Scientific papers converted to audio files so I can listen to papers while I cook / clean / exercise. It's going to skip out on math/diagrams/tables. Even if I could parse the math consistently, I couldn't understand it by just listening anyway. Hopefully the audio files sans math and diagrams are enough to get me the high level content so I know what to spend time on at the office.
- Populate Queue
- Collect PDFs saved to desktop/downloads/documents
- Collect PDFs saved via Google Keep
- Collect PDFs from Google Scholar Recommendations
- Transform PDFs into JSONs with ScienceParse
- Clean and Transform JSON text into SSML
- replace references with "See X by x` Y by y,..."
- strip out spurious stuff like arxiv + date on side of paper
- Spelling things out vs Saying them Phonetically
- Figure out what is best way to handle math / algorithms
- Transform SSML into MP3 with Google Text-to-Speech
- Tag MP3 with album and track info for easy organization in music apps
- Use pip to set up dependencies
pip install requirements.txt
- Set up your google text-to-speach API if you don't already have one
- Save the json to the google-key directory, make sure its the only thing in there
- run test.py, you should get an album of mp3 files of a the test paper
You have a single paper you want to process via
TODOTODO
Just download the paper and let the code do the rest. The process starts when your computer does. It watches a few dir, looks for new PDFs, renames them (since most file names for pdfs of a pre-print server or online journal are just a bunch of numbers and i can never be bothered to type in the title), makes the auidofiles, puts them in a dir that syncs with your mobile device. Then when you're making breakfast or going on a bike ride, the papers you downloaded with the intention of reading are now ready for auditory consumption.