Skip to content

Latest commit

 

History

History
171 lines (107 loc) · 5.61 KB

README.md

File metadata and controls

171 lines (107 loc) · 5.61 KB
Raw

php-druid-query

PHP wrapper around executing HTTP requests to Druid. Usually this will be for queries, but can be used for other purposes (such as ingestion).

Overview

The wrapper lives in the namespace DruidFamiliar. Druid itself was named in the sprit of the D&D character, and that character could have a familiar - a spiritually linked animal companion. This wrapper in a sense lives as a companion to Druid, and thus the name.

I think the repo name should reflect the namespace for clarity, it currently is php-druid-query while the namespace is DruidFamiliar. This would be pretty breaking change and will be saved for the future. If you have other suggestions for naming of project or namespaces, feel free to suggest before then.

Changelog

0.2.1

  • DruidTime/Interval classes added to make it easier to work with varieties of time inputs to Druid compatible time output.

0.2.0 Major refactoring

  • Query and Response Handling separated. All interfaces renamed and redesigned.
    • IDruidConnection is now IDruidQueryExecutor.
    • IDruidQuery is split into IDruidQueryGenerator and IDruidQueryParameters and IDruidQueryResponseHandler.
    • BaseQuery is no longer needed, many similar classes were deprecated or removed.
    • DruidNodeConnection is now DruidNodeDruidQueryExecutor.

0.1.0 Initial release

  • Quick sketch for sharing early.

Typical Use

In general, this wrapper's purpose is to streamline the execution of queries by encapsulating the cruft from the HTTP nature of Druid and the analytical grammar in query configuration.

  1. Instantiate a connection, configured to hit a Druid endpoint.
  2. Instantiate a query generator object for the desired query.
  3. Instantiate a query parameters object, configured with desired query parameters.
  4. Instantiate a result handler to format the results (otherwise use DoNothingResponseHandler)
  5. Combine the connection, query, parameters, and response handler to execute it, getting the result from the result handler.

Interface wise, this looks like:

  1. Instantiate a IDruidQueryExecutor, configured to hit a Druid endpoint.
  2. Instantiate a IDruidQueryGenerator.
  3. Instantiate a IDruidQueryParameters, configured with parameters.
  4. Instantiate a IDruidQueryResponseHandler.
  5. Run the IDruidQueryExecutor's executeQuery function with the IDruidQueryGenerator, IDruidQueryParameters, and the IDruidQueryResponseHandler, getting the result from the IDruidQueryResponseHandler.

Implementation wise, this can look like:

  1. Instantiate a DruidNodeDruidQueryExecutor, configured to hit a Druid endpoint.
  2. Instantiate a SegmentMetadataDruidQuery.
  3. Instantiate a SegmentMetadataDruidQueryParameters, configured with parameters.
  4. Instantiate a SegmentMetadataResponseHandler.
  5. Run the DruidNodeDruidQueryExecutor's executeQuery function with the classes spawned in the previous steps, getting the result from SegmentMetadataResponseHandler.

How to Install

Right now, there is no tagged version. To be ready for it when it comes, branch-aliases are in place.

  • Stable branch: ~1.0@dev
  • Cutting edge: ~1.1@dev

To install, it is suggested to use Composer. If you have it installed, then the following instructions in a composer.json should be all you need to get started:

{
    "require": {
        "r4j4h/php-druid-query": "~1.0@dev"
    }
}

Once that is in, composer install and composer update should work.

Once those are run, require Composer's autoloader and you are off to the races, or tree circles as it were (bad Druid reference):

  1. require 'vendor/autoload.php';
  2. $yay = new \DruidFamiliar\Query\TimeBoundaryDruidQuery('my-cool-data-source');
  3. Refer to the Typical Use section above.

How to Test

Once composer install has finished, from the root directory in a command terminal run:

vendor/bin/phing

or

vendor/bin/phpunit tests

Generate Documentation

From the root directory, in a command terminal run: php vendor/bin/phing docs.

Examples

Examples are located in the \examples folder.

They share connection information from _examples-config.php. Change that match your Druid instance's connection info.

Right now most are designed to run via the CLI, but will work in a browser if a web server running php serves them.

The HTML outputting ones should print the query results to HTML:

Example HTML TimeBoundary Output

How it Works & How to Extend

Please refer to this diagram for an overview of how this works underneath the hood.

Sequence Diagram

(From this Dynamic LucidChart Source URL)

In general, to add support for a new query all you need to do is create a new class wherever you want that implements IDruidQuery.

By wherever you want, that could be in a fork of this repo, or outside of this repo using this repo's interfaces. That is up to you. :)

References

Appendix A. Composer.json example that does not rely on Packagist.org:

{
    "repositories": [
        {
            "type": "vcs",
            "url": "[email protected]:r4j4h/php-druid-query"
        }
    ],
    "require": {
        "r4j4h/php-druid-query": "~1.0@dev"
    }
}