Extract audio track, convert to CAF/Opus, and generate a RSS2 feed for personal hosting podcast!

convert_mp4_to_caf_opus.sh

#!/bin/bash

ffmpeg -y -i "$f" -vn -c:a libfdk_aac -b:a 128k "audio/$(basename $f .mp4).m4a"
afconvert -f caff -d opus -b 8000 "audio/$(basename $f .mp4).m4a" "opus/$(basename $f .mp4).caf"

gen_podcast_feed.py

#!/usr/bin/env python

# In this code snippet, what I'm going to do is
# 1. extract the audio track from .mp4 files and convert it into .m4a with 128kb AAC.
# 2. convert to .caf (Core Audio Format) with 8kb Opus.
# 3. generate a RSS2 feed for self-hosting podcast.

# built-in modules
import datetime
import json
import pathlib
from pathlib import Path
import re
import sys
import uuid

# 3rd-party modules
import plumbum
import arrow
import rfeed
from rfeed import *


def get_items(path = None):
    from plumbum.cmd import afconvert, ffprobe, ffmpeg

    items = []
    if path is None:
        path = '.'
    root = Path(path)

    prog = re.compile('.*(\d{3})\..*title_(.+)_speaker_(.+)\..+$')

    for file_ in root.glob('*.mp4'):
        m = prog.match(str(file_))
        if m is None:
            continue
        #print(m.group(0), '--', m.group(1), m.group(2), m.group(3))
        number = m.group(1)
        title = m.group(2)
        author = m.group(3)

        file_m4a = root.cwd().joinpath('audio', file_.stem + '.m4a')
        # extract audio channel of 128k into .m4a located in /audio/ folder
        if not file_m4a.exists():
            ffmpeg(['-y', '-i', file_, '-vn', '-c:a', 'libfdk_aac', '-b:a', '128k', file_m4a])

        file_caf = root.cwd().joinpath('podcast', number + '.caf')
        # convert to 8k .caf with opus codec located in /podcast/, with simplified filename
        if not file_caf.exists():
            afconvert(['-f', 'caff', '-d', 'opus', '-b', '8000', file_m4a, file_caf])

        # get the duration by ffprobe
        stdout = ffprobe(['-v', 'quiet', '-of', 'json', '-hide_banner', '-show_format', '-i', file_caf])
        json_data = json.loads(stdout)
        duration = datetime.timedelta(0, float(json_data['format']['duration']))

        itunes_item = iTunesItem(
                author = author,
                #image = "",
                duration = duration,
                explicit = "clean",
                subtitle = title,
                order = int(number),
                #summary = ""
                )

        item = Item(
                title = number + '. ' + title + ' | ' + author,
                #link = "http://www.example.com/articles/1",
                #description = "",
                author = author,
                guid = Guid(uuid.uuid3(uuid.NAMESPACE_DNS, str(file_))),
                pubDate = arrow.Arrow.fromtimestamp(file_.stat().st_mtime),
                enclosure = Enclosure(url="https://oo.xx.asdf/"+file_caf.name, length=file_caf.stat().st_size, type='audio/aac'),
                extensions = [itunes_item])

        items.append(item)

    return sorted(items, key=lambda i: i.extensions[0].order, reverse=True)


def main(argv = sys.argv[:]):
    items = get_items()

    itunes = iTunes(
            author = 'John Doe',
            subtitle = 'Some title',
            summary = 'Some summary',
            image = 'https://oo.xx.asdf/cover.jpg',
            explicit = 'clean',
            categories = iTunesCategory(name = 'Education', subcategory = 'Higher Education'),
            owner = iTunesOwner(name = 'John Doe', email = '[email protected]'))

    feed = Feed(
            title = 'Some title',
            link = 'https://oo.xx.asdf',
            description = 'Some description',
            generator = 'rfeed - An extensible RSS 2.0 Generator written in Python',
            docs = 'http://blogs.law.harvard.edu/tech/rss',
            language = 'en-US',
            pubDate = arrow.now(),
            lastBuildDate = arrow.now(),
            items = items,
            extensions = [itunes])

    print(feed.rss())
    return 0


if __name__ == '__main__':
    sys.exit(main())

rfeed_with_itunes.py

# refer: https://github.com/egorsmkv/rfeed

import datetime 
from rfeed import *

 itunes_item = iTunesItem(
    author = "Santiago L. Valdarrama",
    image = "http://www.example.com/artwork.jpg",
    duration = "01:11:02",
    explicit = "clean",
    subtitle = "The subtitle of the podcast episode",
    summary = "Here is the summary of this specific episode")

item = Item(
    title = "Sample article",
    link = "http://www.example.com/articles/1", 
    description = "This is the description of the first article",
    author = "Santiago L. Valdarrama",
    guid = Guid("http://www.example.com/articles/1"),
    pubDate = datetime.datetime(2014, 12, 29, 10, 00),
    extensions = [itunes_item])

itunes = iTunes(
    author = "Santiago L. Valdarrama",
    subtitle = "A sample podcast that will never be produced",
    summary = "This is just a fake description",
    image = "http://www.example.com/artwork.jpg",
    explicit = "clean",
    categories = iTunesCategory(name = 'Technology', subcategory = 'Software How-To'),
    owner = iTunesOwner(name = 'Santiago L. Valdarrama', email = '[email protected]'))

feed = Feed(
    title = "Sample Podcast RSS Feed",
    link = "http://www.example.com/rss"
    description = "An example of how to generate an RSS 2.0 feed",
    language = "en-US",
    lastBuildDate = datetime.datetime.now(),
    items = [item],
    extensions = [itunes])

print feed.rss()

rsync.sh

#!/bin/bash

rsync -r -a -v -e ssh podcast/* oo.xx.asdf:public-html/

sample_podcast_feed_by_feeder.xml

<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0">
    <channel>
        <title>讀書 podcast</title>
        <link>https://oo.xx.asdf/</link>
        <description>大家一起來讀書</description>
        <generator>Feeder 3 3.5.8(3057); Mac OS X Version 10.13.3 (Build 17D47) http://reinventedsoftware.com/feeder/</generator>
        <docs>http://blogs.law.harvard.edu/tech/rss</docs>
        <language>en</language>
        <pubDate>Mon, 19 Feb 2018 11:22:18 +0800</pubDate>
        <lastBuildDate>Mon, 19 Feb 2018 11:22:18 +0800</lastBuildDate>
        <itunes:image href="https://oo.xx.asdf/cover.jpg"/>
        <itunes:block>no</itunes:block>
        <itunes:category text="Education">
            <itunes:category text="Higher Education"/>
        </itunes:category>
        <itunes:type>episodic</itunes:type>
        <item>
            <title>006.《書本六》</title>
            <description><![CDATA[desc blah blah]]></description>
            <pubDate>Mon, 19 Feb 2018 11:22:18 +0800</pubDate>
            <enclosure url="https://oo.xx.asdf/006.m4a" length="1192902" type="audio/aac"/>
            <guid isPermaLink="false">99B07627-1111-2222-3333-DBC6FCEE9372</guid>
            <itunes:author>講者三</itunes:author>
            <itunes:duration>19:52</itunes:duration>
            <itunes:season>0</itunes:season>
            <itunes:episode>6</itunes:episode>
            <itunes:episodeType>full</itunes:episodeType>
            <content:encoded><![CDATA[note blah blah]]></content:encoded>
        </item>
    </channel>
</rss>