Elasticsearch doesn’t offer a lemmatizer for following languages out of the box:

• Bulgarian,
• Czech,
• Estonian,
• French,
• Hungarian,
• Macedonian,
• Persian,
• Polish,
• Romanian,
• Russian,
• Slovak,
• Slovene,
• Serbian,
• Ukrainian.

There is a great plugin LemmaGen that solves this shortcoming. At the time of writing LemmaGen works with ElasticSearch 2.2.0 and older.

We show a simple python example, which connects to Elasticsearch server, initializes index and mapping, adds documents in Slovenian language and executes a search. We add documents with following titles: pes, psa, psi, pse, psovanje, pesem, pesmi, where first 4 titles are about dogs and last 3 have the same first letters, but different meaning. We show how to execute a search with a query pes (a dog) and retrieve only search results about dogs. Further reading: Efficient search in your local language

1. Download elasticsearch, extract the zip and move the elasticsearch directory to some path.
2. Go to that path and install Lemmagen plugin: ./bin/plugin install https://github.com/vhyza/elasticsearch-analysis-lemmagen/releases/download/v2.2.0/elasticsearch-analysis-lemmagen-2.2.0-plugin.zip
3. Download this project and install requirements with: pip install -r requirements.txt

Run bin/elasticsearch to start elasticsearch server. Run python main.py to execute the search.