アーカイブをアクセス #4 & CLIモード #3

getpage.go

@@ -5,8 +5,15 @@ import (
   "fmt"
   "net/http"
   "io"
+  "strings"
 )
 
+func stripurl (url string) string {
+  res := strings.ReplaceAll(url, "?", "")
+  res = strings.ReplaceAll(res, "=", "")
+  return res
+}
+
 func getpage (url string, path string) {
   curl, err := http.Get(url)
   if err != nil {

main.go

@@ -31,7 +31,8 @@ func main () {
       return
     } else {
       if checkprefix(args[1]) {
-        exist := checkexist(args[1], cnf.datapath)
+        eurl := stripurl(args[1])
+        exist := checkexist(eurl, cnf.datapath)
         var confirm string
         if len(exist) > 0 {
           fmt.Println("このページが既に保存されているみたいです。")
@@ -42,10 +43,9 @@ func main () {
           fmt.Scanf("%s", &confirm)
         }
         if len(exist) == 0 || confirm == "y" || confirm == "Y" {
-          path := mkdirs(args[1], cnf.datapath)
+          path := mkdirs(eurl, cnf.datapath)
           getpage(args[1], path)
-          // TODO: ページの確認
-          scanpage(path, args[1], cnf.datapath)
+          scanpage(path, eurl, cnf.datapath)
           fmt.Println(cnf.domain + strings.Replace(path, cnf.datapath, "", 1))
         }
         return

scanpage.go

@@ -5,16 +5,17 @@ import (
   "fmt"
   "strings"
   "net/http"
+  "net/url"
   "io"
   "regexp"
   "errors"
+  "path/filepath"
 )
 
 func scanpage (path string, domain string, thisdomain string) error {
   fn, err := os.ReadFile(path + "/index.html")
   if err != nil {
-    fmt.Println(err)
-    return errors.New("ファイルを開けられなかった：")
+    return err
   }
 
   /* 削除 */
@@ -24,74 +25,131 @@ func scanpage (path string, domain string, thisdomain string) error {
   var video = regexp.MustCompile(`(<video.*</video>)`).ReplaceAllString(string(audio), "")
   var iframe = regexp.MustCompile(`(<iframe.*</iframe>)`).ReplaceAllString(string(video), "")
   /* 追加ダウンロード＋ローカル化 */
-  var ass = regexp.MustCompile(`(<img.*src="|<meta.*content="|<link.*href=")(.*\.)(png|webm|jpg|jpeg|gif|css|js)`)
+  var ass = regexp.MustCompile(`(<img.*src=['"]|<meta.*content=['"]|<link.*href=['"])(.*\.)(png|webm|jpg|jpeg|gif|css|js|ico|svg|tiff|woff2)(\?[^'"]*)?`)
+
   spath := "static/"
   if !strings.HasSuffix(path, "/") {
     spath = "/" + spath
   }
   spath = path + spath
-  err1 := os.Mkdir(spath, 0755)
-  if err1 != nil {
-    fmt.Println(err1)
-    return errors.New("失敗：")
+  err = os.Mkdir(spath, 0755)
+  if err != nil {
+    return err
   }
 
+  repmap := make(map[string]string)
+
   for _, cssx := range ass.FindAllString(iframe, -1) {
-    s := regexp.MustCompile(`(.*src="|.*content="|.*href=")`).Split(cssx, -1)
-    ss := regexp.MustCompile(`(".*)`).Split(s[1], -1)
-    if strings.HasPrefix(ss[0], "http://") || strings.HasPrefix(ss[0], "https://") {
-      // TODO
-    } else {
-      fss := strings.Split(ss[0], "/")
-      filename := fss[len(fss)-1]
+    s := regexp.MustCompile(`(.*src=['"]|.*content=['"]|.*href=['"])`).Split(cssx, -1)
+    ss := regexp.MustCompile(`(['"].*)`).Split(s[1], -1)
 
-      if filename == "" {
-        continue
-      }
-      f, err := os.Create(spath + filename)
-      if err != nil {
-        fmt.Println(err)
-        return errors.New("2. 作成失敗：")
-      }
-      defer f.Close()
-
-      af := domain + ss[0]
-      if !strings.HasPrefix(ss[0], "/") {
-        af = domain + "/" + ss[0]
-      }
-      i, err := http.Get(af)
-      if err != nil {
-        fmt.Println(err)
-        return errors.New("2. ダウンロードに失敗：")
-      }
-      defer i.Body.Close()
-      if strings.HasSuffix(filename, "css") || strings.HasSuffix(filename, "js") {
-        body, err := io.ReadAll(i.Body)
-        if err != nil {
-          fmt.Println(err)
-          return errors.New("2. 読込エラー：")
-        }
-
-        _, err2 := f.WriteString(string(body))
-        if err2 != nil {
-          fmt.Println(err)
-          return errors.New("2. ファイル書込エラー：")
-        }
-      } else {
-        _, err = io.Copy(f, i.Body)
-        if err != nil {
-          fmt.Println(err)
-          return errors.New("2. コピーに失敗：")
-        }
-      }
-
-      iframe = strings.Replace(iframe, ss[0], "/static/" + filename, -1)
+    if strings.HasPrefix(ss[0], "//") {
+      ss[0] = "https:" + ss[0]
+    }
+
+    fss := strings.Split(ss[0], "/")
+    assdom := ""
+    filename := fss[len(fss)-1]
+
+    if strings.HasPrefix(ss[0], "http://") || strings.HasPrefix(ss[0], "https://") {
+      assdom = fss[2]
+    }
+
+    asspath := path + "/static/" + assdom
+    err = os.MkdirAll(asspath, 0755)
+    if err != nil {
+      return err
+    }
+
+    if filename == "" {
+      continue
+    }
+
+    if strings.HasPrefix(ss[0], "http://") || strings.HasPrefix(ss[0], "https://") {
+      err = dlres(ss[0], filepath.Join(asspath, filename))
+      if err != nil {
+        return err
+      }
+    } else {
+      af := domain
+      if strings.HasPrefix(ss[0], "/") {
+        af = af + ss[0]
+      } else {
+        af = af + "/" + ss[0]
+      }
+      err = dlres(af, filepath.Join(asspath, filename))
+      if err != nil {
+        return err
+      }
+    }
+
+    repmap[ss[0]] = filepath.Join("/static", assdom, filename)
+    if assdom == "" {
+      repmap[ss[0]] = filepath.Join("/static", filename)
     }
 
-    err := os.WriteFile(path + "/index.html", []byte(iframe), 0644)
     if err != nil {
       fmt.Println(err)
-      return errors.New("書込に失敗")
+      return errors.New("2. ダウンロードに失敗：")
+    }
+  }
+
+  for ourl, lurl := range repmap {
+    aurl := strings.ReplaceAll(path, thisdomain, "") + stripver(lurl)
+    iframe = strings.ReplaceAll(iframe, ourl, aurl)
+  }
+
+  err = os.WriteFile(path + "/index.html", []byte(iframe), 0644)
+  if err != nil {
+    fmt.Println(err)
+    return errors.New("書込に失敗")
+  }
+
+  return nil
+}
+
+func stripver (durl string) string {
+  u, err := url.Parse(durl)
+  if err != nil {
+    fmt.Println("エラー：", err)
+    return ""
+  }
+  u.RawQuery = ""
+  return u.Path
+}
+
+func dlres (durl string, dest string) error {
+  // ダウンロード
+  res, err := http.Get(stripver(durl))
+  if err != nil {
+    return err
+  }
+  defer res.Body.Close()
+
+  dest = stripver(dest)
+
+  // ファイルを作成
+  f, err := os.Create(dest)
+  if err != nil {
+    return err
+  }
+  defer f.Close()
+
+  // ファイルを書き込む
+  if strings.HasSuffix(dest, "css") || strings.HasSuffix(dest, "js") {
+    body, err := io.ReadAll(res.Body)
+    if err != nil {
+      return err
+    }
+
+    _, err2 := f.WriteString(string(body))
+    if err2 != nil {
+      return err
+    }
+  } else {
+    _, err = io.Copy(f, res.Body)
+    if err != nil {
+      return err
     }
   }
 

srv.go

@@ -35,27 +35,83 @@ func initloc (r *http.Request) string {
   }
 }
 
-func serv (cnf Config, port int) {
-  http.Handle("/static/", http.StripPrefix("/static/", http.FileServer(http.Dir("static"))))
-  ftmpl := []string{cnf.webpath + "/view/index.html", cnf.webpath + "/view/header.html", cnf.webpath + "/view/footer.html"}
-  data := &Page{Ver: version, Ves: strings.ReplaceAll(version, ".", "")}
+func siteHandler (cnf Config) func (http.ResponseWriter, *http.Request) {
+  return func (w http.ResponseWriter, r *http.Request) {
+    ftmpl := []string{cnf.webpath + "/view/index.html", cnf.webpath + "/view/header.html", cnf.webpath + "/view/footer.html"}
+    data := &Page{Ver: version, Ves: strings.ReplaceAll(version, ".", "")}
 
-  http.HandleFunc("/api", func(w http.ResponseWriter, r *http.Request) {
-    w.Header().Set("Content-Type", "application/json; charset=UTF-8")
-    w.WriteHeader(200)
-    buf, _ := json.MarshalIndent(&Stat{Url: cnf.domain, Ver: version}, "", "  ")
-    _, _ = w.Write(buf)
-  })
-
-  http.HandleFunc("/", func(w http.ResponseWriter, r *http.Request) {
     lang := initloc(r)
-    pth := r.URL.Path
 
     data.Lan = lang
     ftmpl[0] = cnf.webpath + "/view/index.html"
     tmpl := template.Must(template.ParseFiles(ftmpl[0], ftmpl[1], ftmpl[2]))
 
-    if strings.HasPrefix(r.URL.Path, "/archive") {
+    data.Tit = getloc("top", lang)
+    if r.Method == "POST" {
+      err := r.ParseForm()
+      if err != nil { fmt.Println(err) }
+
+      // クッキー
+      if r.PostForm.Get("langchange") != "" {
+        cookie, err := r.Cookie("lang")
+        if err != nil || cookie.Value == "ja" {
+          http.SetCookie(w, &http.Cookie {Name: "lang", Value: "en", MaxAge: 31536000, Path: "/"})
+        } else {
+          http.SetCookie(w, &http.Cookie {Name: "lang", Value: "ja", MaxAge: 31536000, Path: "/"})
+        }
+        http.Redirect(w, r, "/", http.StatusSeeOther)
+        return
+      }
+
+      if r.PostForm.Get("hozonsite") != "" {
+        fmt.Println("sasa")
+        url := r.PostForm.Get("hozonsite")
+        fmt.Println(url)
+        // HTTPかHTTPSじゃない場合
+        if !checkprefix(url) {
+          data.Err = getloc("errfuseiurl", lang)
+          ftmpl[0] = cnf.webpath + "/view/404.html"
+        } else {
+          //if r.PostForm.Get("sosin") != "" {}
+        }
+      }
+    }
+
+    tmpl = template.Must(template.ParseFiles(ftmpl[0], ftmpl[1], ftmpl[2]))
+    tmpl.Execute(w, data)
+  }
+}
+
+func apiHandler (cnf Config) func (http.ResponseWriter, *http.Request) {
+  return func (w http.ResponseWriter, r *http.Request) {
+    w.Header().Set("Content-Type", "application/json; charset=UTF-8")
+    w.WriteHeader(200)
+    buf, _ := json.MarshalIndent(&Stat{Url: cnf.domain, Ver: version}, "", "  ")
+    _, _ = w.Write(buf)
+  }
+}
+
+func archiveHandler (cnf Config) func (http.ResponseWriter, *http.Request) {
+  return func (w http.ResponseWriter, r *http.Request) {
+    ftmpl := []string{cnf.webpath + "/view/index.html", cnf.webpath + "/view/header.html", cnf.webpath + "/view/footer.html"}
+    data := &Page{Ver: version, Ves: strings.ReplaceAll(version, ".", "")}
+    lang := initloc(r)
+
+    data.Lan = lang
+    ftmpl[0] = cnf.webpath + "/view/index.html"
+    tmpl := template.Must(template.ParseFiles(ftmpl[0], ftmpl[1], ftmpl[2]))
+    path := strings.TrimPrefix(r.URL.Path, "/archive/")
+
+    if strings.Contains(path, "/static/") {
+      if !strings.HasSuffix(path, ".css") && !strings.HasSuffix(path, ".png") && !strings.HasSuffix(path, ".jpeg") && !strings.HasSuffix(path, ".jpg") && !strings.HasSuffix(path, ".webm") && !strings.HasSuffix(path, ".gif") && !strings.HasSuffix(path, ".js") {
+        http.NotFound(w, r)
+        return
+      }
+
+      fpath := cnf.datapath + "/archive/" + path
+      http.ServeFile(w, r, fpath)
+    } else {
+      pth := r.URL.Path
       if !strings.HasSuffix(pth, "/") && !strings.HasSuffix(pth, "index.html") {
         pth += "/index.html"
       } else if strings.HasSuffix(pth, "/") && !strings.HasSuffix(pth, "index.html") {
@@ -72,40 +128,16 @@ func serv (cnf Config, port int) {
       tmpl = template.Must(template.ParseFiles(cnf.webpath + "/view/archive.html"))
       tmpl.Execute(w, data)
       data = nil
-    } else {
-      data.Tit = getloc("top", lang)
-      if r.Method == "POST" {
-        err := r.ParseForm()
-        if err != nil { fmt.Println(err) }
-
-        // クッキー
-        if r.PostForm.Get("langchange") != "" {
-          cookie, err := r.Cookie("lang")
-          if err != nil || cookie.Value == "ja" {
-            http.SetCookie(w, &http.Cookie {Name: "lang", Value: "en", MaxAge: 31536000, Path: "/"})
-          } else {
-            http.SetCookie(w, &http.Cookie {Name: "lang", Value: "ja", MaxAge: 31536000, Path: "/"})
-          }
-          http.Redirect(w, r, "/", http.StatusSeeOther)
-          return
-        }
-
-        if r.PostForm.Get("hozonsite") != "" {
-          url := r.PostForm.Get("hozonsite")
-          // HTTPかHTTPSじゃない場合
-          if !checkprefix(url) {
-            data.Err = getloc("errfuseiurl", lang)
-            ftmpl[0] = cnf.webpath + "/view/404.html"
-          } else {
-            //if r.PostForm.Get("sosin") != "" {}
-          }
-        }
-      }
     }
+  }
+}
 
-    tmpl = template.Must(template.ParseFiles(ftmpl[0], ftmpl[1], ftmpl[2]))
-    tmpl.Execute(w, data)
-  })
+func serv (cnf Config, port int) {
+  http.Handle("/static/", http.StripPrefix("/static/", http.FileServer(http.Dir("static"))))
+
+  http.HandleFunc("/api/", apiHandler(cnf))
+  http.HandleFunc("/archive/", archiveHandler(cnf))
+  http.HandleFunc("/", siteHandler(cnf))
 
   fmt.Println(fmt.Sprint("http://127.0.0.1:", port, " でサーバーを実行中。終了するには、CTRL+Cを押して下さい。"))
   http.ListenAndServe(fmt.Sprint(":", port), nil)