feat: add support for handling unsupported stream storage in download process

bot/handle_start.go

@@ -1,6 +1,8 @@
 package bot
 
 import (
+	"fmt"
+
 	"github.com/celestix/gotgproto/dispatcher"
 	"github.com/celestix/gotgproto/ext"
 	"github.com/krau/SaveAny-Bot/common"
@@ -17,6 +19,7 @@ func start(ctx *ext.Context, update *ext.Update) error {
 
 const helpText string = `
 Save Any Bot - 转存你的 Telegram 文件
+版本: %s , 提交: %s
 命令:
 /start - 开始使用
 /help - 显示帮助
@@ -32,6 +35,6 @@ Save Any Bot - 转存你的 Telegram 文件
 `
 
 func help(ctx *ext.Context, update *ext.Update) error {
-	ctx.Reply(update, ext.ReplyTextString(helpText), nil)
+	ctx.Reply(update, ext.ReplyTextString(fmt.Sprintf(helpText, common.Version, common.GitCommit[:7])), nil)
 	return dispatcher.EndGroups
 }

bot/utils.go

@@ -9,6 +9,7 @@ import (
 
 	"github.com/celestix/gotgproto/dispatcher"
 	"github.com/celestix/gotgproto/ext"
+	"github.com/gabriel-vasile/mimetype"
 	"github.com/gotd/td/telegram/message/entity"
 	"github.com/gotd/td/telegram/message/styling"
 	"github.com/gotd/td/tg"
@@ -200,13 +201,7 @@ func FileFromMessage(ctx *ext.Context, chatID int64, messageID int, customFileNa
 }
 
 func GetTGMessage(ctx *ext.Context, chatId int64, messageID int) (*tg.Message, error) {
-	key := fmt.Sprintf("message:%d:%d", chatId, messageID)
 	common.Log.Debugf("Fetching message: %d", messageID)
-	var cachedMessage tg.Message
-	err := common.Cache.Get(key, &cachedMessage)
-	if err == nil {
-		return &cachedMessage, nil
-	}
 	messages, err := ctx.GetMessages(chatId, []tg.InputMessageClass{&tg.InputMessageID{ID: messageID}})
 	if err != nil {
 		return nil, err
@@ -219,9 +214,6 @@ func GetTGMessage(ctx *ext.Context, chatId int64, messageID int) (*tg.Message, e
 	if !ok {
 		return nil, fmt.Errorf("unexpected message type: %T", msg)
 	}
-	if err := common.Cache.Set(key, tgMessage, 3600); err != nil {
-		common.Log.Errorf("Failed to cache message: %s", err)
-	}
 	return tgMessage, nil
 }
 
@@ -286,6 +278,19 @@ func GenFileNameFromMessage(message tg.Message, file *types.File) string {
 	if file.FileName != "" {
 		return file.FileName
 	}
+	fileName := genFileNameFromMessageText(message, file)
+	media, ok := message.GetMedia()
+	if !ok {
+		return fileName
+	}
+	ext, ok := extraMediaExt(media)
+	if ok {
+		return fileName + ext
+	}
+	return fileName
+}
+
+func genFileNameFromMessageText(message tg.Message, file *types.File) string {
 	text := strings.TrimSpace(message.GetMessage())
 	if text == "" {
 		return file.Hash()
@@ -297,3 +302,21 @@ func GenFileNameFromMessage(message tg.Message, file *types.File) string {
 	runes := []rune(text)
 	return string(runes[:min(128, len(runes))])
 }
+
+func extraMediaExt(media tg.MessageMediaClass) (string, bool) {
+	switch media := media.(type) {
+	case *tg.MessageMediaDocument:
+		doc, ok := media.Document.AsNotEmpty()
+		if !ok {
+			return "", false
+		}
+		ext := mimetype.Lookup(doc.MimeType).Extension()
+		if ext == "" {
+			return "", false
+		}
+		return ext, true
+	case *tg.MessageMediaPhoto:
+		return ".jpg", true
+	}
+	return "", false
+}

common/cache.go

@@ -21,8 +21,6 @@ func initCache() {
 	gob.Register(types.File{})
 	gob.Register(tg.InputDocumentFileLocation{})
 	gob.Register(tg.InputPhotoFileLocation{})
-	gob.Register(tg.Message{})
-	gob.Register(tg.PeerUser{})
 	Cache = &CommonCache{cache: freecache.NewCache(10 * 1024 * 1024)}
 }
 

core/core.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 
 	"github.com/celestix/gotgproto/ext"
+	"github.com/gotd/td/telegram/downloader"
 	"github.com/gotd/td/tg"
 	"github.com/krau/SaveAny-Bot/common"
 	"github.com/krau/SaveAny-Bot/config"
@@ -13,6 +14,12 @@ import (
 	"github.com/krau/SaveAny-Bot/types"
 )
 
+var Downloader *downloader.Downloader
+
+func init() {
+	Downloader = downloader.NewDownloader().WithPartSize(1024 * 1024)
+}
+
 func worker(queue *queue.TaskQueue, semaphore chan struct{}) {
 	for {
 		semaphore <- struct{}{}

core/download.go

@@ -59,41 +59,50 @@ func processPendingTask(task *types.Task) error {
 
 	downloadBuilder := Downloader.Download(bot.Client.API(), task.File.Location).WithThreads(getTaskThreads(task.File.FileSize))
 
+	notsupportStreamStorage, notsupportStream := taskStorage.(storage.StorageNotSupportStream)
+	cancelMarkUp := getCancelTaskMarkup(task)
 	if config.Cfg.Stream {
+		if !notsupportStream {
+			text, entities := buildProgressMessageEntity(task, 0, task.StartTime, 0)
+			ctx.EditMessage(task.ReplyChatID, &tg.MessagesEditMessageRequest{
+				Message:     text,
+				Entities:    entities,
+				ID:          task.ReplyMessageID,
+				ReplyMarkup: cancelMarkUp,
+			})
 
-		text, entities := buildProgressMessageEntity(task, 0, task.StartTime, 0)
-		ctx.EditMessage(task.ReplyChatID, &tg.MessagesEditMessageRequest{
-			Message:     text,
-			Entities:    entities,
-			ID:          task.ReplyMessageID,
-			ReplyMarkup: getCancelTaskMarkup(task),
-		})
+			pr, pw := io.Pipe()
+			defer pr.Close()
 
-		pr, pw := io.Pipe()
-		defer pr.Close()
+			task.StartTime = time.Now()
+			progressCallback := buildProgressCallback(ctx, task, getProgressUpdateCount(task.File.FileSize))
 
-		task.StartTime = time.Now()
-		progressCallback := buildProgressCallback(ctx, task, getProgressUpdateCount(task.File.FileSize))
+			progressStream := NewProgressStream(pw, task.File.FileSize, progressCallback)
 
-		progressStream := NewProgressStream(pw, task.File.FileSize, progressCallback)
+			eg, uploadCtx := errgroup.WithContext(cancelCtx)
 
-		eg, uploadCtx := errgroup.WithContext(cancelCtx)
-
-		eg.Go(func() error {
-			return taskStorage.Save(uploadCtx, pr, task.StoragePath)
-		})
-		eg.Go(func() error {
-			_, err := downloadBuilder.Stream(uploadCtx, progressStream)
-			if closeErr := pw.CloseWithError(err); closeErr != nil {
-				common.Log.Errorf("Failed to close pipe writer: %v", closeErr)
+			eg.Go(func() error {
+				return taskStorage.Save(uploadCtx, pr, task.StoragePath)
+			})
+			eg.Go(func() error {
+				_, err := downloadBuilder.Stream(uploadCtx, progressStream)
+				if closeErr := pw.CloseWithError(err); closeErr != nil {
+					common.Log.Errorf("Failed to close pipe writer: %v", closeErr)
+				}
+				return err
+			})
+			if err := eg.Wait(); err != nil {
+				return err
 			}
-			return err
-		})
-		if err := eg.Wait(); err != nil {
-			return err
-		}
 
-		return nil
+			return nil
+		}
+		common.Log.Warnf("存储 %s 不支持流式传输: %s", task.StorageName, notsupportStreamStorage.NotSupportStream())
+		ctx.EditMessage(task.ReplyChatID, &tg.MessagesEditMessageRequest{
+			Message:     fmt.Sprintf("存储 %s 不支持流式传输: %s\n正在使用普通下载...", task.StorageName, notsupportStreamStorage.NotSupportStream()),
+			ID:          task.ReplyMessageID,
+			ReplyMarkup: cancelMarkUp,
+		})
 	}
 
 	cacheDestPath := filepath.Join(config.Cfg.Temp.BasePath, task.FileName())
@@ -110,7 +119,7 @@ func processPendingTask(task *types.Task) error {
 		Message:     text,
 		Entities:    entities,
 		ID:          task.ReplyMessageID,
-		ReplyMarkup: getCancelTaskMarkup(task),
+		ReplyMarkup: cancelMarkUp,
 	})
 
 	progressCallback := buildProgressCallback(ctx, task, getProgressUpdateCount(task.File.FileSize))
@@ -188,6 +197,13 @@ func processTelegraph(extCtx *ext.Context, cancelCtx context.Context, task *type
 				common.Log.Errorf("Failed to unmarshal element: %s", err)
 				continue
 			}
+
+			if len(node.Children) != 0 {
+				for _, child := range node.Children {
+					imgs = append(imgs, getNodeImages(child)...)
+				}
+			}
+
 			if node.Tag == "img" {
 				if src, ok := node.Attrs["src"]; ok {
 					imgs = append(imgs, src)

core/download_test.go

@@ -0,0 +1,80 @@
+package core
+
+import (
+	"reflect"
+	"testing"
+
+	"github.com/celestix/telegraph-go/v2"
+)
+
+func TestGetImgSrcs(t *testing.T) {
+	complexStructure := telegraph.NodeElement{
+		Tag: "div",
+		Children: []telegraph.Node{
+			telegraph.NodeElement{
+				Tag: "figure",
+				Children: []telegraph.Node{
+					telegraph.NodeElement{
+						Tag: "img",
+						Attrs: map[string]string{
+							"src": "https://example.com/image1.png",
+						},
+					},
+					telegraph.NodeElement{
+						Tag: "p",
+						Children: []telegraph.Node{
+							"A text node",
+						},
+					},
+					telegraph.NodeElement{
+						Tag: "figure",
+						Children: []telegraph.Node{
+							telegraph.NodeElement{
+								Tag: "img",
+								Attrs: map[string]string{
+									"src": "https://example.com/image2.png",
+								},
+							},
+						},
+					},
+				},
+			},
+			telegraph.NodeElement{
+				Tag: "img",
+				Attrs: map[string]string{
+					"src": "https://example.com/image3.png",
+				},
+			},
+			"text node",
+			telegraph.NodeElement{
+				Tag: "div",
+				Children: []telegraph.Node{
+					telegraph.NodeElement{
+						Tag: "span",
+						Children: []telegraph.Node{
+							telegraph.NodeElement{
+								Tag: "img",
+								Attrs: map[string]string{
+									"src": "https://example.com/image4.png",
+								},
+							},
+						},
+					},
+				},
+			},
+		},
+	}
+
+	expected := []string{
+		"https://example.com/image1.png",
+		"https://example.com/image2.png",
+		"https://example.com/image3.png",
+		"https://example.com/image4.png",
+	}
+
+	got := getNodeImages(complexStructure)
+
+	if !reflect.DeepEqual(expected, got) {
+		t.Errorf("expected %v，got %v", expected, got)
+	}
+}

core/downloader.go

@@ -1,9 +0,0 @@
-package core
-
-import "github.com/gotd/td/telegram/downloader"
-
-var Downloader *downloader.Downloader
-
-func init() {
-	Downloader = downloader.NewDownloader().WithPartSize(1024 * 1024)
-}

core/utils.go

@@ -3,6 +3,7 @@ package core
 import (
 	"bytes"
 	"context"
+	"encoding/json"
 	"fmt"
 	"io"
 	"os"
@@ -10,6 +11,7 @@ import (
 	"time"
 
 	"github.com/celestix/gotgproto/ext"
+	"github.com/celestix/telegraph-go/v2"
 	"github.com/gabriel-vasile/mimetype"
 	"github.com/gotd/td/telegram/message/entity"
 	"github.com/gotd/td/telegram/message/styling"
@@ -22,22 +24,33 @@ import (
 )
 
 func saveFileWithRetry(ctx context.Context, storagePath string, taskStorage storage.Storage, cacheFilePath string) error {
+	file, err := os.Open(cacheFilePath)
+	if err != nil {
+		return fmt.Errorf("failed to open cache file: %w", err)
+	}
+	defer file.Close()
+	fileStat, err := file.Stat()
+	if err != nil {
+		return fmt.Errorf("failed to get file stat: %w", err)
+	}
+	vctx := context.WithValue(ctx, types.ContextKeyContentLength, fileStat.Size())
 	for i := 0; i <= config.Cfg.Retry; i++ {
-		if err := ctx.Err(); err != nil {
+		if err := vctx.Err(); err != nil {
 			return fmt.Errorf("context canceled while saving file: %w", err)
 		}
 		file, err := os.Open(cacheFilePath)
 		if err != nil {
 			return fmt.Errorf("failed to open cache file: %w", err)
 		}
-		if err := taskStorage.Save(ctx, file, storagePath); err != nil {
+		defer file.Close()
+		if err := taskStorage.Save(vctx, file, storagePath); err != nil {
 			if i == config.Cfg.Retry {
 				return fmt.Errorf("failed to save file: %w", err)
 			}
 			common.Log.Errorf("Failed to save file: %s, retrying...", err)
 			select {
-			case <-ctx.Done():
-				return fmt.Errorf("context canceled during retry delay: %w", ctx.Err())
+			case <-vctx.Done():
+				return fmt.Errorf("context canceled during retry delay: %w", vctx.Err())
 			case <-time.After(time.Duration(i*500) * time.Millisecond):
 			}
 			continue
@@ -256,3 +269,27 @@ func NewProgressStream(writer io.Writer, size int64, callback func(bytesRead, co
 		interval: interval,
 	}
 }
+
+func getNodeImages(node telegraph.Node) []string {
+	var srcs []string
+
+	var nodeElement telegraph.NodeElement
+	data, err := json.Marshal(node)
+	if err != nil {
+		return srcs
+	}
+	err = json.Unmarshal(data, &nodeElement)
+	if err != nil {
+		return srcs
+	}
+
+	if nodeElement.Tag == "img" {
+		if src, exists := nodeElement.Attrs["src"]; exists {
+			srcs = append(srcs, src)
+		}
+	}
+	for _, child := range nodeElement.Children {
+		srcs = append(srcs, getNodeImages(child)...)
+	}
+	return srcs
+}

docs/docs/help.md

@@ -4,10 +4,14 @@
 
 Bot 接受两种消息: 文件和链接.
 
-目前, 链接仅支持公开频道 (具有用户名) 的链接, 例如: `https://t.me/acherkrau/1097`.
+支持以下链接:
+
+1. 公开频道 (具有用户名) 的消息链接, 例如: `https://t.me/acherkrau/1097`.
 
 **即使频道禁止了转发和保存, Bot 依然可以下载其文件.**
 
+2. Telegra.ph 的文章链接, Bot 将下载其中的所有图片
+
 ## 静默模式 (silent)
 
 使用 `/silent` 命令可以开关静默模式.
@@ -32,4 +36,6 @@ Bot 接受两种消息: 文件和链接.
 - 网络不稳定时, 任务失败率高.
 - 无法在中间层对文件进行处理, 例如自动文件类型识别.
 
-虽然目前 Bot 适配的所有存储端 (Alist, 本地磁盘, Webdav) 都支持 Stream 模式, 但今后可能会有不支持的存储端, 此时即使开启 Stream 模式, Bot 也会自动切换到普通模式.
+**不支持** Stream 模式的存储端:
+
+- alist

storage/alist/alist.go

@@ -106,6 +106,12 @@ func (a *Alist) Save(ctx context.Context, reader io.Reader, storagePath string)
 	req.Header.Set("Authorization", a.token)
 	req.Header.Set("File-Path", url.PathEscape(storagePath))
 	req.Header.Set("Content-Type", "application/octet-stream")
+	if length := ctx.Value(types.ContextKeyContentLength); length != nil {
+		length, ok := length.(int64)
+		if ok {
+			req.ContentLength = length
+		}
+	}
 
 	resp, err := a.client.Do(req)
 	if err != nil {
@@ -134,6 +140,10 @@ func (a *Alist) Save(ctx context.Context, reader io.Reader, storagePath string)
 	return nil
 }
 
+func (a *Alist) NotSupportStream() string {
+	return "Alist does not support chunked transfer encoding"
+}
+
 func (a *Alist) JoinStoragePath(task types.Task) string {
 	return path.Join(a.config.BasePath, task.StoragePath)
 }

storage/storage.go

@@ -23,6 +23,11 @@ type Storage interface {
 	Save(ctx context.Context, reader io.Reader, storagePath string) error
 }
 
+type StorageNotSupportStream interface {
+	Storage
+	NotSupportStream() string
+}
+
 var Storages = make(map[string]Storage)
 
 var UserStorages = make(map[int64][]Storage)

storage/webdav/client.go

@@ -6,6 +6,8 @@ import (
 	"io"
 	"net/http"
 	"strings"
+
+	"github.com/krau/SaveAny-Bot/types"
 )
 
 type Client struct {
@@ -38,6 +40,11 @@ func (c *Client) doRequest(ctx context.Context, method, url string, body io.Read
 	if c.Username != "" && c.Password != "" {
 		req.SetBasicAuth(c.Username, c.Password)
 	}
+	if length := ctx.Value(types.ContextKeyContentLength); length != nil {
+		if l, ok := length.(int64); ok {
+			req.ContentLength = l
+		}
+	}
 	return c.httpClient.Do(req)
 }
 

types/task.go

@@ -0,0 +1,82 @@
+package types
+
+import (
+	"context"
+	"crypto/md5"
+	"encoding/hex"
+	"fmt"
+	"net/url"
+	"strings"
+	"time"
+
+	"github.com/gotd/td/tg"
+)
+
+type Task struct {
+	Ctx         context.Context
+	Cancel      context.CancelFunc
+	Error       error
+	Status      TaskStatus
+	StorageName string
+	StoragePath string
+	StartTime   time.Time
+
+	File          *File
+	FileMessageID int
+	FileChatID    int64
+
+	IsTelegraph  bool
+	TelegraphURL string
+
+	// to track the reply message
+	ReplyMessageID int
+	ReplyChatID    int64
+	UserID         int64
+}
+
+func (t Task) Key() string {
+	if t.IsTelegraph {
+		return hashStr(t.TelegraphURL)
+	}
+	return fmt.Sprintf("%d:%d", t.FileChatID, t.FileMessageID)
+}
+
+func (t Task) String() string {
+	if t.IsTelegraph {
+		return fmt.Sprintf("[telegraph]:%s", t.TelegraphURL)
+	}
+	return fmt.Sprintf("[%d:%d]:%s", t.FileChatID, t.FileMessageID, t.File.FileName)
+}
+
+func (t Task) FileName() string {
+	if t.IsTelegraph {
+		tgphPath := strings.Split(t.TelegraphURL, "/")[len(strings.Split(t.TelegraphURL, "/"))-1]
+		tgphPathUnescaped, err := url.PathUnescape(tgphPath)
+		if err != nil {
+			return tgphPath
+		}
+		return tgphPathUnescaped
+	}
+	return t.File.FileName
+}
+
+type File struct {
+	Location tg.InputFileLocationClass
+	FileSize int64
+	FileName string
+}
+
+func (f File) Hash() string {
+	locationBytes := []byte(f.Location.String())
+	fileSizeBytes := []byte(fmt.Sprintf("%d", f.FileSize))
+	fileNameBytes := []byte(f.FileName)
+
+	structBytes := append(locationBytes, fileSizeBytes...)
+	structBytes = append(structBytes, fileNameBytes...)
+
+	hash := md5.New()
+	hash.Write(structBytes)
+	hashBytes := hash.Sum(nil)
+
+	return hex.EncodeToString(hashBytes)
+}

types/types.go

@@ -1,20 +1,8 @@
 package types
 
-import (
-	"context"
-	"crypto/md5"
-	"encoding/hex"
-	"fmt"
-	"net/url"
-	"strings"
-	"time"
-
-	"github.com/gotd/td/tg"
-)
-
 type TaskStatus string
 
-var (
+const (
 	Pending   TaskStatus = "pending"
 	Succeeded TaskStatus = "succeeded"
 	Failed    TaskStatus = "failed"
@@ -23,7 +11,7 @@ var (
 
 type StorageType string
 
-var (
+const (
 	StorageTypeLocal  StorageType = "local"
 	StorageTypeWebdav StorageType = "webdav"
 	StorageTypeAlist  StorageType = "alist"
@@ -38,71 +26,8 @@ var StorageTypeDisplay = map[StorageType]string{
 	StorageTypeMinio:  "Minio",
 }
 
-type Task struct {
-	Ctx         context.Context
-	Cancel      context.CancelFunc
-	Error       error
-	Status      TaskStatus
-	StorageName string
-	StoragePath string
-	StartTime   time.Time
+type ContextKey string
 
-	File          *File
-	FileMessageID int
-	FileChatID    int64
-
-	IsTelegraph  bool
-	TelegraphURL string
-
-	// to track the reply message
-	ReplyMessageID int
-	ReplyChatID    int64
-	UserID         int64
-}
-
-func (t Task) Key() string {
-	if t.IsTelegraph {
-		return hashStr(t.TelegraphURL)
-	}
-	return fmt.Sprintf("%d:%d", t.FileChatID, t.FileMessageID)
-}
-
-func (t Task) String() string {
-	if t.IsTelegraph {
-		return fmt.Sprintf("[telegraph]:%s", t.TelegraphURL)
-	}
-	return fmt.Sprintf("[%d:%d]:%s", t.FileChatID, t.FileMessageID, t.File.FileName)
-}
-
-func (t Task) FileName() string {
-	if t.IsTelegraph {
-		tgphPath := strings.Split(t.TelegraphURL, "/")[len(strings.Split(t.TelegraphURL, "/"))-1]
-		tgphPathUnescaped, err := url.PathUnescape(tgphPath)
-		if err != nil {
-			return tgphPath
-		}
-		return tgphPathUnescaped
-	}
-	return t.File.FileName
-}
-
-type File struct {
-	Location tg.InputFileLocationClass
-	FileSize int64
-	FileName string
-}
-
-func (f File) Hash() string {
-	locationBytes := []byte(f.Location.String())
-	fileSizeBytes := []byte(fmt.Sprintf("%d", f.FileSize))
-	fileNameBytes := []byte(f.FileName)
-
-	structBytes := append(locationBytes, fileSizeBytes...)
-	structBytes = append(structBytes, fileNameBytes...)
-
-	hash := md5.New()
-	hash.Write(structBytes)
-	hashBytes := hash.Sum(nil)
-
-	return hex.EncodeToString(hashBytes)
-}
+const (
+	ContextKeyContentLength ContextKey = "content-length"
+)