log_reduce_predict_full_fl()

The function log_reduce_predict_full_fl() parses semi structured textual columns, such as log lines, and for each line it matches the respective pattern from a pretrained model or reports an anomaly if no matching pattern was found. The patterns are retrieved from a pretrained model, generated by log_reduce_train_fl(). The function is similar to log_reduce_predict_fl(), but unlike log_reduce_predict_fl() that outputs a patterns summary table, this function outputs a full table containing the pattern and parameters per each line.

Syntax

T | invoke log_reduce_predict_full_fl(models_tbl, model_name, reduce_col, pattern_col, parameters_col [, anomaly_str ])

Parameters

Function definition

You can define the function by either embedding its code as a query-defined function, or creating it as a stored function in your database, as follows:

Query-defined

Define the function using the following let statement. No permissions are required.

let log_reduce_predict_full_fl=(tbl:(*), models_tbl: (name:string, timestamp: datetime, model:string), 
                           model_name:string, reduce_col:string, pattern_col:string, parameters_col:string, 
                           anomaly_str: string = 'ANOMALY')
{
    let model_str = toscalar(models_tbl | where name == model_name | top 1 by timestamp desc | project model);
    let kwargs = bag_pack('logs_col', reduce_col, 'output_patterns_col', pattern_col,'output_parameters_col', 
                          parameters_col, 'model', model_str, 'anomaly_str', anomaly_str, 'output_type', 'full');
    let code = ```if 1:
        from log_cluster import log_reduce_predict
        result = log_reduce_predict.log_reduce_predict(df, kargs)
    ```;
    tbl
    | evaluate hint.distribution=per_node python(typeof(*), code, kwargs)
};
// Write your query to use the function here.

Stored

Define the stored function once using the following .create function. Database User permissions are required.

.create-or-alter function with (folder = 'Packages\\Text', docstring = 'Apply a trained model to find common patterns in textual logs, output a full table')
log_reduce_predict_full_fl(tbl:(*), models_tbl: (name:string, timestamp: datetime, model:string), 
                           model_name:string, reduce_col:string, pattern_col:string, parameters_col:string, 
                           anomaly_str: string = 'ANOMALY')
{
    let model_str = toscalar(models_tbl | where name == model_name | top 1 by timestamp desc | project model);
    let kwargs = bag_pack('logs_col', reduce_col, 'output_patterns_col', pattern_col,'output_parameters_col', 
                          parameters_col, 'model', model_str, 'anomaly_str', anomaly_str, 'output_type', 'full');
    let code = ```if 1:
        from log_cluster import log_reduce_predict
        result = log_reduce_predict.log_reduce_predict(df, kargs)
    ```;
    tbl
    | evaluate hint.distribution=per_node python(typeof(*), code, kwargs)
}

Example

The following example uses the invoke operator to run the function.

Query-defined

To use a query-defined function, invoke it after the embedded function definition.

let log_reduce_predict_full_fl=(tbl:(*), models_tbl: (name:string, timestamp: datetime, model:string), 
                           model_name:string, reduce_col:string, pattern_col:string, parameters_col:string, 
                           anomaly_str: string = 'ANOMALY')
{
    let model_str = toscalar(models_tbl | where name == model_name | top 1 by timestamp desc | project model);
    let kwargs = bag_pack('logs_col', reduce_col, 'output_patterns_col', pattern_col,'output_parameters_col', 
                          parameters_col, 'model', model_str, 'anomaly_str', anomaly_str, 'output_type', 'full');
    let code = ```if 1:
        from log_cluster import log_reduce_predict
        result = log_reduce_predict.log_reduce_predict(df, kargs)
    ```;
    tbl
    | evaluate hint.distribution=per_node python(typeof(*), code, kwargs)
};
HDFS_log_100k
| extend Patterns='', Parameters=''
| take 10
| invoke log_reduce_predict_full_fl(models_tbl=ML_Models, model_name="HDFS_100K", reduce_col="data", pattern_col="Patterns", parameters_col="Parameters")

Stored

HDFS_log_100k
| extend Patterns='', Parameters=''
| take 10
| invoke log_reduce_predict_full_fl(models_tbl=ML_Models, model_name="HDFS_100K", reduce_col="data", pattern_col="Patterns", parameters_col="Parameters")

Output