Merge pull request #9 from traugdor/dev

potential fix for  #8

README.md

@@ -1,6 +1,6 @@
 # ComfyUI-Riffusion
 
-A ComfyUI extension for Riffusion audio generation.
+A ComfyUI extension for Riffusion audio generation. Requires installation of 3rd-party software FFmpeg.
 
 ## Installation
 
@@ -9,7 +9,56 @@ A ComfyUI extension for Riffusion audio generation.
 OR
 Install via Comfy Registry: https://registry.comfy.org/publishers/traugdor/nodes/riffusion
 
-2. Restart ComfyUI
+2. Download and install FFMPEG by following these instructions:
+
+### Windows users
+
+1. **Download FFmpeg:**
+    * Go to the [official FFmpeg download page](https://ffmpeg.org/download.html).
+    * Click on the "Windows" logo and select a build (usually, the "gpl" builds are recommended). You may find builds hosted by third-party sites like Gyan.dev or BtbN.
+2. **Extract the Files:**
+    * Once downloaded, extract the contents of the ZIP file to a location on your computer (e.g., C:\ffmpeg).
+3. **Add FFmpeg to System PATH:**
+    * Right-click on "This PC" or "Computer" on your desktop or in File Explorer and select "Properties."
+    * Click on "Advanced system settings" on the left side.
+    * In the System Properties window, click on the "Environment Variables" button.
+    * In the Environment Variables window, find and select the "Path" variable in the "System variables" section, then click "Edit."
+    * Click "New" and add the path to the bin directory of the extracted FFmpeg folder (e.g., C:\ffmpeg\bin).
+    * Click "OK" to close all dialog boxes.
+4. **Verify Installation:**
+    * Open Command Prompt (press Win + R, type cmd, and hit Enter).
+    * Type the following command and press Enter:
+    ```bash
+    ffmpeg -version
+    ```
+    * If FFmpeg is correctly installed, you will see version information and other details. You may need to close and reopen your Command Prompt/Powershell/Terminal window to get the latest environment variables.
+5. **Restart ComfyUI**
+
+### Linux users
+
+1. **Open Terminal**
+2. **Install FFmpeg** using your package manager
+    * Debian/Ubuntu-based:
+        ```bash
+        sudo apt update
+        sudo apt install ffmpeg
+        ```
+    * Fedora:
+        ```bash
+        sudo dnf install ffmpeg 
+        ```
+    * Arch:
+        ```bash
+        sudo pacman -S ffmpeg
+        ```
+3. **Verify installation**
+    * After installation is complete, you can verify that FFmpeg is installed correctly by running:
+        ```bash
+        ffmpeg -version
+        ```
+    * If FFmpeg is installed, you will see version information and other details about the installation.
+
+4. **Restart ComfyUI**
 
 ## Usage
 
@@ -18,4 +67,4 @@ The extension adds a new node "Riffusion" under the audio category.
 ## TODO
 - Add support for multiple images (as input) to generate a single audio file for longer audio lengths
 - Add support for multiple images to generate multiple audio files
-- Add support for other audio types (wav, mp3, ogg, etc.)
+- Add FFmpeg installer for Windows/Linux? Current method is to warn user and disable audio formats that require FFmpeg

__init__.py

@@ -2,6 +2,8 @@ import sys
 import subprocess
 from pathlib import Path
 
+FFMPEG_IS_NOT_INSTALLED = False
+
 ### GET PACKAGES ###
 def packages(versions=False):
     try:
@@ -32,7 +34,26 @@ required_packages = read_requirements()
 for package in required_packages:
     if package.lower() not in [pkg.lower() for pkg in installed_packages]:
         install_package(package)
-print("All packages needed by Riffusion are installed.")
+print("All Python packages needed by Riffusion are installed.")
+
+### Ensure FFmpeg is installed ###
+# FFmpeg is a 3rd-party program that is needed for Riffusion to work and it's not a Python library.
+# determine if system is windows or linux??? MacOS is not really supported for Stable Diffusion
+try:
+    result = subprocess.check_output(['ffmpeg', '-version'], stderr=subprocess.STDOUT)
+    print("FFmpeg is installed.")
+except subprocess.CalledProcessError as e:
+    print("[Riffusion WARNING] An error occurred while checking FFmpeg installation:", e.output.decode())
+    print("[Riffusion WARNING] All other audio formats will be disabled until until the previous error is resolved.")
+    FFMPEG_IS_NOT_INSTALLED = True
+except FileNotFoundError: #Windows gives a FileNotFoundError
+    print("[Riffusion WARNING] FFmpeg is not installed.")
+    print("[Riffusion WARNING] Please install FFmpeg from https://ffmpeg.org/download.html. All other audio formats will be disabled until FFmpeg is installed.")
+    # set flag to signal FFMpeg is not installed
+    FFMPEG_IS_NOT_INSTALLED = True
+
+# TODO: Add FFmpeg installer for Windows
+# TODO: Add FFmpeg installer for Linux
 
 ### Don't import nodes until all packages are installed ###
 from .nodes import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS

nodes.py

@@ -23,7 +23,7 @@ class RiffusionNode:
         return {
             "required": {
                 'spectrogram': ("IMAGE",),
-                'filename_prefix': ("STRING", {"default": "Riffusion"}),
+                'filename_prefix': ("STRING", {"default": "audio\Riffusion"}),
 
             },
         }
@@ -132,7 +132,7 @@ class RiffusionNode:
         else:
             max_number = 0
             for file in list_of_files:
-                number = int(file.split("_")[-1].split(".")[0]) 
+                number = int(file.split("_")[-2]) 
                 max_number = max(max_number, number)
             # Increment the number and add it to the filename and add the extension
             new_number = max_number + 1
@@ -155,7 +155,7 @@ class RiffusionNode:
         print(f"Duration of {new_filename}: {duration}")
 
         # get full output path
-        full_output_path = os.path.join(self.output_dir, new_filename)
+        full_output_path = os.path.join(output_dir, new_filename)
 
         # save the audio
         with open(full_output_path, 'wb') as f: