@datafire/google_prediction

Client library for Prediction

Installation and Usage

npm install --save @datafire/google_prediction

let google_prediction = require('@datafire/google_prediction').create({
  access_token: "",
  refresh_token: "",
  client_id: "",
  client_secret: "",
  redirect_uri: ""
});

.then(data => {
  console.log(data);
});

Description

Lets you access a cloud hosted machine learning service that makes it easy to build smart apps

Actions

oauthCallback

Exchange the code passed to your redirect URI for an access_token

google_prediction.oauthCallback({
  "code": ""
}, context)

Input

• input object
  • code required string

Output

• output object
  • access_token string
  • refresh_token string
  • token_type string
  • scope string
  • expiration string

oauthRefresh

Exchange a refresh_token for an access_token

google_prediction.oauthRefresh(null, context)

Input

This action has no parameters

Output

• output object
  • access_token string
  • refresh_token string
  • token_type string
  • scope string
  • expiration string

hostedmodels.predict

Submit input and request an output against a hosted model.

google_prediction.hostedmodels.predict({
  "hostedModelName": "",
  "project": ""
}, context)

Input

• input object
  • body Input
  • hostedModelName required string: The name of a hosted model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Output

trainedmodels.insert

Train a Prediction API model.

google_prediction.trainedmodels.insert({
  "project": ""
}, context)

Input

• input object
  • body Insert
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Insert2

trainedmodels.list

List available models.

google_prediction.trainedmodels.list({
  "project": ""
}, context)

Input

• input object
  • maxResults integer: Maximum number of results to return.
  • pageToken string: Pagination token.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output List

trainedmodels.delete

Delete a trained model.

google_prediction.trainedmodels.delete({
  "id": "",
  "project": ""
}, context)

Input

• input object
  • id required string: The unique name for the predictive model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

Output schema unknown

trainedmodels.get

Check training status of your model.

google_prediction.trainedmodels.get({
  "id": "",
  "project": ""
}, context)

Input

• input object
  • id required string: The unique name for the predictive model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Insert2

trainedmodels.update

Add new data to a trained model.

google_prediction.trainedmodels.update({
  "id": "",
  "project": ""
}, context)

Input

• input object
  • body Update
  • id required string: The unique name for the predictive model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Insert2

trainedmodels.analyze

Get analysis of the model and the data the model was trained on.

google_prediction.trainedmodels.analyze({
  "id": "",
  "project": ""
}, context)

Input

• input object
  • id required string: The unique name for the predictive model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Analyze

trainedmodels.predict

Submit model id and request a prediction.

google_prediction.trainedmodels.predict({
  "id": "",
  "project": ""
}, context)

Input

• input object
  • body Input
  • id required string: The unique name for the predictive model.
  • project required string: The project associated with the model.
  • alt string (values: json): Data format for the response.
  • fields string: Selector specifying which fields to include in a partial response.
  • key string: API key. Your API key identifies your project and provides you with API access, quota, and reports. Required unless you provide an OAuth 2.0 token.
  • oauth_token string: OAuth 2.0 token for the current user.
  • prettyPrint boolean: Returns response with indentations and line breaks.
  • quotaUser string: Available to use for quota purposes for server-side applications. Can be any arbitrary string assigned to a user, but should not exceed 40 characters. Overrides userIp if both are provided.
  • userIp string: IP address of the site where the request originates. Use this if you want to enforce per-user limits.

Output

• output Output

Definitions

Analyze

• Analyze object
  • dataDescription object: Description of the data the model was trained on.
    • features array: Description of the input features in the data set.
      • items object
        • categorical object: Description of the categorical values of this feature.
          • count string: Number of categorical values for this feature in the data.
          • values array: List of all the categories for this feature in the data set.
        • index string: The feature index.
        • numeric object: Description of the numeric values of this feature.
          • count string: Number of numeric values for this feature in the data set.
          • mean string: Mean of the numeric values of this feature in the data set.
          • variance string: Variance of the numeric values of this feature in the data set.
        • text object: Description of multiple-word text values of this feature.
          • count string: Number of multiple-word text values for this feature.
    • outputFeature object: Description of the output value or label.
      • numeric object: Description of the output values in the data set.
        • count string: Number of numeric output values in the data set.
        • mean string: Mean of the output values in the data set.
        • variance string: Variance of the output values in the data set.
      • text array: Description of the output labels in the data set.
        • items object
          • count string: Number of times the output label occurred in the data set.
          • value string: The output label.
  • errors array: List of errors with the data.
    • items object
  • id string: The unique name for the predictive model.
  • kind string: What kind of resource this is.
  • modelDescription object: Description of the model.
    • confusionMatrix object: An output confusion matrix. This shows an estimate for how this model will do in predictions. This is first indexed by the true class label. For each true class label, this provides a pair {predicted_label, count}, where count is the estimated number of times the model will predict the predicted label given the true label. Will not output if more then 100 classes (Categorical models only).
    • confusionMatrixRowTotals object: A list of the confusion matrix row totals.
    • modelinfo Insert2
  • selfLink string: A URL to re-request this resource.

Input

• Input object
  • input object: Input to the model for a prediction.
    • csvInstance array: A list of input features, these can be strings or doubles.

Insert

• Insert object
  • id string: The unique name for the predictive model.
  • modelType string: Type of predictive model (classification or regression).
  • sourceModel string: The Id of the model to be copied over.
  • storageDataLocation string: Google storage location of the training data file.
  • storagePMMLLocation string: Google storage location of the preprocessing pmml file.
  • storagePMMLModelLocation string: Google storage location of the pmml model file.
  • trainingInstances array: Instances to train model on.
    • items object
      • csvInstance array: The input features for this instance.
      • output string: The generic output value - could be regression or class label.
  • utility array: A class weighting function, which allows the importance weights for class labels to be specified (Categorical models only).
    • items object: Class label (string).

Insert2

• Insert2 object
  • created string: Insert time of the model (as a RFC 3339 timestamp).
  • id string: The unique name for the predictive model.
  • kind string: What kind of resource this is.
  • modelInfo object: Model metadata.
    • classWeightedAccuracy string: Estimated accuracy of model taking utility weights into account (Categorical models only).
    • classificationAccuracy string: A number between 0.0 and 1.0, where 1.0 is 100% accurate. This is an estimate, based on the amount and quality of the training data, of the estimated prediction accuracy. You can use this is a guide to decide whether the results are accurate enough for your needs. This estimate will be more reliable if your real input data is similar to your training data (Categorical models only).
    • meanSquaredError string: An estimated mean squared error. The can be used to measure the quality of the predicted model (Regression models only).
    • modelType string: Type of predictive model (CLASSIFICATION or REGRESSION).
    • numberInstances string: Number of valid data instances used in the trained model.
    • numberLabels string: Number of class labels in the trained model (Categorical models only).
  • modelType string: Type of predictive model (CLASSIFICATION or REGRESSION).
  • selfLink string: A URL to re-request this resource.
  • storageDataLocation string: Google storage location of the training data file.
  • storagePMMLLocation string: Google storage location of the preprocessing pmml file.
  • storagePMMLModelLocation string: Google storage location of the pmml model file.
  • trainingComplete string: Training completion time (as a RFC 3339 timestamp).
  • trainingStatus string: The current status of the training job. This can be one of following: RUNNING; DONE; ERROR; ERROR: TRAINING JOB NOT FOUND

List

• List object
  • items array: List of models.
    • items Insert2
  • kind string: What kind of resource this is.
  • nextPageToken string: Pagination token to fetch the next page, if one exists.
  • selfLink string: A URL to re-request this resource.

Output

• Output object
  • id string: The unique name for the predictive model.
  • kind string: What kind of resource this is.
  • outputLabel string: The most likely class label (Categorical models only).
  • outputMulti array: A list of class labels with their estimated probabilities (Categorical models only).
    • items object
      • label string: The class label.
      • score string: The probability of the class label.
  • outputValue string: The estimated regression value (Regression models only).
  • selfLink string: A URL to re-request this resource.

Update

• Update object
  • csvInstance array: The input features for this instance.
  • output string: The generic output value - could be regression or class label.