Project goal is to automate the creation of accessible full-text (sound and text synced) ebooks in epub3 and daisy format.
Aeneas is used to sync the audio and html (text) files automatically.
Aeneas has its own requirements that come with the aeneas-installer but can also be installed seperately using pip.
- eSpeak
- FFmpeg
- Python
- BeautifulSoup4
- lxml
- numpy
If issues arise with CEW then go into System Properties -> Environment Variables,
Add a new variable under System Variables named "AENEAS_WITH_CEW" and set it to "False"
Follow the installation guide found at https://github.com/readbeyond/aeneas/blob/master/wiki/INSTALL.md
Making sure to apt install all requirements such as ffmpeg, espeak, python, pip etc.
Follow the manual procedure 1.
Use the following commands:
- $ wget https://raw.githubusercontent.com/readbeyond/aeneas/master/install_dependencies.sh
- $ bash install_dependencies.sh
Try running "python -m aeneas.diagnostics" to see whether everything is installed correctly.
If issues arrive try also using the manual procedure 3.
Use the following commands:
- $ sudo pip install numpy
- $ sudo pip install aeneas
Rerun "python -m aeneas.diagnostics" to see whether everything is installed correctly.
Daisy Pipeline 2 has numerous scripts available to the user. Their GUI program is capable of running all these scripts though we are using the CLI to automate the process even further.
Before installing the Daisy Pipeline 2, install Java version 14
Daisy Pipeline 2 requires at least version 11 of Java, but there were some issues with using version 11 in early tests.
Follow the installation guide
Make sure the install package includes the CLI, also available as a seperate download.
If the Pipeline installer prompts you to install java 11 through them press "No".
Open up a command line and use the command
- "dp2.exe help" (windows)
- "dp2 help" (linux, add 'export PATH="$PATH:/opt/daisy-pipeline2-cli/"' to ~/.bashrc (remove ' ') to see if the webserver and everything else is working correctly.
If you wish to contribute please contact HBS directly at hbs@hbs.is
- Make sure to have nodejs and npm installed
- Run "npm install" to install all the required dependencies
- Start the server with "npm start"
- Server should now be running on http://localhost:5000
- Upload mp3 files, ncc.html and the book.html to the server
- Make sure that each mp3 file correspondes to a header (h1, h2, etc.) in the book file
- Wait for the script to finish creating the smil files
- Download zip file containing smil files
- Unzip the contents into a folder containing ncc.html, bookname.html, mp3 files and any other necessary files (/images, style.css)
- Open the ncc file in Hindenburg to standardize all the files
- Export epub/epub3/daisy from Hindenburg
- You should now have a working full-text book
- Upload the folder containing a valid daisy 2.02 ebook
- Validate the ebook if any errors occur
- Wait for the Daisy Pipeline 2 to process the book
- Download the epub file and open up in a ebook reader that supports Media Overlay (recommend Thorium Reader Alpha)
- Upload a folder containing nothing but valid daisy 2.02 ebooks
- Wait for the conversion to finish
- Download a zip file containing all books as epub 3 and or log files of failed books
- Some books may need to be fixed before re-running them, also converting them by themselves may be a better option
GNU General Public License v3.0