mirror of
https://github.com/ckaczor/azuredatastudio.git
synced 2026-02-16 10:58:30 -05:00
Machine Learning Services Extension - Predict wizard (#9450)
*MLS extension - Added predict wizard
This commit is contained in:
Leila Lali
@@ -0,0 +1,24 @@
|
||||
/*---------------------------------------------------------------------------------------------
|
||||
* Copyright (c) Microsoft Corporation. All rights reserved.
|
||||
* Licensed under the Source EULA. See License.txt in the project root for license information.
|
||||
*--------------------------------------------------------------------------------------------*/
|
||||
|
||||
export interface PredictColumn {
|
||||
name: string;
|
||||
dataType?: string;
|
||||
displayName?: string;
|
||||
}
|
||||
|
||||
export interface DatabaseTable {
|
||||
databaseName: string | undefined;
|
||||
tableName: string | undefined;
|
||||
schema: string | undefined
|
||||
}
|
||||
|
||||
export interface PredictInputParameters extends DatabaseTable {
|
||||
inputColumns: PredictColumn[] | undefined
|
||||
}
|
||||
|
||||
export interface PredictParameters extends PredictInputParameters {
|
||||
outputColumns: PredictColumn[] | undefined
|
||||
}
|
||||
@@ -0,0 +1,203 @@
|
||||
/*---------------------------------------------------------------------------------------------
|
||||
* Copyright (c) Microsoft Corporation. All rights reserved.
|
||||
* Licensed under the Source EULA. See License.txt in the project root for license information.
|
||||
*--------------------------------------------------------------------------------------------*/
|
||||
|
||||
import * as azdata from 'azdata';
|
||||
|
||||
import { ApiWrapper } from '../common/apiWrapper';
|
||||
import { QueryRunner } from '../common/queryRunner';
|
||||
import * as utils from '../common/utils';
|
||||
import { RegisteredModel } from '../modelManagement/interfaces';
|
||||
import { PredictParameters, PredictColumn, DatabaseTable } from '../prediction/interfaces';
|
||||
import { Config } from '../configurations/config';
|
||||
|
||||
/**
|
||||
* Service to make prediction
|
||||
*/
|
||||
export class PredictService {
|
||||
|
||||
/**
|
||||
* Creates new instance
|
||||
*/
|
||||
constructor(
|
||||
private _apiWrapper: ApiWrapper,
|
||||
private _queryRunner: QueryRunner,
|
||||
private _config: Config) {
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns the list of databases
|
||||
*/
|
||||
public async getDatabaseList(): Promise<string[]> {
|
||||
let connection = await this.getCurrentConnection();
|
||||
if (connection) {
|
||||
return await this._apiWrapper.listDatabases(connection.connectionId);
|
||||
}
|
||||
return [];
|
||||
}
|
||||
|
||||
/**
|
||||
* Generates prediction script given model info and predict parameters
|
||||
* @param predictParams predict parameters
|
||||
* @param registeredModel model parameters
|
||||
*/
|
||||
public async generatePredictScript(
|
||||
predictParams: PredictParameters,
|
||||
registeredModel: RegisteredModel | undefined,
|
||||
filePath: string | undefined
|
||||
): Promise<string> {
|
||||
let connection = await this.getCurrentConnection();
|
||||
let query = '';
|
||||
if (registeredModel && registeredModel.id) {
|
||||
query = this.getPredictScriptWithModelId(
|
||||
registeredModel.id,
|
||||
predictParams.inputColumns || [],
|
||||
predictParams.outputColumns || [],
|
||||
predictParams);
|
||||
} else if (filePath) {
|
||||
let modelBytes = await utils.readFileInHex(filePath || '');
|
||||
query = this.getPredictScriptWithModelBytes(modelBytes, predictParams.inputColumns || [],
|
||||
predictParams.outputColumns || [],
|
||||
predictParams);
|
||||
}
|
||||
let document = await this._apiWrapper.openTextDocument({
|
||||
language: 'sql',
|
||||
content: query
|
||||
});
|
||||
await this._apiWrapper.showTextDocument(document.uri);
|
||||
await this._apiWrapper.connect(document.uri.toString(), connection.connectionId);
|
||||
this._apiWrapper.runQuery(document.uri.toString(), undefined, false);
|
||||
return query;
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns list of tables given database name
|
||||
* @param databaseName database name
|
||||
*/
|
||||
public async getTableList(databaseName: string): Promise<DatabaseTable[]> {
|
||||
let connection = await this.getCurrentConnection();
|
||||
let list: DatabaseTable[] = [];
|
||||
if (connection) {
|
||||
let query = utils.getScriptWithDBChange(connection.databaseName, databaseName, this.getTablesScript(databaseName));
|
||||
let result = await this._queryRunner.safeRunQuery(connection, query);
|
||||
if (result && result.rows && result.rows.length > 0) {
|
||||
result.rows.forEach(row => {
|
||||
list.push({
|
||||
databaseName: databaseName,
|
||||
tableName: row[0].displayValue,
|
||||
schema: row[1].displayValue
|
||||
});
|
||||
});
|
||||
}
|
||||
}
|
||||
return list;
|
||||
}
|
||||
|
||||
/**
|
||||
*Returns list of column names of a database
|
||||
* @param databaseTable table info
|
||||
*/
|
||||
public async getTableColumnsList(databaseTable: DatabaseTable): Promise<string[]> {
|
||||
let connection = await this.getCurrentConnection();
|
||||
let list: string[] = [];
|
||||
if (connection && databaseTable.databaseName) {
|
||||
const query = utils.getScriptWithDBChange(connection.databaseName, databaseTable.databaseName, this.getTableColumnsScript(databaseTable));
|
||||
let result = await this._queryRunner.safeRunQuery(connection, query);
|
||||
if (result && result.rows && result.rows.length > 0) {
|
||||
result.rows.forEach(row => {
|
||||
list.push(row[0].displayValue);
|
||||
});
|
||||
}
|
||||
}
|
||||
return list;
|
||||
}
|
||||
|
||||
private async getCurrentConnection(): Promise<azdata.connection.ConnectionProfile> {
|
||||
return await this._apiWrapper.getCurrentConnection();
|
||||
}
|
||||
|
||||
private getTableColumnsScript(databaseTable: DatabaseTable): string {
|
||||
return `
|
||||
SELECT COLUMN_NAME,*
|
||||
FROM INFORMATION_SCHEMA.COLUMNS
|
||||
WHERE TABLE_NAME='${utils.doubleEscapeSingleQuotes(databaseTable.tableName)}'
|
||||
AND TABLE_SCHEMA='${utils.doubleEscapeSingleQuotes(databaseTable.schema)}'
|
||||
AND TABLE_CATALOG='${utils.doubleEscapeSingleQuotes(databaseTable.databaseName)}'
|
||||
`;
|
||||
}
|
||||
|
||||
private getTablesScript(databaseName: string): string {
|
||||
return `
|
||||
SELECT TABLE_NAME,TABLE_SCHEMA
|
||||
FROM INFORMATION_SCHEMA.TABLES
|
||||
WHERE TABLE_TYPE = 'BASE TABLE' AND TABLE_CATALOG='${utils.doubleEscapeSingleQuotes(databaseName)}'
|
||||
`;
|
||||
}
|
||||
|
||||
private getPredictScriptWithModelId(
|
||||
modelId: number,
|
||||
columns: PredictColumn[],
|
||||
outputColumns: PredictColumn[],
|
||||
databaseNameTable: DatabaseTable): string {
|
||||
return `
|
||||
DECLARE @model VARBINARY(max) = (
|
||||
SELECT artifact_content
|
||||
FROM ${utils.getRegisteredModelsThreePartsName(this._config)}
|
||||
WHERE artifact_id = ${modelId}
|
||||
);
|
||||
WITH predict_input
|
||||
AS (
|
||||
SELECT TOP 1000
|
||||
${this.getColumnNames(columns, 'pi')}
|
||||
FROM [${utils.doubleEscapeSingleBrackets(databaseNameTable.databaseName)}].[${databaseNameTable.schema}].[${utils.doubleEscapeSingleBrackets(databaseNameTable.tableName)}] as pi
|
||||
)
|
||||
SELECT
|
||||
${this.getInputColumnNames(columns, 'predict_input')}, ${this.getColumnNames(outputColumns, 'p')}
|
||||
FROM PREDICT(MODEL = @model, DATA = predict_input)
|
||||
WITH (
|
||||
${this.getColumnTypes(outputColumns)}
|
||||
) AS p
|
||||
`;
|
||||
}
|
||||
|
||||
private getPredictScriptWithModelBytes(
|
||||
modelBytes: string,
|
||||
columns: PredictColumn[],
|
||||
outputColumns: PredictColumn[],
|
||||
databaseNameTable: DatabaseTable): string {
|
||||
return `
|
||||
WITH predict_input
|
||||
AS (
|
||||
SELECT TOP 1000
|
||||
${this.getColumnNames(columns, 'pi')}
|
||||
FROM [${utils.doubleEscapeSingleBrackets(databaseNameTable.databaseName)}].[${databaseNameTable.schema}].[${utils.doubleEscapeSingleBrackets(databaseNameTable.tableName)}] as pi
|
||||
)
|
||||
SELECT
|
||||
${this.getInputColumnNames(columns, 'predict_input')}, ${this.getColumnNames(outputColumns, 'p')}
|
||||
FROM PREDICT(MODEL = ${modelBytes}, DATA = predict_input)
|
||||
WITH (
|
||||
${this.getColumnTypes(outputColumns)}
|
||||
) AS p
|
||||
`;
|
||||
}
|
||||
|
||||
private getColumnNames(columns: PredictColumn[], tableName: string) {
|
||||
return columns.map(c => {
|
||||
return c.displayName ? `${tableName}.${c.name} AS ${c.displayName}` : `${tableName}.${c.name}`;
|
||||
}).join(',\n');
|
||||
}
|
||||
|
||||
private getInputColumnNames(columns: PredictColumn[], tableName: string) {
|
||||
return columns.map(c => {
|
||||
return c.displayName ? `${tableName}.${c.displayName}` : `${tableName}.${c.name}`;
|
||||
}).join(',\n');
|
||||
}
|
||||
|
||||
private getColumnTypes(columns: PredictColumn[]) {
|
||||
return columns.map(c => {
|
||||
return `${c.name} ${c.dataType}`;
|
||||
}).join(',\n');
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user