kbbi: The command line interface and Go library for kbbi.kemdikbud.go.id

1 files changed, 304 insertions, 0 deletions

diff --git a/client.go b/client.go
new file mode 100644
index 0000000..69401cd
--- /dev/null
+++ b/client.go
@@ -0,0 +1,304 @@
+// Copyright 2020, Shulhan <m.shulhan@gmail.com>. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package kbbi
+
+import (
+	"bytes"
+	"fmt"
+	"io/ioutil"
+	"log"
+	"net/http"
+	"net/http/cookiejar"
+	"net/url"
+	"strconv"
+	"strings"
+	"time"
+
+	"golang.org/x/net/html"
+	"golang.org/x/net/publicsuffix"
+)
+
+const (
+	maxPageNumber = 501
+	defTimeout    = 20 * time.Second
+)
+
+//
+// Client for KBBI web using HTTP.
+//
+type Client struct {
+	cookieURL *url.URL
+	httpc     *http.Client
+}
+
+//
+// New create and initialize new client for KBBI web.
+//
+func New(cookies []*http.Cookie) (cl *Client, err error) {
+	cookieURL, err := url.Parse(baseURL)
+	if err != nil {
+		return nil, fmt.Errorf("New: %w", err)
+	}
+
+	jarOpt := &cookiejar.Options{
+		PublicSuffixList: publicsuffix.List,
+	}
+
+	jar, err := cookiejar.New(jarOpt)
+	if err != nil {
+		return nil, err
+	}
+
+	if cookies != nil {
+		jar.SetCookies(cookieURL, cookies)
+	}
+
+	cl = &Client{
+		cookieURL: cookieURL,
+		httpc: &http.Client{
+			Jar:     jar,
+			Timeout: defTimeout,
+		},
+	}
+
+	return cl, nil
+}
+
+//
+// ListKataDasar get list of kata dasar
+//
+func (cl Client) ListKataDasar() (kataDasar daftarKata, err error) {
+	params := url.Values{
+		paramNameMasukan:        []string{paramValueDasar},
+		paramNameMasukanLengkap: []string{paramValueDasar},
+	}
+
+	urlPage := baseURL + "/Cari/Jenis?"
+
+	kataDasar = make(daftarKata)
+
+	for pageNumber := 1; pageNumber <= maxPageNumber; pageNumber++ {
+		params.Set(paramNamePage, strconv.Itoa(pageNumber))
+
+		req, err := http.NewRequest(http.MethodGet, urlPage+params.Encode(), nil)
+		if err != nil {
+			return kataDasar, err
+		}
+
+		res, err := cl.httpc.Do(req)
+		if err != nil {
+			return kataDasar, fmt.Errorf("ListKataDasar: page %d: %w",
+				pageNumber, err)
+		}
+
+		defer res.Body.Close()
+
+		body, err := ioutil.ReadAll(res.Body)
+		if err != nil {
+			return kataDasar, fmt.Errorf("ListKataDasar: page %d: %w",
+				pageNumber, err)
+		}
+
+		got, err := cl.parseHTMLKataDasar(body)
+		if err != nil {
+			return kataDasar, fmt.Errorf("ListKataDasar: page %d: %w",
+				pageNumber, err)
+		}
+		if len(got) == 0 {
+			break
+		}
+
+		kataDasar.merge(got)
+
+		log.Printf("ListKataDasar: halaman %d, jumlah kata %d, total kata %d",
+			pageNumber, len(got), len(kataDasar))
+	}
+
+	return kataDasar, nil
+}
+
+//
+// Login authenticate the client using username and password.
+//
+func (cl *Client) Login(user, pass string) (cookies []*http.Cookie, err error) {
+	tokenLogin, err := cl.preLogin()
+	if err != nil {
+		return nil, fmt.Errorf("Login: %w", err)
+	}
+
+	params := url.Values{
+		paramNameRequestVerificationToken: []string{tokenLogin},
+		paramNamePosel:                    []string{user},
+		paramNameKataSandi:                []string{pass},
+		paramNameIngatSaya:                []string{paramValueFalse},
+	}
+
+	reqBody := strings.NewReader(params.Encode())
+
+	req, err := http.NewRequest(http.MethodPost, loginURL, reqBody)
+	if err != nil {
+		return nil, fmt.Errorf("Login: %w", err)
+	}
+
+	req.Header.Set(headerNameContentType, headerValueContentType)
+
+	res, err := cl.httpc.Do(req)
+	if err != nil {
+		return nil, fmt.Errorf("Login: %w", err)
+	}
+
+	defer res.Body.Close()
+
+	resBody, err := ioutil.ReadAll(res.Body)
+	if err != nil {
+		return nil, fmt.Errorf("Login: %w", err)
+	}
+
+	if res.StatusCode >= http.StatusBadRequest {
+		return nil, fmt.Errorf("Login: %d %s", res.StatusCode, resBody)
+	}
+
+	cookies = cl.httpc.Jar.Cookies(cl.cookieURL)
+
+	return cookies, nil
+}
+
+//
+// SetCookies for HTTP request that need an authentication.
+//
+func (cl *Client) SetCookies(cookies []*http.Cookie) {
+	if len(cookies) > 0 {
+		cl.httpc.Jar.SetCookies(cl.cookieURL, cookies)
+	}
+}
+
+func (cl Client) parseHTMLKataDasar(htmlBody []byte) (kataDasar daftarKata, err error) {
+	node, err := html.Parse(bytes.NewReader(htmlBody))
+	if err != nil {
+		return nil, err
+	}
+
+	kataDasar = make(daftarKata)
+
+	var prev *html.Node
+
+	for {
+		switch {
+		case node.FirstChild != nil && node.FirstChild != prev && node.LastChild != prev:
+			node = node.FirstChild
+		case node.NextSibling != nil:
+			node = node.NextSibling
+		default:
+			prev = node
+			node = node.Parent
+		}
+		if node == nil {
+			break
+		}
+
+		if node.Type != html.ElementNode {
+			continue
+		}
+		if node.Data != elementTypeAnchor {
+			continue
+		}
+		for _, attr := range node.Attr {
+			if attr.Key != attrNameHref {
+				continue
+			}
+			if !strings.HasPrefix(attr.Val, entriURL) {
+				continue
+			}
+			k := strings.TrimSpace(node.FirstChild.Data)
+			kataDasar[k] = struct{}{}
+		}
+	}
+
+	return kataDasar, nil
+}
+
+//
+// parseHTMLLogin get the token at the form login.
+//
+func (cl Client) parseHTMLLogin(htmlBody []byte) (token string, err error) {
+	node, err := html.Parse(bytes.NewReader(htmlBody))
+	if err != nil {
+		return "", err
+	}
+
+	var prev *html.Node
+
+	for {
+		switch {
+		case node.FirstChild != nil && node.FirstChild != prev && node.LastChild != prev:
+			node = node.FirstChild
+		case node.NextSibling != nil:
+			node = node.NextSibling
+		default:
+			prev = node
+			node = node.Parent
+		}
+		if node == nil {
+			break
+		}
+
+		if node.Type != html.ElementNode {
+			continue
+		}
+		if node.Data != elementTypeInput {
+			continue
+		}
+		for _, attr := range node.Attr {
+			if attr.Key != attrNameName {
+				continue
+			}
+
+			token = getAttrValue(node.Attr)
+			if len(token) > 0 {
+				return token, nil
+			}
+		}
+	}
+
+	return "", fmt.Errorf("token login not found")
+}
+
+//
+// preLogin initialize the client to get the first cookie.
+//
+func (cl *Client) preLogin() (token string, err error) {
+	req, err := http.NewRequest(http.MethodGet, loginURL, nil)
+	if err != nil {
+		return "", err
+	}
+
+	res, err := cl.httpc.Do(req)
+	if err != nil {
+		return "", err
+	}
+
+	defer res.Body.Close()
+
+	body, err := ioutil.ReadAll(res.Body)
+	if err != nil {
+		return "", err
+	}
+
+	token, err = cl.parseHTMLLogin(body)
+	if err != nil {
+		return "", err
+	}
+
+	return token, nil
+}
+
+func getAttrValue(attrs []html.Attribute) string {
+	for _, attr := range attrs {
+		if attr.Key == attrNameValue {
+			return attr.Val
+		}
+	}
+	return ""
+}