package google.cloud.bigquery.v2
Get desktop application:
View/edit binary Protocol Buffers messages
service ModelService
model.proto:34rpc GetModel (, )
model.proto:43Gets the specified model resource by model ID.
message
model.proto:578string project_id = 1
Required. Project ID of the requested model.
string dataset_id = 2
Required. Dataset ID of the requested model.
string model_id = 3
Required. Model ID of the requested model.
rpc ListModels (, )
model.proto:49Lists all models in the specified dataset. Requires the READER dataset role.
message
model.proto:616string project_id = 1
Required. Project ID of the models to list.
string dataset_id = 2
Required. Dataset ID of the models to list.
optional max_results = 3
The maximum number of results to return in a single response page. Leverage the page tokens to iterate through the entire collection.
string page_token = 4
Page token, returned by a previous call to request the next page of results
message
model.proto:632repeated models = 1
Models in the requested dataset. Only the following fields are populated: model_reference, model_type, creation_time, last_modified_time and labels.
string next_page_token = 2
A token to request the next page of results.
rpc PatchModel (, )
model.proto:54Patch specific fields in the specified model.
message
model.proto:589string project_id = 1
Required. Project ID of the model to patch.
string dataset_id = 2
Required. Dataset ID of the model to patch.
string model_id = 3
Required. Model ID of the model to patch.
optional model = 4
Required. Patched model. Follows RFC5789 patch semantics. Missing fields are not updated. To clear a field, explicitly set to default value.
rpc DeleteModel (, )
model.proto:59Deletes the model specified by modelId from the dataset.
message
model.proto:605string project_id = 1
Required. Project ID of the model to delete.
string dataset_id = 2
Required. Dataset ID of the model to delete.
string model_id = 3
Required. Model ID of the model to delete.
message
encryption_config.proto:28Used in:
optional kms_key_name = 1
Optional. Describes the Cloud KMS encryption key that will be used to protect destination BigQuery table. The BigQuery Service Account associated with your project requires access to this encryption key.
message
model.proto:64Used as response type in: ModelService.GetModel, ModelService.PatchModel
Used as field type in: ,
string etag = 1
Output only. A hash of this resource.
optional model_reference = 2
Required. Unique identifier for this model.
int64 creation_time = 5
Output only. The time when this model was created, in millisecs since the epoch.
int64 last_modified_time = 6
Output only. The time when this model was last modified, in millisecs since the epoch.
string description = 12
Optional. A user-friendly description of this model.
string friendly_name = 14
Optional. A descriptive name for this model.
map<string, string> labels = 15
The labels associated with this model. You can use these to organize and group your models. Label keys and values can be no longer than 63 characters, can only contain lowercase letters, numeric characters, underscores and dashes. International characters are allowed. Label values are optional. Label keys must start with a letter and each label in the list must have a different key.
int64 expiration_time = 16
Optional. The time when this model expires, in milliseconds since the epoch. If not present, the model will persist indefinitely. Expired models will be deleted and their storage reclaimed. The defaultTableExpirationMs property of the encapsulating dataset can be used to set a default expirationTime on newly created models.
string location = 13
Output only. The geographic location where the model resides. This value is inherited from the dataset.
optional encryption_configuration = 17
Custom encryption configuration (e.g., Cloud KMS keys). This shows the encryption configuration of the model data while stored in BigQuery storage.
Model.ModelType model_type = 7
Output only. Type of the model resource.
repeated training_runs = 9
Output only. Information for all training runs in increasing order of start_time.
repeated feature_columns = 10
Output only. Input feature columns that were used to train this model.
repeated label_columns = 11
Output only. Label columns that were used to train this model. The output of the model will have a "predicted_" prefix to these columns.
message
model.proto:107Aggregate metrics for classification/classifier models. For multi-class models, the metrics are either macro-averaged or micro-averaged. When macro-averaged, the metrics are calculated for each label and then an unweighted average is taken of those values. When micro-averaged, the metric is calculated globally by counting the total number of correctly predicted rows.
Used in: ,
optional precision = 1
Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
optional recall = 2
Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
optional accuracy = 3
Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
optional threshold = 4
Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
optional f1_score = 5
The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
optional log_loss = 6
Logarithmic Loss. For multiclass this is a macro-averaged metric.
optional roc_auc = 7
Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
message
model.proto:140Evaluation metrics for binary classification/classifier models.
Used in:
optional aggregate_classification_metrics = 1
Aggregate classification metrics.
repeated binary_confusion_matrix_list = 2
Binary confusion matrix at multiple thresholds.
string positive_label = 3
Label representing the positive class.
string negative_label = 4
Label representing the negative class.
message
model.proto:142Confusion matrix for binary classification models.
Used in:
optional positive_class_threshold = 1
Threshold value used when computing each of the following metric.
optional true_positives = 2
Number of true samples predicted as true.
optional false_positives = 3
Number of false samples predicted as true.
optional true_negatives = 4
Number of true samples predicted as false.
optional false_negatives = 5
Number of false samples predicted as false.
optional precision = 6
The fraction of actual positive predictions that had positive actual labels.
optional recall = 7
The fraction of actual positive labels that were given a positive prediction.
optional f1_score = 8
The equally weighted average of recall and precision.
optional accuracy = 9
The fraction of predictions given the correct label.
message
model.proto:226Evaluation metrics for clustering models.
Used in:
optional davies_bouldin_index = 1
Davies-Bouldin index.
optional mean_squared_distance = 2
Mean of squared distances between each sample to its cluster centroid.
repeated clusters = 3
[Beta] Information for all clusters.
message
model.proto:228Message containing the information about one cluster.
Used in:
int64 centroid_id = 1
Centroid id.
repeated feature_values = 2
Values of highly variant features for this cluster.
optional count = 3
Count of training data rows that were assigned to this cluster.
message
model.proto:230Representative value of a single feature within the cluster.
Used in:
string feature_column = 1
The feature column name.
oneof value
numerical_value = 2
The numerical feature value. This is the centroid value for this feature.
categorical_value = 3
The categorical feature value.
message
model.proto:232Representative value of a categorical feature.
Used in:
repeated category_counts = 1
Counts of all categories for the categorical feature. If there are more than ten categories, we return top ten (by count) and return one more CategoryCount with category "_OTHER_" and count as aggregate counts of remaining categories.
message
model.proto:234Represents the count of a single category within the cluster.
Used in:
string category = 1
The name of category.
optional count = 2
The count of training samples matching the category within the cluster.
Indicates the method to split input data into multiple tables.
Used in:
DATA_SPLIT_METHOD_UNSPECIFIED = 0
RANDOM = 1
Splits data randomly.
CUSTOM = 2
Splits data with the user provided tags.
SEQUENTIAL = 3
Splits data sequentially.
NO_SPLIT = 4
Data split will be skipped.
AUTO_SPLIT = 5
Splits data automatically: Uses NO_SPLIT if the data size is small. Otherwise uses RANDOM.
Distance metric used to compute the distance between two points.
Used in:
DISTANCE_TYPE_UNSPECIFIED = 0
EUCLIDEAN = 1
Eculidean distance.
COSINE = 2
Cosine distance.
message
model.proto:286Evaluation metrics of a model. These are either computed on all training data or just the eval data based on whether eval data was used during training. These are not present for imported models.
Used in:
oneof metrics
regression_metrics = 1
Populated for regression models and explicit feedback type matrix factorization models.
binary_classification_metrics = 2
Populated for binary classification/classifier models.
multi_class_classification_metrics = 3
Populated for multi-class classification/classifier models.
clustering_metrics = 4
Populated for clustering models.
message
model.proto:65(message has no fields)
Indicates the method used to initialize the centroids for KMeans clustering algorithm.
Used in:
KMEANS_INITIALIZATION_METHOD_UNSPECIFIED = 0
RANDOM = 1
Initializes the centroids randomly.
CUSTOM = 2
Initializes the centroids using data specified in kmeans_initialization_column.
Indicates the learning rate optimization strategy to use.
Used in:
LEARN_RATE_STRATEGY_UNSPECIFIED = 0
LINE_SEARCH = 1
Use line search to determine learning rate.
CONSTANT = 2
Use a constant learning rate.
Loss metric to evaluate model training performance.
Used in:
LOSS_TYPE_UNSPECIFIED = 0
MEAN_SQUARED_LOSS = 1
Mean squared loss, used for linear regression.
MEAN_LOG_LOSS = 2
Mean log loss, used for logistic regression.
Indicates the type of the Model.
Used in:
MODEL_TYPE_UNSPECIFIED = 0
LINEAR_REGRESSION = 1
Linear regression model.
LOGISTIC_REGRESSION = 2
Logistic regression based classification model.
KMEANS = 3
K-means clustering model.
TENSORFLOW = 6
[Beta] An imported TensorFlow model.
message
model.proto:187Evaluation metrics for multi-class classification/classifier models.
Used in:
optional aggregate_classification_metrics = 1
Aggregate classification metrics.
repeated confusion_matrix_list = 2
Confusion matrix at different thresholds.
message
model.proto:189Confusion matrix for multi-class classification models.
Used in:
optional confidence_threshold = 1
Confidence threshold used when computing the entries of the confusion matrix.
repeated rows = 2
One row per actual label.
message
model.proto:191A single entry in the confusion matrix.
Used in:
string predicted_label = 1
The predicted label. For confidence_threshold > 0, we will also add an entry indicating the number of items under the confidence threshold.
optional item_count = 2
Number of items being predicted as this label.
message
model.proto:202A single row in the confusion matrix.
Used in:
string actual_label = 1
The original label of this row.
repeated entries = 2
Info describing predicted label distribution.
Indicates the optimization strategy used for training.
Used in:
OPTIMIZATION_STRATEGY_UNSPECIFIED = 0
BATCH_GRADIENT_DESCENT = 1
Uses an iterative batch gradient descent algorithm.
NORMAL_EQUATION = 2
Uses a normal equation to solve linear regression problem.
message
model.proto:84Evaluation metrics for regression and explicit feedback type matrix factorization models.
Used in:
optional mean_absolute_error = 1
Mean absolute error.
optional mean_squared_error = 2
Mean squared error.
optional mean_squared_log_error = 3
Mean squared log error.
optional median_absolute_error = 4
Median absolute error.
optional r_squared = 5
R^2 score.
message
model.proto:304Information about a single training query run for the model.
Used in:
optional training_options = 1
Options that were used for this training run, includes user specified and default options that were used.
optional start_time = 8
The start time of this training run.
repeated results = 6
Output of each iteration run, results.size() <= max_iterations.
optional evaluation_metrics = 7
The evaluation metrics over training/eval data that were computed at the end of training.
message
model.proto:392Information about a single iteration of the training run.
Used in:
optional index = 1
Index of the iteration, 0 based.
optional duration_ms = 4
Time taken to run the iteration in milliseconds.
optional training_loss = 5
Loss computed on the training data at the end of iteration.
optional eval_loss = 6
Loss computed on the eval data at the end of iteration.
double learn_rate = 7
Learn rate used for this iteration.
repeated cluster_infos = 8
Information about top clusters for clustering models.
message
model.proto:394Information about a single cluster for clustering model.
Used in:
int64 centroid_id = 1
Centroid id.
optional cluster_radius = 2
Cluster radius, the average distance from centroid to each point assigned to the cluster.
optional cluster_size = 3
Cluster size, the total number of points assigned to the cluster.
message
model.proto:305Used in:
int64 max_iterations = 1
The maximum number of iterations in training. Used only for iterative training algorithms.
LossType loss_type = 2
Type of loss function used during training run.
double learn_rate = 3
Learning rate in training. Used only for iterative training algorithms.
optional l1_regularization = 4
L1 regularization coefficient.
optional l2_regularization = 5
L2 regularization coefficient.
optional min_relative_progress = 6
When early_stop is true, stops training when accuracy improvement is less than 'min_relative_progress'. Used only for iterative training algorithms.
optional warm_start = 7
Whether to train a model from the last checkpoint.
optional early_stop = 8
Whether to stop early when the loss doesn't improve significantly any more (compared to min_relative_progress). Used only for iterative training algorithms.
repeated string input_label_columns = 9
Name of input label columns in training data.
DataSplitMethod data_split_method = 10
The data split type for training and evaluation, e.g. RANDOM.
double data_split_eval_fraction = 11
The fraction of evaluation data over the whole input data. The rest of data will be used as training data. The format should be double. Accurate to two decimal places. Default value is 0.2.
string data_split_column = 12
The column to split data with. This column won't be used as a feature. 1. When data_split_method is CUSTOM, the corresponding column should be boolean. The rows with true value tag are eval data, and the false are training data. 2. When data_split_method is SEQ, the first DATA_SPLIT_EVAL_FRACTION rows (from smallest to largest) in the corresponding column are used as training data, and the rest are eval data. It respects the order in Orderable data types: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties
LearnRateStrategy learn_rate_strategy = 13
The strategy to determine learn rate for the current iteration.
double initial_learn_rate = 16
Specifies the initial learning rate for the line search learn rate strategy.
map<string, double> label_class_weights = 17
Weights associated with each label class, for rebalancing the training data. Only applicable for classification models.
DistanceType distance_type = 20
Distance type for clustering models.
int64 num_clusters = 21
Number of clusters for clustering models.
string model_uri = 22
[Beta] Google Cloud Storage URI from which the model was imported. Only applicable for imported models.
OptimizationStrategy optimization_strategy = 23
Optimization strategy for training linear regression models.
KmeansEnums.KmeansInitializationMethod kmeans_initialization_method = 33
The method used to initialize the centroids for kmeans algorithm.
string kmeans_initialization_column = 34
The column used to provide the initial centroids for kmeans algorithm when kmeans_initialization_method is CUSTOM.
message
model_reference.proto:28Id path of a model.
Used in:
string project_id = 1
Required. The ID of the project containing this model.
string dataset_id = 2
Required. The ID of the dataset containing this model.
string model_id = 3
Required. The ID of the model. The ID must contain only letters (a-z, A-Z), numbers (0-9), or underscores (_). The maximum length is 1,024 characters.
message
standard_sql.proto:37The type of a variable, e.g., a function argument. Examples: INT64: {type_kind="INT64"} ARRAY<STRING>: {type_kind="ARRAY", array_element_type="STRING"} STRUCT<x STRING, y ARRAY<DATE>>: {type_kind="STRUCT", struct_type={fields=[ {name="x", type={type_kind="STRING"}}, {name="y", type={type_kind="ARRAY", array_element_type="DATE"}} ]}}
Used in:
StandardSqlDataType.TypeKind type_kind = 1
Required. The top level type of this field. Can be any standard SQL data type (e.g., "INT64", "DATE", "ARRAY").
oneof sub_type
array_element_type = 2
The type of the array's elements, if type_kind = "ARRAY".
struct_type = 3
The fields of this struct, in order, if type_kind = "STRUCT".
Used in:
TYPE_KIND_UNSPECIFIED = 0
Invalid type.
INT64 = 2
Encoded as a string in decimal format.
BOOL = 5
Encoded as a boolean "false" or "true".
FLOAT64 = 7
Encoded as a number, or string "NaN", "Infinity" or "-Infinity".
STRING = 8
Encoded as a string value.
BYTES = 9
Encoded as a base64 string per RFC 4648, section 4.
TIMESTAMP = 19
Encoded as an RFC 3339 timestamp with mandatory "Z" time zone string: 1985-04-12T23:20:50.52Z
DATE = 10
Encoded as RFC 3339 full-date format string: 1985-04-12
TIME = 20
Encoded as RFC 3339 partial-time format string: 23:20:50.52
DATETIME = 21
Encoded as RFC 3339 full-date "T" partial-time: 1985-04-12T23:20:50.52
GEOGRAPHY = 22
Encoded as WKT
NUMERIC = 23
Encoded as a decimal string.
ARRAY = 16
Encoded as a list with types matching Type.array_type.
STRUCT = 17
Encoded as a list with fields of type Type.struct_type[i]. List is used because a JSON object cannot have duplicate field names.
message
standard_sql.proto:98A field or a column.
Used in: ,
string name = 1
Optional. The name of this field. Can be absent for struct fields.
optional type = 2
Optional. The type of this parameter. Absent if not explicitly specified (e.g., CREATE FUNCTION statement can omit the return type; in this case the output parameter does not have this "type" field).
message
standard_sql.proto:108Used in:
repeated fields = 1