fuzzy.ai

Interface to the fuzzy.ai API for machine intelligence.

License

Copyright 2014-2018 Fuzzy.ai node@fuzzy.ai

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Overview

var FuzzyAIClient = require('fuzzy.ai');

var apiKey = 'API key from fuzzy.ai';

var client = new FuzzyAIClient({key: apiKey});

var agentID = 'ID from fuzzy.ai';

var inputs = {temperature: 87};

client.evaluate(agentID, inputs, function(err, outputs) {
  if (err) {
    console.error(err);
  } else {
    console.log("Fan speed is " + outputs.fanSpeed);
  }
});

FuzzyAIClient

This is the main class; it's what's returned from the require().

• FuzzyAIClient(options) Takes an options argument. This is an object with the following properties:

  • key: API key to use. You have to get an apiKey from https://fuzzy.ai/ . Keep this secret, by the way.

  • root is the root of the API server. It has the correct default 'https://api.fuzzy.ai' but if you're doing some testing with a mock, it can be useful.

  • queueLength is the length of the request queue; requests are parallelized and a maximum of this number of requests will be done concurrently. This has a reasonable default of 32.

  • maxWait is the maximum time in seconds to wait for a response before returning an error. This has a reasonable default of 10 seconds.

  • timeout is how long persistent connections will stick around. The default is 0, meaning no persistent connections. If you are going to do a lot of requests, set this to something reasonable, like 1000 or 5000. Infinity means never disconnect (although the server will, eventually). Note that if you use persistent connections, you'll want to use stop() at the end of your program so the connections are cleaned up.

This is the main method you need to use:

• evaluate(agentID, inputs, [meta], callback) Does a single inference. agentID is on the main page for the agent on http://fuzzy.ai/ . The inputs is an object, mapping input names to numeric values. meta is a string or boolean value; if provided and truthy, output will include a meta property with meta information, or a property with the same name as the string value, if you need to avoid using "meta" as a property.

  callback is a function with the signature function(err, outputs), where outputs is an object mapping output names to numeric values. It may also have a meta or other specified property depending on the meta flag.

To train for better results, use the feedback method with your success metric.

• feedback(evaluationID, metrics, callback) Trains the agent based on the evaluationID is returned in the output for the evaluate() method (see above). The success is an object, mapping success metric names to numeric values. Defining the success metric is up to you, but closer to zero = better. callback is a function with the signature function(err, metrics), where metrics is an object mapping output names to numeric values, which is just what you passed in plus some extra housekeeping data.

These might be useful but you normally don't need to mess with them.

• getAgents(callback) userID is the user ID, not the API key. callback is a function with the signature function(err, agents), where agents is an array of objects with id and name properties, one for each agent the user has.

• newAgent(agent, callback) userID is the user ID. agent is an agent object with at least properties inputs, outputs, rules. callback is a function with the signature function(err, agent) which returns the fully-realized agent with all its properties like timestamps and IDs.

• getAgent(agentID, callback) Gets a single agent by ID. callback is a function with the signature function(err, agent).

• putAgent(agentID, agent, callback) Updates an agent. callback has the signature function(err, agent) which will return the updated version.

• deleteAgent(agentID, callback) Deletes an agent. This method will permanently delete an agent and all related data; use with caution. callback has the signature function(err) which returns an error if there was a problem.

• evaluation(evaluationID, callback) Gets the audit data for the evaluation. evaluationID is returned in the output for the evaluate() method (see above). callback is a function with the signature function(err, audit), where audit is an object mapping audit data names to data. Important values here:

  • input: the input values that were passed in
  • crisp: the output values that were the results of the evaluation
  • rules: The index of the rules that fired

  Most of the rest is fuzzy logic stuff that you probably don't care about.

• getAgentVersion(versionID, callback) For older versions of an agent, you can get the properties of the older version.

• apiVersion(callback) Get the version data from the server. This is a good "smoke test" method to see if the server is responding at all. callback has the signature function(err, versionData), where versionData is an object with the following properties:

  • name: name of the API server. Usually api.
  • version: (Semver)[http://semver.org/] version for the API server software.
  • controllerVersion: Semver version for the fuzzy controller software.

• stop(callback) Clean up any persistent connections. You only need to call this if you provided a non-zero timeout to the constructor.