Fix --speak-diacritics (#12)

.github/FUNDING.yml

@@ -1 +1,2 @@
 github: [c-loftus]
+custom: ["https://www.paypal.com/paypalme/coltonloftus"]

.github/workflows/go_tests.yml

@@ -0,0 +1,32 @@
+name: Go Test
+
+on: 
+  push:
+    # only trigger on branches, not on tags
+    branches: '**'
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v2
+
+      - name: Set up Go
+        uses: actions/setup-go@v2
+        with:
+          go-version: '1.22'
+
+      - name: Install Calibre
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y calibre
+
+      - name: Ensure ebook-convert is in PATH
+        run: |
+          echo "$(dirname $(which ebook-convert)) is in PATH"
+          ebook-convert --version  # This will fail if ebook-convert is not installed correctly
+
+      - name: Run Go tests
+        run: go test ./... -p 1 -count=1

.vscode/launch.json

@@ -4,13 +4,22 @@
     // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
     "version": "0.2.0",
     "configurations": [
+
         {
-            "name": "Launch Package",
+            "name": "Convert Remote",
             "type": "go",
             "request": "launch",
             "mode": "auto",
-            "program": "${fileDirname}",
+            "program": "${workspaceFolder}",
             "args": ["https://example-files.online-convert.com/document/txt/example.txt"],
-        }
+        },
+        {
+            "name": "Convert Chinese",
+            "type": "go",
+            "request": "launch",
+            "mode": "auto",
+            "program": "${workspaceFolder}",
+            "args": ["--model=zh_CN-huayan-medium.onnx", "lib/test_chinese.txt", "--speak-utf-8"]
+        },
     ]
 }

.vscode/settings.json

@@ -1,5 +1,7 @@
 {
     "editor.formatOnType": true,
-    // hide git ignored files
-    "explorer.excludeGitIgnore": false
+    "explorer.excludeGitIgnore": false,
+    // disable caching for go tests; don't run in parallel since we need to access
+    // files on the local filesystem
+    "go.testFlags": ["-count=1", "-p", "1"] 
 }

Dockerfile

@@ -1,19 +1,18 @@
-## This dockerfile is primarily for testing. It can be used like the following:
+# This dockerfile can be used to build a binary for use with the QuickPiperAudiobook command. 
+# You can use it for testing, or other architectures that don't have a piper build.
 # docker build -t quickpiperaudiobook .
 # docker run quickpiperaudiobook /app/examples/lorem_ipsum.txt
 
-FROM --platform=linux/amd64 golang:latest as build
+FROM --platform=linux/amd64 golang:1.22 as build
 
 WORKDIR /app
 
-# Copy all the code from the current directory
 COPY . .
 
 # Install Go dependencies and build the binary
 RUN go mod tidy && \
     go build -o QuickPiperAudiobook .
 
-# Final stage
 FROM --platform=linux/amd64 ubuntu:latest
 
 # Install runtime dependencies

cli/cli.go

@@ -14,11 +14,11 @@ import (
 )
 
 type CLI struct {
-	Input           string `arg:"" help:"Local path or URL to the input file"`
-	Output          string `help:"Directory in which to save the converted ebook file"`
-	Model           string `help:"Local path to the onnx model for piper to use"`
-	SpeakDiacritics bool   `help:"Speak diacritics from the input file"`
-	ListModels      bool   `help:"List available models"`
+	Input      string `arg:"" help:"Local path or URL to the input file"`
+	Output     string `help:"The directory in which to save the output audiobook"`
+	Model      string `help:"Local path to the onnx model for piper to use"`
+	SpeakUTF8  bool   `help:"Speak UTF-8 characters; Necessary for many non-English languages."`
+	ListModels bool   `help:"List piper models which are installed locally"`
 }
 
 // package level variables we want to expose for testing
@@ -32,15 +32,15 @@ const defaultModel = "en_US-hfc_male-medium.onnx"
 func RunCLI() {
 
 	if homedir_err != nil {
-		fmt.Printf("Error getting user home directory: %v\n", homedir_err)
-		return
+		fmt.Fprintf(os.Stderr, "Error getting user home directory: %v\n", homedir_err)
+		os.Exit(1)
 	}
 
 	var config CLI
 
 	if err := lib.CreateConfigIfNotExists(configFile, configDir, defaultModel); err != nil {
-		fmt.Printf("Error: %v\n", err)
-		return
+		fmt.Fprintf(os.Stderr, "Error creating default config file: %v\n", err)
+		os.Exit(1)
 	}
 
 	parser, _ := kong.New(&config, kong.Configuration(kongyaml.Loader, configFile))
@@ -49,8 +49,8 @@ func RunCLI() {
 		_, err := parser.Parse([]string{name})
 
 		if err != nil {
-			fmt.Println("Error parsing the value for", name, "in your config file at:", configFile)
-			return
+			fmt.Fprintf(os.Stderr, "Error parsing the value for %s in your config file at: %s\n", name, configFile)
+			os.Exit(1)
 		}
 	}
 
@@ -60,9 +60,7 @@ func RunCLI() {
 	if cli.ListModels {
 		models, err := lib.FindModels(configDir)
 		if err != nil {
-			fmt.Printf("Error: %v\n", err)
 			ctx.FatalIfErrorf(err)
-			return
 		}
 
 		if len(models) == 0 {
@@ -74,7 +72,7 @@ func RunCLI() {
 	}
 
 	if cli.Output == "" && config.Output != "" {
-		fmt.Println("No output value specified, default from config file: " + config.Output)
+		fmt.Println("No output directory specified, default from config file: " + config.Output)
 		cli.Output = config.Output
 		// if output is not set and config is not set, default to current directory
 	} else if cli.Output == "" && config.Output == "" {
@@ -98,9 +96,7 @@ func RunCLI() {
 	if _, err := os.Stat(cli.Output); os.IsNotExist(err) {
 		err := os.MkdirAll(cli.Output, os.ModePerm)
 		if err != nil {
-			fmt.Printf("Error: %v\n", err)
 			ctx.FatalIfErrorf(err)
-			return
 		}
 	}
 
@@ -109,52 +105,54 @@ func RunCLI() {
 		if err := lib.CheckEbookConvertInstalled(); err != nil {
 			fmt.Printf("Error: %v\n", err)
 			ctx.FatalIfErrorf(err)
-			return
 		}
 	}
 
 	if !lib.PiperIsInstalled(configDir) {
 		if err := lib.InstallPiper(configDir); err != nil {
 			ctx.FatalIfErrorf(err)
-			return
 		}
 	} else {
 		slog.Debug("Piper install detected in " + configDir)
 	}
 
-	modelPath, err := lib.ExpandModelPath(cli.Model, configDir)
+	modelPath, modelPathErr := lib.ExpandModelPath(cli.Model, configDir)
 
-	if err != nil {
+	// Some errors above are fine; (we can just download the corresponding model)
+	// but others that pertain to having the model but not the corresponding metadata are
+	// an error that should be fatal
+	if modelPathErr != nil && strings.Contains(modelPathErr.Error(), "but the corresponding") {
+		ctx.FatalIfErrorf(modelPathErr)
+	}
+
+	if modelPathErr != nil {
 		// if the path can't be expanded, it doesn't exist and we need to download it
 		err := lib.DownloadModelIfNotExists(cli.Model, configDir)
+
+		if err != nil && modelPathErr != nil {
+			fmt.Printf("Error: %v\n", modelPathErr)
+		}
+
 		if err != nil {
-			fmt.Printf("Error: %v\n", err)
 			ctx.FatalIfErrorf(err)
-			return
 		}
 		modelPath, err = lib.ExpandModelPath(cli.Model, configDir)
-
 		if err != nil {
-			fmt.Printf("Error could not find the model path after downloading it: %v\n", err)
-			ctx.FatalIfErrorf(err)
-			return
+			ctx.FatalIfErrorf(fmt.Errorf("error could not find the model path after downloading it: %v", err))
 		}
 	}
 
 	data, err := lib.GetConvertedRawText(cli.Input)
 
 	if err != nil {
-		fmt.Printf("Error: %v\n", err)
 		ctx.FatalIfErrorf(err)
-	} else {
-		fmt.Println("Text conversion completed successfully.")
+	} else if data == nil {
+		ctx.FatalIfErrorf(fmt.Errorf("after converting %s to txt, no data was generated", cli.Input))
 	}
 
-	if !cli.SpeakDiacritics {
+	if !cli.SpeakUTF8 {
 		if data, err = lib.RemoveDiacritics(data); err != nil {
-			fmt.Printf("Error: %v\n", err)
 			ctx.FatalIfErrorf(err)
-			return
 		}
 
 	}

cli/e2e_test.go

@@ -2,6 +2,7 @@ package cli
 
 import (
 	"os"
+	"path/filepath"
 	"testing"
 )
 
@@ -12,3 +13,43 @@ func TestCLI(t *testing.T) {
 	os.Args = append(os.Args, "https://example-files.online-convert.com/document/txt/example.txt")
 	RunCLI()
 }
+
+func TestCLIWithDiacritics(t *testing.T) {
+	// reset all cli args, since the golang testing framework changes them
+	os.RemoveAll(configDir)
+	origArgs := os.Args
+	os.Args = append(origArgs[:1], "https://example-files.online-convert.com/document/txt/example.txt", "--speak-utf-8")
+	RunCLI()
+
+	// make sure that after running you can run the list models command and it will work
+	os.Args = append(origArgs[:1], "https://example-files.online-convert.com/document/txt/example.txt", "--list-models")
+	RunCLI()
+}
+
+// Test that the cli works with chinese language text
+func TestChinese(t *testing.T) {
+	// reset all cli args, since the golang testing framework changes them
+	os.RemoveAll(configDir)
+	origArgs := os.Args
+	// get the file located at ../lib/test_chinese.txt
+	os.Args = append(origArgs[:1], "../lib/test_chinese.txt", "--model=zh_CN-huayan-medium.onnx", "--speak-utf-8")
+	RunCLI()
+
+	// check if there is a file at ~/Audiobooks/test_chinese.wav and make sure it's not empty
+	homedir, err := os.UserHomeDir()
+	if err != nil {
+		t.Fatalf("error getting user home directory: %v", err)
+	}
+	testFile := filepath.Join(homedir, "Audiobooks", "test_chinese.wav")
+	defer os.Remove(testFile)
+
+	if info, err := os.Stat(testFile); err != nil {
+		if os.IsNotExist(err) {
+			t.Fatalf("file not created: %v", err)
+		}
+		t.Fatalf("error getting file info: %v", err)
+	} else if info.Size() == 0 {
+		t.Fatalf("file is empty")
+	}
+
+}

lib/models.go

@@ -15,6 +15,10 @@ var ModelToURL = map[string]string{
 	"en_US-hfc_female-medium.onnx":            "https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/hfc_female/medium/en_US-hfc_female-medium.onnx",
 	"en_US-lessac-medium.onnx":                "https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx",
 	"en_GB-northern_english_male-medium.onnx": "https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_GB/northern_english_male/medium/en_GB-northern_english_male-medium.onnx",
+	// Below is an example of a non-English model
+	// I happily accept PRs for others here. It is just a bit tedious to enumerate them all
+	// since some do not follow the same pattern.
+	"zh_CN-huayan-medium.onnx": "https://huggingface.co/rhasspy/piper-voices/resolve/main/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx",
 }
 
 func ExpandModelPath(modelName string, defaultModelDir string) (string, error) {
@@ -25,14 +29,15 @@ func ExpandModelPath(modelName string, defaultModelDir string) (string, error) {
 		if _, err := os.Stat(modelName + ".json"); err == nil {
 			return modelName, nil
 		}
-		return "", fmt.Errorf("onnx for model: %s was found but the corresponding onnx.json was not", modelName)
+		return "", fmt.Errorf("onnx for model '%s' was found but the corresponding onnx.json was not", modelName)
 	}
+
 	if _, err := os.Stat(filepath.Join(defaultModelDir, modelName)); err == nil {
 		if _, err := os.Stat(filepath.Join(defaultModelDir, modelName) + ".json"); err == nil {
 			return filepath.Join(defaultModelDir, modelName), nil
 		}
-		return "", fmt.Errorf("onnx for model: %s was found in the model directory: %s but the corresponding onnx.json was not", modelName, defaultModelDir)
+		return "", fmt.Errorf("onnx for model '%s' was found in the model directory: '%s' but the corresponding onnx.json was not", modelName, defaultModelDir)
 
 	}
-	return "", fmt.Errorf("model not found: %s", modelName)
+	return "", fmt.Errorf("model: %s", modelName)
 }

lib/models_test.go

@@ -0,0 +1,51 @@
+package lib
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestExpandModelPath(t *testing.T) {
+	// Create a temporary directory for testing
+	tempDir := t.TempDir()
+	modelName := "test_model"
+	modelPath := filepath.Join(tempDir, modelName)
+	modelJSONPath := modelPath + ".json"
+
+	// Test case 1: Both ONNX and JSON files are present
+	os.WriteFile(modelPath, []byte("dummy ONNX model"), 0644)
+	os.WriteFile(modelJSONPath, []byte("dummy JSON"), 0644)
+
+	result, err := ExpandModelPath(modelName, tempDir)
+	if err != nil || result != modelPath {
+		t.Errorf("Expected %s, got %s, error: %v", modelPath, result, err)
+	}
+
+	// Test case 2: ONNX file is present, but JSON file is missing
+	os.Remove(modelJSONPath) // remove the JSON file
+
+	result, err = ExpandModelPath(modelName, tempDir)
+	if err == nil || result != "" {
+		t.Errorf("Expected error for missing JSON file, got: %v, result: %s", err, result)
+	}
+
+	// Test case 3: Model not found
+	result, err = ExpandModelPath("non_existent_model", tempDir)
+	if err == nil || result != "" {
+		t.Errorf("Expected error for non-existent model, got: %v, result: %s", err, result)
+	}
+
+	// Test case 4: Model found in the default model directory
+	modelNameInDir := "another_model"
+	modelPathInDir := filepath.Join(tempDir, modelNameInDir)
+	modelJSONPathInDir := modelPathInDir + ".json"
+
+	os.WriteFile(modelPathInDir, []byte("dummy ONNX model"), 0644)
+	os.WriteFile(modelJSONPathInDir, []byte("dummy JSON"), 0644)
+
+	result, err = ExpandModelPath(modelNameInDir, tempDir)
+	if err != nil || result != modelPathInDir {
+		t.Errorf("Expected %s, got %s, error: %v", modelPathInDir, result, err)
+	}
+}

lib/piper.go

@@ -16,8 +16,11 @@ import (
 	"github.com/gen2brain/beeep"
 )
 
-func RunPiper(filename string, // we need to have the filename here since the file passed in is a tmp file and a dummy name
-	model string, file *os.File, outdir string) error {
+func RunPiper(
+	filename string, // filename must be specified since the file passed in is a tmp file and a dummy name
+	model string, // piper model used
+	file *os.File, // file with text to convert
+	outdir string) error {
 
 	// Debugging: Read file content to check if it's empty
 	fileContent, err := io.ReadAll(file)

lib/test_chinese.txt

@@ -0,0 +1 @@
+彩虹，又稱天弓、天虹、絳等，簡稱虹，是氣象中的一種光學現象，當太陽 光照射到半空中的水滴，光線被折射及反射，在天空上形成拱形的七彩光譜，由外 圈至内圈呈紅、橙、黃、綠、蓝、靛蓝、堇紫七种颜色（霓虹則相反）。