeap) Pop() any {
+	old := *h
+	n := len(old)
+	x := old[n-1]
+	*h = old[0 : n-1]
+	return x
+}
+
+// push reader back to the heap if it still has lines
+// and update its lastLine value
+// otherwise do nothing
+// returns any errors that occured while reading (expect io.EOF)
+func pushIfNotLast(h *readerHeap, reader *LineReader) error {
+	line, err := (*reader).ReadLine()
+	if err != nil {
+		if err != io.EOF {
+			return err
+		}
+		heap.Push(h, readerHeapItem{reader, line, true})
+		return nil
+	}
+	heap.Push(h, readerHeapItem{reader, line, false})
+	return nil
 }
 
 func Merge(w LineWriter, readers ...LineReader) error {
-	panic("implement me")
+	h := make(readerHeap, 0, len(readers))
+	heap.Init(&h)
+	// push initial values to heap
+	for _, reader := range readers {
+		err := pushIfNotLast(&h, &reader)
+		if err != nil {
+			return err
+		}
+	}
+	// write lines from the top of the heap while
+	// the heap is not empty
+	for len(h) > 0 {
+		it := heap.Pop(&h).(readerHeapItem)
+		w.Write(it.line)
+		if it.lastLine {
+			continue
+		}
+		err := pushIfNotLast(&h, it.reader)
+		if err != nil {
+			return err
+		}
+	}
+	return nil
 }
 
 func Sort(w io.Writer, in ...string) error {
-	panic("implement me")
+	readers := make([]LineReader, 0, len(in))
+	for _, fn := range in {
+		b, err := os.ReadFile(fn)
+		if err != nil {
+			return err
+		}
+		s := strings.TrimSuffix(string(b), "\n")
+		// do not sort empty files
+		if len(s) == 0 {
+			continue
+		}
+		lines := strings.Split(s, "\n")
+		sort.Strings(lines)
+		s = strings.Join(lines, "\n")
+		readers = append(readers, NewReader(strings.NewReader(s)))
+	}
+	return Merge(NewWriter(w), readers...)
 }