Merge remote-tracking branch 'origin/main' into feat/videos-only-mode

Reviewed-on: #5

README.md

@@ -21,24 +21,39 @@ go build -o canvasarchiver ./cmd/canvasarchiver
    ./canvasarchiver
    ```
 
-   Or 
+   Or for files-only mode (all files flat, no videos):
    ```bash
    ./canvasarchiver -fo
    ```
-   For files-only mode.
+
+   Or for videos-only mode (all Panopto videos flat, no files):
+   ```bash
+   ./canvasarchiver -vo
+   ```
 
 2. On first run, you'll be prompted to authenticate:
    - Visit the provided OAuth URL
    - Authorize the application
    - Copy the authorization code back to the terminal
 
-3. Enter your Course ID when prompted
+3. Enter your Course ID when prompted (or use `-me` to download all enrolled courses)
 
 4. The tool will download:
    - Regular course files (to `Course Files/`)
    - Module content (to `Modules/`)
    - Panopto recordings (to `Recordings/`)
 
+   In `-vo` mode, only videos are downloaded — all into the course root directory (no subdirectories).
+
+### Flags
+
+| Flag | Description |
+|------|-------------|
+| `-fo` | Files only — download all files flat into one directory; skips videos and module structure |
+| `-vo` | Videos only — scan recordings and all module video items, download everything flat into one directory; skips regular files |
+| `-me` | Download all enrolled courses |
+| `-n` | Prefix modules with order numbers `[1]`, `[2]`, etc. |
+
 
 ## Configuration
 

cmd/canvasarchiver/main.go

@@ -14,6 +14,9 @@ import (
 
 func main() {
 	filesOnly := flag.Bool("fo", false, "Files only mode - download all files to a single directory without module structure")
+	videosOnly := flag.Bool("vo", false, "Videos only mode - download only Panopto videos to a single directory")
+	me := flag.Bool("me", false, "Download all enrolled courses")
+	moduleNumbers := flag.Bool("n", false, "Prefix modules with order numbers [1], [2], etc.")
 	flag.Parse()
 
 	httpClient := &http.Client{}
@@ -25,11 +28,31 @@ func main() {
 		return
 	}
 
+	if *me {
+		canvasClient := canvas.NewClient(httpClient, accessToken, "", *filesOnly, *videosOnly, *moduleNumbers)
+		courses, err := canvasClient.GetEnrolledCourses()
+		if err != nil {
+			fmt.Printf("Error fetching courses: %v\n", err)
+			return
+		}
+
+		fmt.Printf("[+] Found %d enrolled courses\n", len(courses))
+		for _, course := range courses {
+			fmt.Printf("  -> Downloading: %s (ID: %d)\n", course.Name, course.ID)
+			downloadCourse(httpClient, accessToken, fmt.Sprintf("%d", course.ID), *filesOnly, *videosOnly, *moduleNumbers)
+		}
+		return
+	}
+
 	var courseID string
 	fmt.Print("Enter Course ID: ")
 	fmt.Scanln(&courseID)
 
-	canvasClient := canvas.NewClient(httpClient, accessToken, courseID, *filesOnly)
+	downloadCourse(httpClient, accessToken, courseID, *filesOnly, *videosOnly, *moduleNumbers)
+}
+
+func downloadCourse(httpClient *http.Client, accessToken, courseID string, filesOnly, videosOnly, moduleNumbers bool) {
+	canvasClient := canvas.NewClient(httpClient, accessToken, courseID, filesOnly, videosOnly, moduleNumbers)
 
 	if err := canvasClient.GetCourseInfo(); err != nil {
 		fmt.Printf("Error: %v\n", err)
@@ -44,7 +67,8 @@ func main() {
 
 	canvasClient.DownloadModules(courseRoot)
 
-	if !*filesOnly {
-		panopto.DownloadMainRecordings(httpClient, accessToken, courseID, courseRoot)
+	// Run recordings: always in -vo mode; skipped in -fo mode; normal otherwise.
+	if videosOnly || !filesOnly {
+		panopto.DownloadMainRecordings(httpClient, accessToken, courseID, courseRoot, videosOnly)
 	}
 }

internal/auth/auth.go

@@ -1,11 +1,17 @@
 package auth
 
 import (
+	"bufio"
+	"crypto/rand"
+	"crypto/sha256"
+	"encoding/base64"
 	"encoding/json"
 	"fmt"
+	"io"
 	"net/http"
 	"net/url"
 	"os"
+	"strings"
 
 	"git.directme.in/Joren/CanvasArchiver/internal/config"
 	"git.directme.in/Joren/CanvasArchiver/internal/models"
@@ -24,28 +30,7 @@ func NewAuthenticator(client *http.Client) *Authenticator {
 func (a *Authenticator) GetAccessToken() (string, error) {
 	creds, err := LoadCredentials()
 	if err != nil {
-
-		fmt.Println("--- Initial Canvas Login Required ---")
-		fmt.Printf("Visit: %s/login/oauth2/auth?client_id=%s&response_type=code&redirect_uri=%s\n",
-			config.BaseURL, config.ClientID, url.QueryEscape(config.RedirectURI))
-		fmt.Print("Enter Code: ")
-		var code string
-		fmt.Scanln(&code)
-
-		tr, err := a.doTokenRequest(url.Values{
-			"grant_type":    {"authorization_code"},
-			"client_id":     {config.ClientID},
-			"client_secret": {config.ClientSecret},
-			"redirect_uri":  {config.RedirectURI},
-			"code":          {code},
-		})
-		if err != nil {
-			return "", err
-		}
-
-		SaveCredentials(&models.Credentials{RefreshToken: tr.RefreshToken})
-		fmt.Println("[+] Login successful.")
-		return tr.AccessToken, nil
+		return a.login()
 	}
 
 	fmt.Println("[*] Reusing saved refresh token...")
@@ -54,11 +39,49 @@ func (a *Authenticator) GetAccessToken() (string, error) {
 		"client_id":     {config.ClientID},
 		"client_secret": {config.ClientSecret},
 		"refresh_token": {creds.RefreshToken},
+		"redirect_uri":  {config.RedirectURI},
+	})
+	if err != nil {
+		fmt.Printf("[!] Refresh failed: %v\n", err)
+		return a.login()
+	}
+	fmt.Println("[+] Session refreshed.")
+	return tr.AccessToken, nil
+}
+
+func (a *Authenticator) login() (string, error) {
+	codeVerifier, codeChallenge, err := generatePKCEPair()
+	if err != nil {
+		return "", err
+	}
+
+	fmt.Println("--- Initial Canvas Login Required ---")
+	fmt.Printf("Visit: %s\n", buildAuthURL(codeChallenge))
+	fmt.Print("Enter Code or redirect URL: ")
+	input, err := bufio.NewReader(os.Stdin).ReadString('\n')
+	if err != nil && err != io.EOF {
+		return "", err
+	}
+
+	code, err := extractCode(input)
+	if err != nil {
+		return "", err
+	}
+
+	tr, err := a.doTokenRequest(url.Values{
+		"grant_type":    {"authorization_code"},
+		"client_id":     {config.ClientID},
+		"client_secret": {config.ClientSecret},
+		"redirect_uri":  {config.RedirectURI},
+		"code":          {code},
+		"code_verifier": {codeVerifier},
 	})
 	if err != nil {
 		return "", err
 	}
-	fmt.Println("[+] Session refreshed.")
+
+	SaveCredentials(&models.Credentials{RefreshToken: tr.RefreshToken})
+	fmt.Println("[+] Login successful.")
 	return tr.AccessToken, nil
 }
 
@@ -70,14 +93,75 @@ func (a *Authenticator) doTokenRequest(v url.Values) (*models.TokenResponse, err
 	defer resp.Body.Close()
 
 	if resp.StatusCode != 200 {
-		return nil, fmt.Errorf("token request failed: %d", resp.StatusCode)
+		body, _ := io.ReadAll(resp.Body)
+		return nil, fmt.Errorf("token request failed: %d: %s", resp.StatusCode, string(body))
 	}
 
 	var tr models.TokenResponse
-	json.NewDecoder(resp.Body).Decode(&tr)
+	if err := json.NewDecoder(resp.Body).Decode(&tr); err != nil {
+		return nil, err
+	}
 	return &tr, nil
 }
 
+func buildAuthURL(codeChallenge string) string {
+	u, _ := url.Parse(config.BaseURL + "/login/oauth2/auth")
+	q := u.Query()
+	q.Set("client_id", config.ClientID)
+	q.Set("response_type", "code")
+	q.Set("mobile", "1")
+	q.Set("purpose", config.OAuthPurpose)
+	q.Set("code_challenge", codeChallenge)
+	q.Set("code_challenge_method", "S256")
+	q.Set("redirect_uri", config.RedirectURI)
+	if config.AuthProvider != "" {
+		q.Set("authentication_provider", config.AuthProvider)
+	}
+	u.RawQuery = q.Encode()
+	return u.String()
+}
+
+func generatePKCEPair() (string, string, error) {
+	randomBytes := make([]byte, 64)
+	if _, err := rand.Read(randomBytes); err != nil {
+		return "", "", err
+	}
+
+	verifier := base64.RawURLEncoding.EncodeToString(randomBytes)
+	challengeBytes := sha256.Sum256([]byte(verifier))
+	challenge := base64.RawURLEncoding.EncodeToString(challengeBytes[:])
+	return verifier, challenge, nil
+}
+
+func extractCode(input string) (string, error) {
+	input = strings.TrimSpace(input)
+	if input == "" {
+		return "", fmt.Errorf("empty authorization code")
+	}
+
+	for _, field := range strings.Fields(input) {
+		if strings.Contains(field, "code=") {
+			input = field
+			break
+		}
+	}
+
+	parsedURL, err := url.Parse(input)
+	if err == nil {
+		if code := parsedURL.Query().Get("code"); code != "" {
+			return code, nil
+		}
+	}
+
+	if values, err := url.ParseQuery(input); err == nil {
+		if code := values.Get("code"); code != "" {
+			return code, nil
+		}
+	}
+
+	return input, nil
+}
+
 func SaveCredentials(creds *models.Credentials) {
 	data, _ := json.MarshalIndent(creds, "", "  ")
 	os.WriteFile(config.CredsFile, data, 0o644)

internal/canvas/client.go

@@ -17,19 +17,25 @@ import (
 )
 
 type Client struct {
-	HTTPClient  *http.Client
-	AccessToken string
-	CourseID    string
-	CourseName  string
-	FilesOnly   bool
+	HTTPClient      *http.Client
+	AccessToken     string
+	CourseID        string
+	CourseName      string
+	FilesOnly       bool
+	VideosOnly      bool
+	ModuleNumbers   bool
+	downloadedFiles map[string]bool
 }
 
-func NewClient(httpClient *http.Client, accessToken, courseID string, filesOnly bool) *Client {
+func NewClient(httpClient *http.Client, accessToken, courseID string, filesOnly, videosOnly, moduleNumbers bool) *Client {
 	return &Client{
-		HTTPClient:  httpClient,
-		AccessToken: accessToken,
-		CourseID:    courseID,
-		FilesOnly:   filesOnly,
+		HTTPClient:      httpClient,
+		AccessToken:     accessToken,
+		CourseID:        courseID,
+		FilesOnly:       filesOnly,
+		VideosOnly:      videosOnly,
+		ModuleNumbers:   moduleNumbers,
+		downloadedFiles: make(map[string]bool),
 	}
 }
 
@@ -49,7 +55,25 @@ func (c *Client) GetCourseInfo() error {
 	return nil
 }
 
+func (c *Client) GetEnrolledCourses() ([]models.Course, error) {
+	req, _ := http.NewRequest("GET", fmt.Sprintf("%s/api/v1/courses?enrollment_state=active&per_page=100", config.BaseURL), nil)
+	req.Header.Set("Authorization", "Bearer "+c.AccessToken)
+	resp, err := c.HTTPClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	var courses []models.Course
+	json.NewDecoder(resp.Body).Decode(&courses)
+	return courses, nil
+}
+
 func (c *Client) DownloadCourseFiles(root string) {
+	if c.VideosOnly {
+		fmt.Println("\n[*] Skipping regular course files (videos only mode)")
+		return
+	}
 	fmt.Println("\n[*] Fetching regular course files...")
 
 	fReq, _ := http.NewRequest("GET", fmt.Sprintf("%s/api/v1/courses/%s/folders?per_page=100", config.BaseURL, c.CourseID), nil)
@@ -82,12 +106,23 @@ func (c *Client) DownloadCourseFiles(root string) {
 	fileCount := 0
 	for _, file := range files {
 
+		if c.FilesOnly {
+			if c.downloadedFiles[file.DisplayName] {
+				continue
+			}
+			c.downloadedFiles[file.DisplayName] = true
+		}
+
 		rawFolderPath := folderMap[file.FolderID]
 		safeFolderPath := utils.SanitizePath(rawFolderPath)
 
 		subDir := root
 		if !c.FilesOnly {
-			subDir = filepath.Join(root, "Course Files", safeFolderPath)
+			if safeFolderPath != "" && strings.ToLower(safeFolderPath) != "course files" {
+				subDir = filepath.Join(root, "Course Files", safeFolderPath)
+			} else {
+				subDir = filepath.Join(root, "Course Files")
+			}
 		}
 		os.MkdirAll(subDir, 0o755)
 		path := filepath.Join(subDir, utils.Sanitize(file.DisplayName))
@@ -129,15 +164,25 @@ func (c *Client) DownloadModules(courseRoot string) {
 	json.NewDecoder(resp.Body).Decode(&modules)
 	resp.Body.Close()
 
-	for _, mod := range modules {
+	for i, mod := range modules {
+		modName := mod.Name
+		if c.ModuleNumbers {
+			modName = fmt.Sprintf("[%d] %s", i+1, mod.Name)
+		}
+
+		// In videos-only mode everything goes flat into courseRoot.
+		// In files-only mode everything goes flat into courseRoot.
+		// Otherwise use the structured Modules/<name> path.
 		modBaseDir := courseRoot
-		if !c.FilesOnly {
-			modBaseDir = filepath.Join(courseRoot, "Modules", utils.Sanitize(mod.Name))
+		if !c.FilesOnly && !c.VideosOnly {
+			modBaseDir = filepath.Join(courseRoot, "Modules", utils.Sanitize(modName))
 		}
 		os.MkdirAll(modBaseDir, 0o755)
 
-		if !c.FilesOnly {
-			fmt.Printf("\n[Module] %s\n", mod.Name)
+		if !c.FilesOnly && !c.VideosOnly {
+			fmt.Printf("\n[Module] %s\n", modName)
+		} else if c.VideosOnly {
+			fmt.Printf("\n[Module] %s (scanning for videos)\n", modName)
 		}
 
 		subHeaderStack := []string{}
@@ -145,15 +190,16 @@ func (c *Client) DownloadModules(courseRoot string) {
 
 		for _, item := range mod.Items {
 
+			// In videos-only mode always download to the flat courseRoot.
 			targetDir := modBaseDir
-			if len(subHeaderStack) > 0 && !c.FilesOnly {
+			if len(subHeaderStack) > 0 && !c.FilesOnly && !c.VideosOnly {
 				targetDir = filepath.Join(modBaseDir, filepath.Join(subHeaderStack...))
 			}
 			os.MkdirAll(targetDir, 0o755)
 
 			switch item.Type {
 			case "SubHeader":
-				if c.FilesOnly {
+				if c.FilesOnly || c.VideosOnly {
 					continue
 				}
 				currentIndent := item.Indent
@@ -173,6 +219,9 @@ func (c *Client) DownloadModules(courseRoot string) {
 				fmt.Printf("%s--- %s ---\n", indent, item.Title)
 
 			case "File":
+				if c.VideosOnly {
+					continue
+				}
 				c.downloadModuleFile(item, targetDir)
 
 			case "ExternalTool":
@@ -223,11 +272,18 @@ func (c *Client) downloadModuleFile(item models.ModuleItem, dir string) {
 		return
 	}
 
+	if c.FilesOnly {
+		if c.downloadedFiles[fileMeta.DisplayName] {
+			return
+		}
+		c.downloadedFiles[fileMeta.DisplayName] = true
+	}
+
 	ext := filepath.Ext(fileMeta.DisplayName)
 	origBase := strings.TrimSuffix(fileMeta.DisplayName, ext)
 
 	fileName := fileMeta.DisplayName
-	if !strings.EqualFold(origBase, item.Title) && item.Title != "" {
+	if !c.FilesOnly && !strings.EqualFold(origBase, item.Title) && item.Title != "" {
 		fileName = fmt.Sprintf("%s (%s)%s", origBase, item.Title, ext)
 	}
 

internal/config/config.go

@@ -4,7 +4,9 @@ const (
 	BaseURL      = "https://canvas.vub.be"
 	ClientID     = "170000000000044"
 	ClientSecret = "3sxR3NtgXRfT9KdpWGAFQygq6O9RzLN021h2lAzhHUZEeSQ5XGV41Ddi5iutwW6f"
-	RedirectURI  = "urn:ietf:wg:oauth:2.0:oob"
+	RedirectURI  = "https://sso.canvaslms.com/canvas/login"
+	OAuthPurpose = "CanvasArchiver"
+	AuthProvider = "microsoft"
 	CredsFile    = "credentials.json"
 	PanoptoID    = "15"
 	UserAgent    = "Mozilla/5.0 (Linux; Android 10; K) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/138.0.0.0 Mobile Safari/537.36"

internal/models/models.go

@@ -10,6 +10,7 @@ type TokenResponse struct {
 }
 
 type Course struct {
+	ID   int    `json:"id"`
 	Name string `json:"name"`
 }
 

internal/panopto/downloader.go

@@ -17,6 +17,27 @@ import (
 	"git.directme.in/Joren/CanvasArchiver/internal/utils"
 )
 
+// normalizePanoptoURL converts the query-param form that Canvas stores
+// (List.aspx?folderID=X) to the fragment form that yt-dlp's PanoptoList
+// extractor understands (List.aspx#folderID="X"). Without this yt-dlp
+// ignores the folder filter and downloads the entire Panopto instance.
+func normalizePanoptoURL(rawURL string) string {
+	parsed, err := url.Parse(rawURL)
+	if err != nil {
+		return rawURL
+	}
+	if strings.Contains(parsed.Path, "List.aspx") {
+		folderID := parsed.Query().Get("folderID")
+		if folderID != "" {
+			// Strip query, set fragment: List.aspx#folderID="<id>"
+			parsed.RawQuery = ""
+			parsed.Fragment = fmt.Sprintf(`folderID="%s"`, folderID)
+			return parsed.String()
+		}
+	}
+	return rawURL
+}
+
 func getYoutubeDLCommand() string {
 	exePath, err := os.Executable()
 	if err == nil {
@@ -192,13 +213,35 @@ func DownloadVideo(httpClient *http.Client, accessToken, courseID, modDir, input
 		fmt.Printf("    [*] Downloading video: %s\n", title)
 
 		ytCmd := getYoutubeDLCommand()
-		cmd := exec.Command(ytCmd,
-			"--no-playlist",
-			"--cookies", cookieFile,
-			"--referer", config.BaseURL+"/",
-			"-P", modDir,
-			"-o", utils.Sanitize(title)+".%(ext)s",
-			targetURL)
+
+		// Normalize folder URLs so yt-dlp scopes to the right folder.
+		normalizedURL := normalizePanoptoURL(targetURL)
+
+		// Folder/list URLs are intentional playlists; don't pass --no-playlist.
+		isList := strings.Contains(normalizedURL, "List.aspx")
+		var outputTpl string
+		var args []string
+		if isList {
+			outputTpl = utils.Sanitize(title) + "/%(title)s.%(ext)s"
+			args = []string{
+				"--cookies", cookieFile,
+				"--referer", config.BaseURL + "/",
+				"-P", modDir,
+				"-o", outputTpl,
+				normalizedURL,
+			}
+		} else {
+			outputTpl = utils.Sanitize(title) + ".%(ext)s"
+			args = []string{
+				"--no-playlist",
+				"--cookies", cookieFile,
+				"--referer", config.BaseURL + "/",
+				"-P", modDir,
+				"-o", outputTpl,
+				normalizedURL,
+			}
+		}
+		cmd := exec.Command(ytCmd, args...)
 
 		cmd.Stdout = os.Stdout
 		cmd.Stderr = os.Stderr
@@ -212,7 +255,7 @@ func DownloadVideo(httpClient *http.Client, accessToken, courseID, modDir, input
 	}
 }
 
-func DownloadMainRecordings(httpClient *http.Client, accessToken, courseID, root string) {
+func DownloadMainRecordings(httpClient *http.Client, accessToken, courseID, root string, videosOnly bool) {
 	fmt.Println("\n[*] Checking for main Panopto recordings...")
 
 	jar, _ := cookiejar.New(nil)
@@ -300,8 +343,18 @@ func DownloadMainRecordings(httpClient *http.Client, accessToken, courseID, root
 		}
 		os.WriteFile(cookieFile, []byte(cData), 0o644)
 
-		recordingsDir := filepath.Join(root, "Recordings")
-		os.MkdirAll(recordingsDir, 0o755)
+		var downloadDir string
+		var outputTemplate string
+
+		if videosOnly {
+			// Flat: all videos go directly into course root
+			downloadDir = root
+			outputTemplate = "%(title)s.%(ext)s"
+		} else {
+			downloadDir = filepath.Join(root, "Recordings")
+			os.MkdirAll(downloadDir, 0o755)
+			outputTemplate = "%(playlist_title)s/%(title)s.%(ext)s"
+		}
 
 		fmt.Println("[*] Starting yt-dlp download for main recordings...")
 
@@ -310,8 +363,8 @@ func DownloadMainRecordings(httpClient *http.Client, accessToken, courseID, root
 		cmd := exec.Command(ytCmd,
 			"--cookies", cookieFile,
 			"--referer", config.BaseURL+"/",
-			"-P", recordingsDir,
-			"-o", "%(playlist_title)s/%(title)s.%(ext)s",
+			"-P", downloadDir,
+			"-o", outputTemplate,
 			decodedURL,
 		)
 		cmd.Stdout = os.Stdout
@@ -323,3 +376,4 @@ func DownloadMainRecordings(httpClient *http.Client, accessToken, courseID, root
 		fmt.Println("[!] No main recordings available or handshake failed")
 	}
 }
+