Giter Club home page Giter Club logo

mondrian-rest's Introduction

Overview

This repository contains a Spring Boot -based REST API for querying relational OLAP data sources via Mondrian.

Table of Contents

Motivation

The Mondrian library is a terrific way to submit MDX queries to relational data sources. Great interface platforms, like Saiku and Pivot4J provide full-featured ad-hoc queries for end-users comfortable with pivot tables. There are also toolkits like Pentaho CTools for developing dashboards.

What has been missing is a simple, basic, and clean REST API wrapped around the Mondrian library. Each of the tools just mentioned uses its own approach for programmatic access to Mondrian, and does not really make that approach available (in an easy way) outside of that tool. Thus, we decided to implement the mondrian-rest API.

Mondrian versions

Initial development of mondrian-rest supported version 4.x of the Mondrian library. However, Mondrian 4.x does not appear to be under active development by the core Mondrian maintainers, so we have abandoned it as a dependency. The last version of mondrian-rest that supports version 4.x schemas is version 1.5.0. PRs for the 1.x line of mondrian-rest will be considered, but active development of mondrian-rest by the core committers is focused on Mondrian 8.x (and, at some point, version 9.x). There are significant schema differences between 4.x and 8.x+, however for the most part, the mondrian-rest API is consistent between mondrian-rest v2.x and v1.x.

General Usage

You can run the API in Docker or in an existing J2EE servlet container. Instructions for each option appear below.

The API web application includes the Mondrian FoodMart demonstration database and schema, and a small database and schema that we use for testing. By default, these connections will be available when the API application is started. To turn them off, set an application property removeDemoConnections to true. Instructions for accomplishing this appear in the Docker and J2EE container sections below.

Configuration

Configuration of the controller occurs via editing of the Spring application.properties file. The file contains documentation for each of the available properties. To configure an instance of the controller, edit the file and place it on Tomcat's classpath (typically this is accomplished by enabling Tomcat's shared loader to point to a directory, and then placing the edited properties file in that directory). If you run the application directly via the executable war file, you can set any of the available properties with Java system properties as well.

Configuration allows you to control the following behavior of the controller:

  • User authentication and mapping of users to Mondrian roles (for enabling schema security)
  • Whether the included demo connections / data sources are enabled
  • Whether the simple query UI is enabled
  • Pre-caching of metadata (to trade off longer controller startup time for quicker response to the first /query or /getMetadata call)
  • Setting a timeout for queries

Via Docker

There are pre-built Docker images on DockerHub with the API installed, one with default (i.e., no) security, and one configured with Bearer Token authentication and role-mapping. The unsecured image is ojbc/mondrian-rest:no-auth (which is also tagged as ojbc/mondrian-rest:latest), and the secured image is ojbc/mondrian-rest:bearer-auth. (For more on authentication and role-mapping, see the security section below.)

The images expose the Mondrian demo FoodMart database and schema, and a small database/schema used for testing.

To run it in a local Docker environment:

docker run -d --name=mondrian-rest -p 58080:80 ojbc/mondrian-rest (replace 58080 with whatever port you prefer)

The API will then be running in Docker, exposed on that port, within Apache Tomcat.

To deploy connections with your custom schema and database information (i.e., for a "real" scenario), do the following in a derived image Dockerfile:

  1. Create a *mondrian-connections.json file with the appropriate connection information for your database and schema, and place this file on the Tomcat shared classpath directory at /opt/tomcat/shared/. You can see examples in the codebase. Important: For the application to find them, the files must end in the string mondrian-connections.json and must be on the classpath. So /opt/tomcat/shared/my-application-mondrian-connections.json will work, but .../my-application.json and .../my-application-connections.json will not. Note that the reference to the Mondrian schema in the connection config must be preceded by a forward slash, assuming that you place the schema in the shared Tomcat directory root per the next step. Be sure to set the IsDemo property in the json structure to false, otherwise if you set removeDemoConnections to true via application properties, your connection will be removed.
  2. Place your Mondrian schema in /opt/tomcat/shared and ensure it has the same name as the schema reference in your *mondrian-connections.json file addressed in the prior step.

We typically create a custom Docker bridge network, and deploy the database and API in separate containers. This enables the API container to reference the database via DNS, by container name.

If you want to disable the demo data connections, in your derived image Dockerfile, create an application.properties file with a line removeDemoConnections=true in it, and COPY this file to /opt/tomcat/shared/config.

Via J2EE Container or Spring Boot

If you don't want to run the API from Docker, just drop mondrian-rest.war (see Building from Source below) into your favorite J2EE servlet container (e.g., Tomcat's webapps directory).

To remove demo connections and deploy your custom connections, follow instructions similar to those in the Docker section above as appropriate for your J2EE Servlet container. For Tomcat, you can establish a directory as a "shared classpath" directory by modifying catalina.properties and setting the shared.loader property to the desired location.

You can also run the web application as a Spring Boot standalone application, with embedded Tomcat. From the root directory (where pom.xml is), run mondrian-rest-executable.war like this:

$ java -jar target/mondrian-rest-executable.war

  .   ____          _            __ _ _
 /\\ / ___'_ __ _ _(_)_ __  __ _ \ \ \ \
( ( )\___ | '_ | '_| | '_ \/ _` | \ \ \ \
 \\/  ___)| |_)| | | | | || (_| |  ) ) ) )
  '  |____| .__|_| |_|_| |_\__, | / / / /
 =========|_|==============|___/=/_/_/_/
 :: Spring Boot ::        (v2.2.2.RELEASE)

[      main] org.ojbc.mondrian.rest.Application       INFO  Starting Application v2.0.1 on smc-mbp.local with PID 73198 (/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war started by scott in /Users/scott/git-repos/ojbc/mondrian-rest)
[      main] org.ojbc.mondrian.rest.Application       INFO  No active profile set, falling back to default profiles: default
[      main] boot.web.embedded.tomcat.TomcatWebServer INFO  Tomcat initialized with port(s): 8080 (http)
[      main] g.apache.coyote.http11.Http11NioProtocol INFO  Initializing ProtocolHandler ["http-nio-8080"]
[      main] org.apache.catalina.core.StandardService INFO  Starting service [Tomcat]
[      main] org.apache.catalina.core.StandardEngine  INFO  Starting Servlet engine: [Apache Tomcat/9.0.29]
[      main] e.ContainerBase.[Tomcat].[localhost].[/] INFO  Initializing Spring embedded WebApplicationContext
[      main] pringframework.web.context.ContextLoader INFO  Root WebApplicationContext: initialization completed in 2443 ms
[      main] mondrian.olap.MondrianProperties         INFO  Mondrian: properties loaded from 'jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes!/mondrian.properties'
[      main] mondrian.olap.MondrianProperties         INFO  Mondrian: loaded 0 system properties
[      main] jbc.mondrian.rest.MondrianRestController INFO  Initializing controller, Mondrian version is: 8.0
[      main] jbc.mondrian.rest.MondrianRestController INFO  No query timeout specified
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Working around Spring Boot / Tomcat bug that occurs in standalone mode, to adjust file path found via PathMatchingResourcePatternResolver
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Processing connection definition json found at jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes/mondrian-connections.json
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Adding valid connection test: connection string=jdbc:hsqldb:res:test, Mondrian schema path=jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes!/test.xml
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Adding valid connection foodmart: connection string=jdbc:hsqldb:res:foodmart;set schema "foodmart", Mondrian schema path=https://raw.githubusercontent.com/pentaho/mondrian/master/demo/FoodMart.xml
[      main] org.ehcache.core.EhcacheManager          INFO  Cache 'metadata-cache' created in EhcacheManager.
[      main] org.ehcache.core.EhcacheManager          INFO  Cache 'query-cache' created in EhcacheManager.
[      main] jbc.mondrian.rest.MondrianRestController INFO  Successfully registered request authorizer class org.ojbc.mondrian.rest.DefaultRequestAuthorizer
[      main] duling.concurrent.ThreadPoolTaskExecutor INFO  Initializing ExecutorService 'applicationTaskExecutor'
[      main] g.apache.coyote.http11.Http11NioProtocol INFO  Starting ProtocolHandler ["http-nio-8080"]
[      main] boot.web.embedded.tomcat.TomcatWebServer INFO  Tomcat started on port(s): 8080 (http) with context path ''
[      main] org.ojbc.mondrian.rest.Application       INFO  Started Application in 5.702 seconds (JVM running for 6.699)

Then the API will be available at http://localhost:8080/... (e.g., http://localhost:8080/getConnections, http://localhost:8080/query, etc.)

If you want to disable the demo data connections, pass removeDemoConnections=true as a system property:

$ java -DremoveDemoConnections=true -jar target/mondrian-rest-executable.war

To load custom connections, run the executable warfile and pass a system property shared.loader as described here, to add that directory to the Spring Boot classpath. Then place the *mondrian-connections.json file and Mondrian schema XML in that directory, similar to the manner described in the Docker section above.

API Usage

There are currently four main operations available in the API:

/getConnections: Returns information about the olap4j connections currently available to the API

Example:

$: curl http://localhost:58080/mondrian-rest/getConnections
{
  "test" : {
    "JdbcDrivers" : "org.hsqldb.jdbc.JDBCDriver",
    "Jdbc" : "jdbc:hsqldb:res:test",
    "Description" : "Main version of test connection",
    "ConnectionDefinitionSource" : "/opt/tomcat/webapps/mondrian-rest/WEB-INF/classes/mondrian-connections.json",
    "MondrianSchemaUrl" : "file:/opt/tomcat/webapps/mondrian-rest/WEB-INF/classes/test.xml"
  },
  "foodmart" : {
    "JdbcDrivers" : "org.hsqldb.jdbc.JDBCDriver",
    "Jdbc" : "jdbc:hsqldb:res:foodmart;set schema \"foodmart\"",
    "Description" : "Pentaho/Hyde FoodMart Database",
    "ConnectionDefinitionSource" : "/opt/tomcat/webapps/mondrian-rest/WEB-INF/classes/mondrian-connections.json",
    "MondrianSchemaUrl" : "https://raw.githubusercontent.com/pentaho/mondrian/lagunitas/demo/FoodMart.mondrian.xml"
  }
}
$:

/getSchema?connectionName=[name]: Retrieves the XML for the Mondrian schema for the specified connection

Example:

$: curl -s http://localhost:58080/mondrian-rest/getSchema?connectionName=foodmart | head -30
<?xml version="1.0"?>
<Schema name="FoodMart">
<!--
  == This software is subject to the terms of the Eclipse Public License v1.0
  == Agreement, available at the following URL:
  == http://www.eclipse.org/legal/epl-v10.html.
  == You must accept the terms of that agreement to use this software.
  ==
  == Copyright (C) 2000-2005 Julian Hyde
  == Copyright (C) 2005-2019 Hitachi Vantara and others
  == All Rights Reserved.
  -->

<!-- Shared dimensions -->

  <Dimension name="Store">
    <Hierarchy hasAll="true" primaryKey="store_id">
      <Table name="store"/>
      <Level name="Store Country" column="store_country" uniqueMembers="true"/>
      <Level name="Store State" column="store_state" uniqueMembers="true"/>
      <Level name="Store City" column="store_city" uniqueMembers="false"/>
      <Level name="Store Name" column="store_name" uniqueMembers="true">
        <Property name="Store Type" column="store_type"/>
        <Property name="Store Manager" column="store_manager"/>
        <Property name="Store Sqft" column="store_sqft" type="Numeric"/>
        <Property name="Grocery Sqft" column="grocery_sqft" type="Numeric"/>
        <Property name="Frozen Sqft" column="frozen_sqft" type="Numeric"/>
        <Property name="Meat Sqft" column="meat_sqft" type="Numeric"/>
        <Property name="Has coffee bar" column="coffee_bar" type="Boolean"/>
        <Property name="Street address" column="store_street_address" type="String"/>
$:

/getMetadata?connectionName=[name]: Retrieves the olap4j metadata structure for the specified connection

This contains much the same information as the /getSchema method. However, /getMetadata actually retrieves the metadata from a working connection, which enhances the returned information with data pulled from the underlying database source. Importantly, this includes the Members within each Level in each Dimension's Hierarchy.

Example:

$ curl -s http://localhost:8080/getMetadata?connectionName=foodmart | head -50
{
  "name" : "FoodMart",
  "connectionName" : "foodmart",
  "cubes" : [ {
    "name" : "Warehouse",
    "caption" : "Warehouse",
    "measures" : [ {
      "name" : "Store Invoice",
      "caption" : "Store Invoice",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Supply Time",
      "caption" : "Supply Time",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Warehouse Cost",
      "caption" : "Warehouse Cost",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Warehouse Sales",
      "caption" : "Warehouse Sales",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Units Shipped",
      "caption" : "Units Shipped",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Units Ordered",
      "caption" : "Units Ordered",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Warehouse Profit",
      "caption" : "Warehouse Profit",
      "visible" : true,
      "calculated" : false
    }, {
      "name" : "Fact Count",
      "caption" : "Fact Count",
      "visible" : false,
      "calculated" : false
    }, {
      "name" : "Average Warehouse Sale",
      "caption" : "Average Warehouse Sale",
      "visible" : true,
$:

/query: POST operation that takes an object with three properties:

  • connectionName specifies the connection to query
  • query specifies the MDX for the query
  • tidy is a contained object to request "tidying" of the output, as explained below; the tidy object contains three properties:
    • enabled is a boolean to indicate whether to tidy the output
    • simplifyNames is a boolean to indicate whether to return simple names for dimension levels
    • levelNameTranslationMap is a map that associates level unique names with arbitrary simple names

Non-tidy example:

$: curl -s -X POST -H 'Content-Type: application/json' http://localhost:58080/mondrian-rest/query -d '{
>   "connectionName" : "foodmart",
>   "query" : "select { [Measures].[Units Shipped] } on columns, NON EMPTY [Store].[Store Type].members on rows from Warehouse"
> }'
{
  "cells" : [ {
    "formattedValue" : "207726.0",
    "value" : 207726,
    "ordinal" : 0,
    "coordinates" : [ 0, 0 ],
    "error" : null
  }, {
    "formattedValue" : "64804.0",
    "value" : 64804.0,
    "ordinal" : 1,
    "coordinates" : [ 0, 1 ],
    "error" : null
  }, {
    "formattedValue" : "10759.0",
    "value" : 10759.0,
    "ordinal" : 2,
    "coordinates" : [ 0, 2 ],
    "error" : null
  }, {
    "formattedValue" : "10589.0",
    "value" : 10589.0,
    "ordinal" : 3,
    "coordinates" : [ 0, 3 ],
    "error" : null
  }, {
    "formattedValue" : "5904.0",
    "value" : 5904.0,
    "ordinal" : 4,
    "coordinates" : [ 0, 4 ],
    "error" : null
  }, {
    "formattedValue" : "115670.0",
    "value" : 115670.0,
    "ordinal" : 5,
    "coordinates" : [ 0, 5 ],
    "error" : null
  } ],
  "axes" : [ {
    "ordinal" : 0,
    "name" : "COLUMNS",
    "positions" : [ {
      "memberDimensionNames" : [ "Measures" ],
      "memberDimensionCaptions" : [ "Measures" ],
      "memberDimensionValues" : [ "Units Shipped" ]
    } ]
  }, {
    "ordinal" : 1,
    "name" : "ROWS",
    "positions" : [ {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "All Store Types" ]
    }, {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "Deluxe Supermarket" ]
    }, {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "Gourmet Supermarket" ]
    }, {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "Mid-Size Grocery" ]
    }, {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "Small Grocery" ]
    }, {
      "memberDimensionNames" : [ "Store" ],
      "memberDimensionCaptions" : [ "Store" ],
      "memberDimensionValues" : [ "Supermarket" ]
    } ]
  } ]
}
$:

Authentication and Mondrian Security Integration

The API supports authentication of users and mapping of user credentials to roles defined in each connection's Mondrian schema. Out of the box, the API supports Bearer Token authentication (and role-mapping by token), and SAML Assertion authentication via Shibboleth (with role-mapping by an assertion attribute with the Name gfipm:2.0:user:FederationId). Implementers can inject other authentication/role-mapping approaches via Spring (more details below).

To authenticate and map via Bearer Token authentication:

  1. Run the API web application with an application property of requestAuthorizerBeanName set to the default bean name for the authorizer class. The default bean name for the Bearer Token authorizer is bearerTokenRequestAuthorizer and for the SAML assertion authorizer is samlAssertionRequestAuthorizer. (To "run the application with an application property", see instructions in prior sections for setting the removeDemoConnections property.)
  2. Place an appropriate authorizer configuration file on the classpath (again, following instructions for where to place *mondrian-connections.json files on the classpath under the different deployment scenarios in prior sections). The configuration file identifies the credentials that are authorized to use each connection, and maps each credential to a Mondrian schema role. The format of this file will be specific to each authorizer implementation; consult the codebase for examples of Bearer Token and SAML Assertion configurations.

The API application includes a Bearer Token authentication configuration designed to work with the two roles defined in the Mondrian demo FoodMart schema ("California Manager" and "No HR Cube" roles). To see this in action (using the Spring Boot deployment scenario here for illustration):

$ java -DrequestAuthorizerBeanName=bearerTokenRequestAuthorizer -jar target/mondrian-rest-executable.war

  .   ____          _            __ _ _
 /\\ / ___'_ __ _ _(_)_ __  __ _ \ \ \ \
( ( )\___ | '_ | '_| | '_ \/ _` | \ \ \ \
 \\/  ___)| |_)| | | | | || (_| |  ) ) ) )
  '  |____| .__|_| |_|_| |_\__, | / / / /
 =========|_|==============|___/=/_/_/_/
 :: Spring Boot ::        (v2.2.2.RELEASE)

[      main] org.ojbc.mondrian.rest.Application       INFO  Starting Application v2.0.1 on smc-mbp.local with PID 78583 (/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war started by scott in /Users/scott/git-repos/ojbc/mondrian-rest)
[      main] org.ojbc.mondrian.rest.Application       INFO  No active profile set, falling back to default profiles: default
[      main] boot.web.embedded.tomcat.TomcatWebServer INFO  Tomcat initialized with port(s): 8080 (http)
[      main] g.apache.coyote.http11.Http11NioProtocol INFO  Initializing ProtocolHandler ["http-nio-8080"]
[      main] org.apache.catalina.core.StandardService INFO  Starting service [Tomcat]
[      main] org.apache.catalina.core.StandardEngine  INFO  Starting Servlet engine: [Apache Tomcat/9.0.29]
[      main] e.ContainerBase.[Tomcat].[localhost].[/] INFO  Initializing Spring embedded WebApplicationContext
[      main] pringframework.web.context.ContextLoader INFO  Root WebApplicationContext: initialization completed in 2076 ms
[      main] mondrian.olap.MondrianProperties         INFO  Mondrian: properties loaded from 'jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes!/mondrian.properties'
[      main] mondrian.olap.MondrianProperties         INFO  Mondrian: loaded 0 system properties
[      main] jbc.mondrian.rest.MondrianRestController INFO  Initializing controller, Mondrian version is: 8.0
[      main] jbc.mondrian.rest.MondrianRestController INFO  No query timeout specified
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Working around Spring Boot / Tomcat bug that occurs in standalone mode, to adjust file path found via PathMatchingResourcePatternResolver
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Processing connection definition json found at jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes/mondrian-connections.json
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Adding valid connection test: connection string=jdbc:hsqldb:res:test, Mondrian schema path=jar:file:/Users/scott/git-repos/ojbc/mondrian-rest/target/mondrian-rest-executable.war!/WEB-INF/classes!/test.xml
[      main] .ojbc.mondrian.MondrianConnectionFactory INFO  Adding valid connection foodmart: connection string=jdbc:hsqldb:res:foodmart;set schema "foodmart", Mondrian schema path=https://raw.githubusercontent.com/pentaho/mondrian/master/demo/FoodMart.xml
[      main] org.ehcache.core.EhcacheManager          INFO  Cache 'metadata-cache' created in EhcacheManager.
[      main] org.ehcache.core.EhcacheManager          INFO  Cache 'query-cache' created in EhcacheManager.
[      main] jbc.mondrian.rest.MondrianRestController INFO  Successfully registered request authorizer class org.ojbc.mondrian.rest.BearerTokenRequestAuthorizer << ***
[      main] duling.concurrent.ThreadPoolTaskExecutor INFO  Initializing ExecutorService 'applicationTaskExecutor'
[      main] g.apache.coyote.http11.Http11NioProtocol INFO  Starting ProtocolHandler ["http-nio-8080"]
[      main] boot.web.embedded.tomcat.TomcatWebServer INFO  Tomcat started on port(s): 8080 (http) with context path ''
[      main] org.ojbc.mondrian.rest.Application       INFO  Started Application in 4.703 seconds (JVM running for 5.287)

Note the log message, indicated with << ***, that tells us which request authorizer is being used.

Trying to access the API without a bearer token results in a 403 (forbidden):

$ curl -X POST   http://localhost:8080/query   -H 'Content-Type: application/json'   -d '{
> "connectionName": "foodmart",
> "query" : "select { [Measures].[Org Salary] } on columns, NON EMPTY [Employee].[Employees].children on rows from HR",
>     "tidy" : {
>       "enabled" : true,
>       "simplifyNames" : true
> }
> }' -i
HTTP/1.1 403
Content-Length: 0
Date: Tue, 13 Feb 2018 21:20:53 GMT

On the server side, we also see a message in the logs:

...
[080-exec-2] jbc.mondrian.rest.MondrianRestController WARN  Authentication failed, no bearer authentication header present in request.
...

Using an unrecognized token also results in a 403:

$ curl -X POST   http://localhost:8080/query   -H 'Content-Type: application/json' -H 'Authorization: Bearer UnrecognizedToken' -d '{
> "connectionName": "foodmart",
> "query" : "select { [Measures].[Org Salary] } on columns, NON EMPTY [Employee].[Employees].children on rows from HR",
>     "tidy" : {
>       "enabled" : true,
>       "simplifyNames" : true
> }
> }' -i
HTTP/1.1 403
Content-Length: 0
Date: Tue, 13 Feb 2018 21:20:53 GMT

But with a slightly different log message:

...
[080-exec-4] jbc.mondrian.rest.MondrianRestController WARN  Authentication failed.  Token UnrecognizedToken not found in config.
...

And if we authenticate with a token that is mapped to a Mondrian role that does not authorize access to the query, a 500 status results, with an error message sent back in the body:

$ curl -X POST \
>   http://localhost:8080/query \
>   -H 'Authorization: Bearer NO_HR_CUBE_TOKEN' \
>   -H 'Content-Type: application/json' \
>   -d '{
> "connectionName": "foodmart",
> "query" : "select { [Measures].[Org Salary] } on columns, NON EMPTY [Employee].[Employees].children on rows from HR",
>     "tidy" : {
>       "enabled" : true,
>       "simplifyNames" : true
> }
> }' -i
HTTP/1.1 500
Content-Type: application/json;charset=UTF-8
Content-Length: 147
Date: Tue, 13 Feb 2018 21:23:44 GMT
Connection: close

{
  "reason" : "mondrian gave exception while parsing query",
  "rootCauseReason" : "Mondrian Error:MDX cube 'HR' not found",
  "SQLState" : null
}

And the corresponding log message:

...
[080-exec-2] jbc.mondrian.rest.MondrianRestController INFO  Token NO_HR_CUBE_TOKEN with role No HR Cube executing query on connection foodmart with tidy=true: select { [Measures].[Org Salary] } on columns, NON EMPTY [Employee].[Employees].children on rows from HR
[080-exec-2] jbc.mondrian.rest.MondrianRestController WARN  OlapException occurred processing query.  Stack trace follows (if debug logging).
[080-exec-2] jbc.mondrian.rest.MondrianRestController WARN  Exception root cause: Mondrian Error:MDX cube 'HR' not found...

(Note that the token NO_HR_CUBE_TOKEN is mapped to the "No HR Cube" Mondrian role in the bearer token authentication configuration file here).

If we authenticate with a token that is mapped to a Mondrian role that is authorized to access the query, we get a successful result:

$ curl -X POST \
>   http://localhost:8080/query \
>   -H 'Authorization: Bearer TOKEN1' \
>   -H 'Content-Type: application/json' \
>   -d '{
> "connectionName": "foodmart",
> "query" : "select { [Measures].[Org Salary] } on columns, NON EMPTY [Employee].[Employees].children on rows from HR",
>     "tidy" : {
>       "enabled" : true,
>       "simplifyNames" : true
> }
> }' -i
HTTP/1.1 200
mondrian-rest-cached-result: true
Content-Type: application/json;charset=UTF-8
Content-Length: 92
Date: Tue, 13 Feb 2018 21:29:19 GMT

{
  "values" : [ {
    "Employee Id" : "Sheri Nowmer",
    "Org Salary" : 39431.6712
  } ]
}

Note that there is a "special" Mondrian role, not defined in any schema, named org.ojbc.mondrian.rest.RequestAuthorizer-All-Access, that can be used to grant full access to a connection's schema. This is essentially a way to authenticate a user, but map that user to the default Mondrian role (which is no role), enabling full access. In the example just above, the Bearer Token TOKEN1 is mapped to this role for the FoodMart connection, which is why the request is successful.

Implementing a custom request authorizer

It is straightforward to implement new request authorizers for new authentication schemes by implementing the org.ojbc.mondrian.rest.RequestAuthorizer interface, found here. This interface contains a single method, authorizeRequest(HttpServletRequest request, QueryRequest queryRequest): RequestAuthorizationStatus. The bearer token and SAML assertion implementations should serve as good examples to follow.

Tidying

You can also include a property, named tidy, in the query request. This causes the service to return more "tabularized" data, suitable for loading into a data frame or using in a visualization tool like Vega-lite. Example:

$: curl -s -X POST -H 'Content-Type: application/json' http://localhost:58080/mondrian-rest/query -d '{
>   "connectionName" : "foodmart",
>   "query" : "select { [Measures].[Units Shipped] } on columns, NON EMPTY [Store].[Store Type].children on rows from Warehouse",
>   "tidy" : {
>     "enabled" : true
>   }
> }'
{
  "values" : [ {
    "[Store].[Store Type].[Store Type]" : "Deluxe Supermarket",
    "Units Shipped" : 64804.0
  }, {
    "[Store].[Store Type].[Store Type]" : "Gourmet Supermarket",
    "Units Shipped" : 10759.0
  }, {
    "[Store].[Store Type].[Store Type]" : "Mid-Size Grocery",
    "Units Shipped" : 10589.0
  }, {
    "[Store].[Store Type].[Store Type]" : "Small Grocery",
    "Units Shipped" : 5904.0
  }, {
    "[Store].[Store Type].[Store Type]" : "Supermarket",
    "Units Shipped" : 115670.0
  } ]
}
$:

Setting the simplifyNames property on the tidy object causes the service to remove the level's ancestors, and the brackets, from the level name:

$: curl -s -X POST -H 'Content-Type: application/json' http://localhost:58080/mondrian-rest/query -d '{
>    "connectionName" : "foodmart",
>    "query" : "select { [Measures].[Units Shipped] } on columns, NON EMPTY [Store].[Store Type].children on rows from Warehouse",
>    "tidy" : {
>      "enabled" : true,
>      "simplifyNames" : true
>    }
> }'
{
  "values" : [ {
    "Units Shipped" : 64804.0,
    "Store Type" : "Deluxe Supermarket"
  }, {
    "Units Shipped" : 10759.0,
    "Store Type" : "Gourmet Supermarket"
  }, {
    "Units Shipped" : 10589.0,
    "Store Type" : "Mid-Size Grocery"
  }, {
    "Units Shipped" : 5904.0,
    "Store Type" : "Small Grocery"
  }, {
    "Units Shipped" : 115670.0,
    "Store Type" : "Supermarket"
  } ]
}
$:

And by including a levelNameTranslationMap property, mapping level unique names to arbitrary names, it is possible to control the output even more:

$: curl -s -X POST -H 'Content-Type: application/json' http://localhost:58080/mondrian-rest/query -d '{
>     "connectionName" : "foodmart",
>     "query" : "select { [Measures].[Units Shipped] } on columns, NON EMPTY [Store].[Store Type].children on rows from Warehouse",
>     "tidy" : {
>       "enabled" : true,
>       "simplifyNames" : true,
>       "levelNameTranslationMap" : {
>         "[Store].[Store Type].[Store Type]" : "Kind of establishment"
>       }
>     }
> }'
{
  "values" : [ {
    "Units Shipped" : 64804.0,
    "Kind of establishment" : "Deluxe Supermarket"
  }, {
    "Units Shipped" : 10759.0,
    "Kind of establishment" : "Gourmet Supermarket"
  }, {
    "Units Shipped" : 10589.0,
    "Kind of establishment" : "Mid-Size Grocery"
  }, {
    "Units Shipped" : 5904.0,
    "Kind of establishment" : "Small Grocery"
  }, {
    "Units Shipped" : 115670.0,
    "Kind of establishment" : "Supermarket"
  } ]
}
$:

When specifying a levelNameTranslationMap, any unique names not found in the map are simplified as if no map were provided at all (i.e., by returning the child level's name only).

Caching

The service uses ehCache to provide on-heap caching of queries. By default, 500 queries are cached. To change caching behavior, place a file named ehcache-config.xml on the classpath.

The /flushCache operation in the API will clear all cached queries from the cache.

Building from source

To build the application with Maven, just go into the root directory (where the pom.xml is) and run:

mvn install

It will produce files mondrian-rest.war and mondrian-rest-executable.war in the target directory.

mondrian-rest's People

Contributors

aowen avatar scottcame avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

mondrian-rest's Issues

Change root of tidy json

To better accommodate vega-lite, change root of tidy json output to:

{
    "values" : [ {...}, {...}, ...]
}

Caching

Enable caching of queries and an operation to clear the cache

Mondrian security

Need to add support for assigning roles to connections, so Mondrian will respect constraints defined in schema

Simplify measures

Currently, measures are treated like any other axis position. Add an option to the /query operation to enable "collapsing" or "simplifying" measures, where all measures are included in the same cell, and change the "CellValue" property to using the name of the measure as the key.

Repeatable builds

Hello,

I'm trying to build from sources but the instructions in the Readme fail me: the pom.xml seems to be referring to a deprecated public nexus repository.

Even by following the indications on pentaho's website (https://repo.pentaho.org/), I've got the error:

[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 01:55 min
[INFO] Finished at: 2018-07-16T17:16:51+02:00
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project mondrian-rest: Could not resolve dependencies for project org.ojbc:mondrian-rest:jar:0.0.1-SNAPSHOT: Could not find artifact pentaho:mondrian:jar:4.4-SNAPSHOT in pentaho-public (http://nexus.pentaho.org/content/groups/omni) -> [Help 1]

Seems like pentaho does not provide SNAPSHOT artifacts in their repo (at least for this 4.4-SNAPSHOT version).

Also, It doesn't feel right to have a stable version number (mondrian-rest 1.3.0) depend on a snapshot dependency (mondrian 4.4-SNAPSHOT) since this means the builds of 1.3.0 will use different versions of mondrian over time.

the rest api is too slow for get connection

hi,Mr Scott:
I am a developer from China,I have two questions on below:
1.we want to import mondrian-rest to our project ,but I found the rest api is too slow,all the request entry,they all need to get connection again and again, so whether we could make a Connection Pool to save time?

2 .I see there is a dependency in pom.xml:
<dependency>
<groupId>net.hydromatic</groupId>
<artifactId>foodmart-data-hsqldb</artifactId>
<version>${foodmart-hsqldb-version}</version>
</dependency>

if I need mysql, I have to change the jar to below, I can't connect the databases without it:

<dependency>
<groupId>net.hydromatic</groupId>
<artifactId>foodmart-data-mysql</artifactId>
<version>0.10</version>
</dependency>

the question is the jar was named "foodmart-data-mysql" ,so I guess it can be only work for foodmart, so how can I get connection on my own database ? does I need to overwrite the jar? or does there have a common jar for all database for connection?

thanks for your time and God bless all Americans for healthy

Dimension label naming

Change dimension labels from using the "unique name" to just the name
Change /query operation to enable user to supply a mapping from unique name to an alternative simplified name

can not connect to mysql databases

I want to run the API based Docker, and my mondrian-connections.json as follow in docker:

{
       "AppTraffic" : {
                "JdbcDriver" : "org.mysql.jdbc.Driver",
                "Jdbc" : "jdbc:mysql://192.168.98.213/AppTraffic",
                "Catalog" : "/AppTraffic.xml",
                "Description" : "App Traffic Database",
                "IsDemo" : false
        }
}

I want know , is ondrian-rest can run with mysql ? and how i should do, thank you for your time.

error response is :

{"timestamp":1548229223099,"status":500
,"error":"Internal Server Error",
"exception":"mondrian.olap.MondrianException",
"message":"Mondrian Error:Internal error: Error while creating SQL connection: Jdbc=jdbc:mysql://192.168.98.213/AppTraffic",
"path":"/mondrian-rest/query"}

Dynamic connections

Currently, available connections have to be defined in the config file at application startup.

We need to create an API endpoint allowing the client to create a new connection dynamically. This will involve sending up the mondrian schema as well as supplying all the properties on the connection (e.g., db connect info, etc.)

Metadata caching and pre-populating

  • Cache metadata objects once created (they can't change during a controller's lifetime, at least not until we do #14)
  • Add an application property to control whether metadata are populated at controller start time

Connection to postgresql

Hi.
Is not really a issue, but a problem that i am having trying to connect to Postgresql my date base named ep.
My coonection JSON is as follow:
`{
"RnMonitorMeasurements": {
"JdbcDriver": "org.postgresql.Driver",
"Jdbc": "jdbc:postgresql://localhost/ep",
"JdbcUser": "postgres",
"JdbcPassword": "%DW_TOKEN%",
"Catalog": "/ep.xml",
"Description": "Simple test",
"IsDemo": false

}
}`
Is not working and gives me a timeout when i try to do a POST to /query

I need help this error

java.util.NoSuchElementException: Could not create a validated object, cause: enabling autocommit is not supported

I want to connect hive jdbc, but this error it makes me hard.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.