Project spotlight: PySolarized Solr library

Jernej Virag

April 13, 2014

There are numerous Python/Solr libraries out there, each having a different subset of functionality. Obviously, as per Murphy's law, none of them had a set of features I required. So I rolled my own - PySolarized!

I wrote PySolarized because I needed a Solr connector which would dispatch and query documents to multiple cores.

To use PySolarized just grab it from PyPi:

pip install pysolarized

1. Create an instance of Solr object.

For a simple single-server configuration pass server endpoint as a string

import pysolarized
solr = pysolarized.solr.Solr("http://localhost:8080/solr/core1")

If you want to use multi-language / multicore configuration pass dictionary of cores with their language identifiers:

import pysolarized
solr = pysolarized.solr.Solr({"en": "http://localhost:8080/solr/core-en", "si": "http://localhost:8080/solr/core-si"}, default_endpoint="en")

Afterwards, you use this solr instance to communicate with Solr server.

2. Query

Use query method on the created solr instance.

results = solr.query("Ljubljana", 
                    filters = {"country": "Slovenia" },
                    columns = ["id", "city_name"],
                    sort = ["city_name desc"],  
                    start = 0,
                    num_rows = 20)                  

The query method will automatically query all configured Solr cores and return aggregated results.

3. Insert documents

To send new documents to Solr server you just pass a list of documents in dictionary form to the add method:

solr.add([ { "id": "c1en", "city_name": "Vienna", "country": "Austria", "language": "en" }, 
           { "id": "c1si", "city_name": "Dunaj", "country": "Avstrija", "language": "si"}])
solr.commit()

PySolarized will use value of the language field to choose which core to send the document to - e.g. "language": "si" field in the document will cause PySolarized to send document to core with si identifier. Documents with unknown or missing language value will be sent to default endpoint. Don't forget to call commit to commit changes in documents.

PySolarized can also delete documents, clear are cores and do “More Like This” queries. Project is available on GitHub with a detailed documentation for rest of the functionality.