The documentation for creating and editing Notifications Rules for matching application notifications and defining the test-to-speech output is a little hard to find in the app - so here it is for easier reference:
Notification Rules are used to define the desired text-to-speech behaviour for system and application notifications.
Several default rules are pre-configured. As new application notifications are discovered, the rules can be modified to change the rule matching criteria and text-to-speech output.
The text-to-speech pattern is read when a system or application notification is displayed. Data items from the notification can be used to construct the speech output using the substitution patterns described below
_APPNAME_- The application name.
_TICKER_- The ticker text displayed across the top of the screen when a notification occurs.
_TITLE_- The title shown on the notification panel.
_TEXT_- The second line of text shown on the notification panel. Only the first line of text is read to prevent excess text-to-speech output. If the full text output is require use
_ALLTEXT- The full text of the second line on the notification panel. Use with caution, some applications include a lot of output in the Text field, for example the entire contents of a email.
_SUBTEXT_- The third line of text shown on the notification panel.
_SUMMARY_- The text at the bottom of the notification panel.
_INFO_- The small text on the bottom right notification panel.
An example notification text-to-speech pattern can be constructed as:
Notification from _APP_NAME_. New message from _TEXT_. _TITLE_.
Some applications use multi-line notifications to consolidate multiple alerts into a single notification panel. When creating a text-to-speech pattern for a multiline rule the following additional substitutions can be used:
_LINES_- All the individual notification items.
_FIRSTLINE_- Just the first entry in the list of notification items.
_LASTLINE_- Just the last entry in the list of notification items.\<
Notification Matching Patterns
The notification rule used to determine the desired text-to-speech output is selected based on pattern matching against the notification details. The match patterns may can be selected from some common selections like 'Begins with', 'Contains' or on match is the field is blank etc. More complex matching rules can be defined using the 'Match Pattern' section allows for Regular Expression syntax (regex) pattern matching.
- ID - an application specific Id assigned to the notification. This is usually fixed, but some applications generate a new value for each notification. If you find multiple new rules are generated for the same notification try removing this value from the rule pattern.
- TAG - an application specific Tag assigned to the notification. This is usually fixed, but some applications generate a new value for each notification.
- TICKER - the ticker text displayed for the notification.
- TITLE - the title text displayed for the notification.
- TEXT - the second line of text displayed for the notification.
- SUBTEXT - the third line of text displayed for the notification.
- SUMMARY - the summary text displayed for the notification.
- INFO - the info text displayed for the notification.
Adjusting the RULE PRIORITY can be used to ensure more specific rules are matched in priority over generic matching patterns. If multiple rules match the same notification then the rule with the lowest priority value will be used. The default Rule priority for new rules is 100