Command Line Artificial Intelligence or CLAI is an open-sourced project from IBM Research aimed to bring the power of AI to the command line interface.
Related pull-requests: #63 (this PR supersedes #63)
:tophat: What is the goal?
The tellina skills provides our own deployed version of the Tellina service, which CLAI calls. This is intended as a broader-domain expansion of the nlc2cmd skill that works with a much wider variety of Bash utilities and flags.
:memo: How is it being implemented?
The tellina skill is implemented by calling a POST server that exposes a copy of the Tellina service running on an IBM YellowZone machine. The skill passes the natural language invocation from the CLI to the service; which returns the top translation and a 0.0 confidence value. The skill then passes that translation back to the CLI.
:tv: Screenshot or gif showing the result.
:boom: How can it be tested?
Currently only functional testing can be performed, since the tellina skill returns a default confidence value of 0.0. After installing CLAI and installing/activating the new (tellina) skill, there are two possible cases:
The skill must be explicitly invoked, e.g. as below:
>> clai tellina <natural language invocation>
The skill must be the only one active (this can be achieved by deactivating all other skills), and then it can be invoked as shown in the gif above, as:
:pushpin: References
:tophat: What is the goal?
The
tellinaskills provides our own deployed version of theTellinaservice, whichCLAIcalls. This is intended as a broader-domain expansion of thenlc2cmdskill that works with a much wider variety of Bash utilities and flags.:memo: How is it being implemented?
The
tellinaskill is implemented by calling aPOSTserver that exposes a copy of theTellinaservice running on an IBM YellowZone machine. The skill passes the natural language invocation from the CLI to the service; which returns the top translation and a0.0confidence value. The skill then passes that translation back to the CLI.:tv: Screenshot or gif showing the result.
:boom: How can it be tested?
Currently only functional testing can be performed, since the
tellinaskill returns a default confidence value of0.0. After installingCLAIand installing/activating the new (tellina) skill, there are two possible cases:The skill must be explicitly invoked, e.g. as below:
>> clai tellina <natural language invocation>The skill must be the only one active (this can be achieved by deactivating all other skills), and then it can be invoked as shown in the gif above, as:
>> clai <natural language invocation>