Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Natural Language AIML amendments:
- When an NLTK responses isn't meaningful enough (response_weight) enough with existing corpa, AIML will provide a response
- Create PATTERN/TEMPLATE (+catch all) pairs in AIML that will render nltk responses from the grammar engine of good weight
- The AIML XML must be able to specify calls to python nltk api
- hubot internals + AIML -> python nltk Will hook into the appropriate nltk api call just before rendering. AIML interpreters can provide pre processing functions to expand abbreviations, remove misspellings, etc. in our case call nltk api functions:
<category> <pattern>WHAT IS YOUR NAME</pattern> <template><![CDATA[My name is <bot action="grammar.compose(question = 'What is your name')"/> ]]></template> </category> <category> <pattern>WHAT ARE YOU CALLED</pattern> <!-- Different ways of asking --> <template> <srai>what is your name</srai> </template> </category> <category> <pattern>WHAT DID JOHN DO</pattern> <template><![CDATA[<bot action="grammar.compose(question = 'What did john do')"/> ]]></template> </category>
must be able to interpret: "yooo codi, set up my symfony in the spaz room (where spaz room tag is known by codi from previous chats to be in /home/charl/spaz)"
Codi's response (after some pebble consumption): found symfony2, downloading, extracting, read docs, check config you have a symfony application deployed in /home/charl/spaz check documentation on how to run your first app ... opens documentation or points you to it.
Create Debugger Assistant Pebble:
as part of an interface to octoplasm (do action)
octoplasm mvc style dispatcher to debugger assistant actions
use case: during app development you get a standard app domain specific error that codi keeps track of using salsa. Codi either prompts or you switch to it asking "Whats wrong with my symfony app in spaz?" Codi prints the computed structured error response using contextual lexicon from a knowledge base entry. e.g. Slight miss configuration, your logs indicated permissions for cache error This is now corrected as per FilesystemKB #12388
Codi then makes use of the debugger_assistant pebble to try and correct it for you.