feat: Benchmark

benchmark/bfs.go

@@ -0,0 +1,60 @@
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"time"
+)
+
+type Site struct {
+	Title      string   `json:"title"`
+	Body       string   `json:"body"`
+	Links      []int    `json:"links"`
+	ParentSite string   `json:"parent_site"`
+	SiteURL    string   `json:"site_url"`
+	LinkURL    []string `json:"link_urls"`
+	Slug       string   `json:"slug"`
+}
+
+func main() {
+	// fmt.Println("Hello, World!")
+	jsonFile, err := os.Open("../site/data/site_data.json")
+	if err != nil {
+		fmt.Println(err)
+	}
+	// fmt.Println("opened jsonfile")
+
+	defer jsonFile.Close()
+	byteValue, _ := ioutil.ReadAll(jsonFile)
+	var sites []Site
+	json.Unmarshal(byteValue, &sites)
+
+	var visited [1000000]int
+	t1 := time.Now()
+	for i := 0; i < len(sites); i++ {
+		if visited[i] == 0 {
+			queue := []int{i}
+			queueBeg := 0
+			for queueBeg < len(queue) {
+
+				for size := len(queue) - queueBeg; size > 0; size-- {
+					front := queue[queueBeg]
+					queueBeg++
+					if visited[front] == 1 {
+						continue
+					}
+					visited[front] = 1
+					// fmt.Println("visiting", front)
+					for j := 0; j < len(sites[front].Links); j++ {
+						queue = append(queue, sites[front].Links[j])
+					}
+				}
+			}
+		}
+	}
+	t2 := time.Now()
+	elapsed := t2.Sub(t1)
+	fmt.Println(elapsed)
+}

benchmark/dfs.go

@@ -0,0 +1,58 @@
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"time"
+)
+
+type Site struct {
+	Title      string   `json:"title"`
+	Body       string   `json:"body"`
+	Links      []int    `json:"links"`
+	ParentSite string   `json:"parent_site"`
+	SiteURL    string   `json:"site_url"`
+	LinkURL    []string `json:"link_urls"`
+	Slug       string   `json:"slug"`
+}
+
+func main() {
+	// fmt.Println("Hello, World!")
+	jsonFile, err := os.Open("../site/data/site_data.json")
+	if err != nil {
+		fmt.Println(err)
+	}
+	// fmt.Println("opened jsonfile")
+
+	defer jsonFile.Close()
+	byteValue, _ := ioutil.ReadAll(jsonFile)
+	var sites []Site
+	json.Unmarshal(byteValue, &sites)
+
+	var visited [1000000]int
+	var stack []int
+	t1 := time.Now()
+	for i := 0; i < len(sites); i++ {
+		if visited[i] == 0 {
+			stack = append(stack, i)
+			for len(stack) > 0 {
+				top := stack[len(stack)-1]
+				// pop here
+				stack = stack[:len(stack)-1]
+				if visited[top] == 0 {
+					visited[top] = 1
+					// fmt.Println("visiting", top)
+					for j := 0; j < len(sites[top].Links); j++ {
+						stack = append(stack, sites[top].Links[j])
+					}
+				}
+			}
+
+		}
+	}
+	t2 := time.Now()
+	elapsed := t2.Sub(t1)
+	fmt.Println(elapsed)
+}

benchmark/map.go

@@ -0,0 +1,165 @@
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"regexp"
+	"runtime"
+	"strings"
+	//"time"
+)
+
+type Site struct {
+	Title      string   `json:"title"`
+	Body       string   `json:"body"`
+	Links      []int    `json:"links"`
+	ParentSite string   `json:"parent_site"`
+	SiteURL    string   `json:"site_url"`
+	LinkURL    []string `json:"link_urls"`
+	Slug       string   `json:"slug"`
+}
+
+const (
+	ALBHABET_SIZE = 26
+)
+
+/*
+type trieNode struct {
+	childrens [ALBHABET_SIZE]*trieNode
+	isWordEnd bool
+}
+
+type trie struct {
+	root *trieNode
+}
+
+func initTrie() *trie {
+	return &trie{
+		root: &trieNode{},
+	}
+}
+
+func (t *trie) insert(word string) {
+	wordLength := len(word)
+	current := t.root
+	for i := 0; i < wordLength; i++ {
+		index := word[i] - 'a'
+		if current.childrens[index] == nil {
+			current.childrens[index] = &trieNode{}
+		}
+		current = current.childrens[index]
+	}
+	current.isWordEnd = true
+}
+
+func (t *trie) find(word string) bool {
+	wordLength := len(word)
+	current := t.root
+	for i := 0; i < wordLength; i++ {
+		index := word[i] - 'a'
+		if current.childrens[index] == nil {
+			return false
+		}
+		current = current.childrens[index]
+	}
+	if current.isWordEnd {
+		return true
+	}
+	return false
+}
+*/
+func PrintMemUsage() {
+	var m runtime.MemStats
+	runtime.ReadMemStats(&m)
+	// For info on each, see: https://golang.org/pkg/runtime/#MemStats
+	fmt.Printf("HeapAlloc = %v MiB", bToMb(m.HeapAlloc))
+	fmt.Printf("\tTotalAlloc = %v MiB", bToMb(m.TotalAlloc))
+	fmt.Printf("\tSys = %v MiB", bToMb(m.Sys))
+	fmt.Printf("\tNumGC = %v\n", m.NumGC)
+}
+
+func bToMb(b uint64) uint64 {
+	return b / 1024 / 1024
+}
+
+func main() {
+	jsonFile, err := os.Open("../site/data/site_data.json")
+	if err != nil {
+		fmt.Println(err)
+	}
+
+	defer jsonFile.Close()
+	byteValue, _ := ioutil.ReadAll(jsonFile)
+	var sites []Site
+	json.Unmarshal(byteValue, &sites)
+
+	myMap := make(map[string]bool)
+	//	trie := initTrie()
+	//var sum time.Duration
+
+	//no_of_words, err := strconv.Atoi(os.Args[1])
+	//k := 0
+	//fl := 0
+	runtime.GC()
+	PrintMemUsage()
+	for i := 0; i < len(sites); i++ {
+		sentence := strings.Split(sites[i].Body, " ")
+		for j := 0; j < len(sentence); j++ {
+			sentence[j] = strings.ToLower(sentence[j])
+			matched, _ := regexp.MatchString(`^[a-z]*$`, sentence[j])
+			// fmt.Println(matched, sentence[j])
+			// t1 := time.Now()
+			if matched {
+				//trie.insert(sentence[j])
+				//	myMap[sentence[j]] = append(myMap[sentence[j]], sites[i].SiteURL)
+				myMap[sentence[j]] = true
+				//k++
+			}
+			/*if k >= no_of_words+100 {
+				fl = 1
+				break
+			}
+			*/
+			// t2 := time.Now()
+			// elapsed := t2.Sub(t1)
+			// sum += elapsed
+		}
+	}
+	runtime.GC()
+	PrintMemUsage()
+
+	//	t1 := time.Now()
+	/*
+		k = 0
+		fl = 0
+		for i := 0; i < len(sites); i++ {
+			sentence := strings.Split(sites[i].Body, " ")
+			for j := 0; j < len(sentence); j++ {
+				sentence[j] = strings.ToLower(sentence[j])
+				matched, _ := regexp.MatchString(`^[a-z]*$`, sentence[j])
+				if matched {
+					t1 := time.Now()
+					if myMap[sentence[j]] {
+					}
+					t2 := time.Now()
+					elapsed := t2.Sub(t1)
+					sum += elapsed
+					k++
+					if k >= no_of_words {
+						fl = 1
+						break
+					}
+				}
+			}
+			if fl == 1 {
+				break
+			}
+		}
+	*/
+	//fmt.Println(sum)
+
+	//	t2 := time.Now()
+	//	fmt.Println(elapsed)
+}