From the 11.-13. August we had a fun, inspiring and productive codesprint at the dkd office in Frankfurt.
At the beginning we started by talking about expectations, ideas and formed teams for the tasks.
What we did
After we formed small working groups, the groups decided on which task they want to work. In the following paragraphs I will summarize the results that we achieve.
Prepare the migration to ApacheSolr 6.1
During the migration to Apache Solr 6.1 we solved the following steps:
Switch to Apache Solr „configSets“ to simplify the installation. A configSet is now one single unit with:
- All schemata
- The solr config.xml
- All required custom jar files (access filter) to run EXT:solr
Simplify the old installer
We decided to replace the old installer with a new script, that should be used for development use only.
- On demand downloads of language files and jar files have been removed and are now shipped with the configSet
- The installer can be executed without sudo writes, as long as you have write permissions in the installation folder.
Prepared Docker containers
Based on the official Apache Solr Docker container from Lucidworks we created two docker images:
- BaseImage: Ships a configured docker container with all configSets, but without any default cores. This container can be used to spin up your own container and create your cores afterwards with a REST call
- FullImage: Ships a configured container, preconfigured with a core for each supported language. This is the easy way to create your solr instance ready with all languages.
Make EXT:solr ready for ApacheSolr 6.1
During the development our integration test indicated that there is a different behaviour in Apache Solr 6.1 compared to 4.10.4.
- The format of the spellcheck response was changed and adapted to the latest version
- FieldCollapsing only supports string and numeric fields (no text fields). This was documented and adapted in the tests.
- The deprecated components in the schemata have been replaced by theire new implementation. The help of our japanese native speakers was usefull here :)
Get value from the statistic recorded by EXT:solr
In EXT:solr you are able to track statistics from the searches that are done in the frontend for a couple of time. Now you can get a value out of this in the backend and see:
- Most frequent searches together with the number of hits
- Search terms, that are used in the fronted but do not show any hits.
This can help to optimize your configuration and content according to the interests of your website users
During the codesprint we:
- Added missing SVG icons
- Adjusted CSS to optimize the backend layout
Beside the technical stuff we had a great and inspiring time and took the change to learn more about each other.
And what do 6 hungry developers need after coding?
That’s why we decided to take our dinner on friday at the „Blockhouse Steakhouse“ and afterwards enjoy the music and conversations in the „O Reilly`s Irish Pub“...
How to get it?
With the release of Version 5.1 we will provide the "new backend statistics module" and the „ux improvements“. Beside that, version 5.1, will contain the possibility to configure „document variants“ aka „field collapsing“, which can be used as one approach of de-duplication for solrfal or to render variants of documents.
The integration of Apache Solr 6.1 and Docker images is planned for Version EXT:solr 6.0 that is planned to be released by the end of 2016. You can try it right now, by installing our „solr6-docker“ branch as "technology preview“. We are happy to get your feedback on the current state.
TYPO3 Version: 7.6 LTS
Apache Solr Version: 4.10.4
Estimated Release: 9/2016
Focus: Codesprint Results / Document Variants / UX Improvements
TYPO3 Version: 7.6 LTS
Apache Solr Version: 6.1.0
Estimated Release: 12/2016
Focus: New Solr Version / Docker Images
TYPO3 Version: 8.2 LTS
Apache Solr Version: 6.1.0
Estimated Release: 3-4/2016 (Depends on LTS Release)
Focus: Fluid as Default Templating Engine in EXT:solr
First of all, we had a great team and i want to thank all participants that you’ve joined the codesprint and really worked hard to bring "Apache Solr For TYPO3" to a new level:
- Ingo Renner
- Markus Friedrich
- Patrick Oberdorf
- Thomas Hohn
- Timo Hund
- Olivier Dobberkau
- Thomas Janke
In addition we want to say thank you to:
- Raphael GmbH: For sending Patrick Oberdorf to support us
- Systime.dk: For sending Thomas Hohn to support us
- Jochend Weiland: For sponsoring a lunch!
- Phorax: For sponsoring a lunch as well!
- dkd Internet Service GmbH: For hosting the sprint and sponsoring hotel, dinner, drinks, coffee and lunches!
- Ingo Renner for the Donuts!
- All our EB Partners: For investing into Apache Solr for TYPO3!
Looking forward for the next codesprint! Maybe you want to join as well?
We are hiring! https://www.dkd.de/de/karriere/typo3-senior-entwickler/