diff options
author | Vasil Zlatanov <v@skozl.com> | 2017-10-17 19:12:39 +0100 |
---|---|---|
committer | Vasil Zlatanov <v@skozl.com> | 2017-10-17 19:12:39 +0100 |
commit | 11201dc8dbb4a2fd47436ac755dbcfd0727573f2 (patch) | |
tree | 2f9ee0104bfd72d44e5ecbf91dd70c5bbf076dbc | |
parent | 4e4734e285a4b43d3fe98378071ad48734ba8ffa (diff) | |
download | speech-11201dc8dbb4a2fd47436ac755dbcfd0727573f2.tar.gz speech-11201dc8dbb4a2fd47436ac755dbcfd0727573f2.tar.bz2 speech-11201dc8dbb4a2fd47436ac755dbcfd0727573f2.zip |
add README
-rw-r--r-- | README | 44 |
1 files changed, 44 insertions, 0 deletions
@@ -0,0 +1,44 @@ +Info +==== +This project leverages the google speech to text api, which returns a +structure of results ranked by confidence. This JSON structure gets +passed to a procces.pl which has a long list of regular expressions. + +Questions like "How much battery do I have left." will match and evoke +either tts or other tools to perform the required action. This script is +very custom and will likely completely and should be adapter to the user +own system. The conditions and actions are easy to modify. + +If there none of the results returned in the JSON structure match the +regex, then the most likely input is passed to wolfram alpha api, the +output of which is returned to the text to speech and and the browser is +opened. + +How to use +========== +Executing run.sh will start sox which should record voice until there is +a pause of silence. If sox doesn't detect silence either microphone +volume is to loud or the constants passed to `rec' need to be increase +from 1% in run.sh + +Example's to ask: + +What is your name? +Who created you? +How are you? +Greetings my lovely computer! +Where is Canada? + + +Dependencies +============ +sox - for recording with microphone +festival - for text to speech +cpan JSON::MaybeXS - to parse JSON + +Optional +============ +cpan Forecast::IO - to get weather forecast +firefox - to open websites +xkvbd - to insert keystrokes +ratpoison - to manage the windows |