Collect news in a single place for Reddit subs and RSS feeds. The Text-to-Speech (TTS) helps you listen to your news. The TTS management settings help you configure the TTS sounds available on your device. Additionally, you can set pitch, speed and interval configuration.
The script fetches the destinations using php-curl, processes the content, and saves them in json files in the temp directory. On the frontend, VueJS preares the virtual DOM and uses the json files to list the items listed in the json file. The TTS uses the voices available on your device. If your device has a natural voice engine, it is a much more pleasant experience.
The feed.php file is the processing hub of this project. It contains the sub reddits and feeds that are to be fetched and processed. A cron job is set to run in an interval to retrive the data and process it into json files in the \temp folder.
The core functions are present in functions.php. For setting up, you do not need to edit the functions.php file. By default, the top list of the subreddit is requested.
The $subs variable is an array that contains the name of the subreddits to be requested.
For subreddits, the name of the subreddit is used in the parsed json file.
E.g. For worldnews subreddit, the file would be saved as worldnews-parsed.json.
For parsing feeds, there is some rudimentary name extraction process to use the domain name for the parsed file name. It might not work with sub-domains and complicated URLs. It is easier if you a specific name is provided using the parsefeed(feed_link, name) syntax. Which takes two arguments, the RSS feed link and the name prepended to the parsed json file.
E.g.
parseFeed('http://feeds.bbci.co.uk/news/technology/rss.xml', 'bbctech');
For the above code, it will fetch the RSS URL and the parsed file will be saved as bbctech-parsed.json in the temp folder
Note: The sparser.php is a simple html parser to parse HTML into json made for sites that do not provide RSS feeds.
An simple URL query string parameter k is used to prevent unauthorized requests for processing the feeds
To use it in production, it is important that you have SWEN_KEY environment variable set. For local testing, if the environement variable is not set or is empty, the update will run with with just /feed.php?k. If more secure method can be used, please raise an issue.
With SWEN_KEY set to run123
https://domain.com/feed.php?k=run123
Without any environment variable set
https://domain.com/feed.php?k
On the index.html, first level of tabs are represented in <chitab title=News :selected=true>. The title is what get displayed in the frontend. Second level tabs are represented in <tab name="npr" title="NPR"></tab> format. The name attribute of the tab represents the json file to call.
For tab with the name npr, it will call the npr-parsed.json file and process it when that tab is clicked.
The title is what's displayed on the tab display.
You can run cron jobs on your own linux system with crontab -e or use external services.
php <script-location>/feed.php?k=run123 > dev/null 2>&1
- Setting a global timer on from line 31 to line 41 and 193
Common Issues (PC/Mobile)
- Cannot select default TTS voice on first load on some devices
Common Mobile Issues
-
When a tab has a lot of text, the TTS fails to initiate and speak. (solution: reduce the number of items to be read from the gear icon, like 1 to 15)
-
Text to Speech FAB doesn't disappear on scroll to bottom
-
Older devices, if capable of web speech synthesis, take time to pause and stop
Firefox Mobile (Android)
- Cannot pause (related to speechsynthesis API not being completely supported)
Laracasts youtube video on VueJS and Tabs