AsyncIO Python client for Apache Solr

This project requires Python 3.7+

To install with pip

pip install aiosolr

The connection to the Solr backend is defined during object initialization. The accepted kwargs to init are scheme, host, port, and collection.

collection may optionally be passed at query time

import aiosolr

client = aiosolr.Client(host=localhost, collection="example", port=8983)

Alternatively you may instantiate via passing connection_url like:

import aiosolr

client = aiosolr.Client(connection_url="http://host:1234/path/to/solr/collection")

Once you have your aiosolr.Client instance, set up the session:

await client.setup()

There are methods available for querying. You can use Solr's built-in get handler with the get method to retrieve a single document:

await client.get(document_id)

You can use a pre-defined suggestions handler by using the suggestions method:

await client.suggestions("suggest_handler", query="asdf")

You can also use the suggestions method to build your suggestions:

await client.suggestions("suggest_handler", build=True)

handler is a required argument for suggestions unlike for get or query

You can use the query method to query your search handler. The default handler used is select. If you would like spellcheck suggestion turned on, pass spellcheck=True (default is False).

await client.query(handler="my_handler", query="asdf", spellcheck=True)

If spellcheck is True the query method returns a tuple with the first element being an array of documents and the 2nd element being an array of spellcheck suggestions. Otherwise, the query method returns a simple array of documents.

You can use the update method to access Solr's built-in update handler like:

await client.update(my_data)

At any point that you need to commit data to your collection you can use the commit method. Arguments should be the handler (update by default) and soft as a boolean indicating whether it should be a hard or soft commit (defaults to False).

There is one more method you might want to use before querying Solr especially if the query is coming from an untrusted end user. There is a clean_query method which can be used to strip out unwanted characters. Use it like:

trusted_query = aiosolr.clean_query(users_query)

Once you are finished with the Solr instance, you should call the method close to cleanup sessions like:

await client.close()

You can initialize the client with read_timeout and write_timeout to limit how long to wait for requests to complete. read_timeout applies to get and query whereas write_timeout applies to update:

import aiosolr

client = aiosolr.Client(connection_url=connection_url, read_timeout=5, write_timeout=30)

You can override the timeouts for a specific request:

await client.get(document_id, read_timeout=1)  # I'm in a hurry
await client.update(doc, write_timeout=60)  # this is a large request so we expect it to take a long time

aiosolr uses asyncio.wait_for internally, so if a timeout occurs the exception raised is asyncio.TimeoutError.

To get more information from the Client you can initialize with debug=True:

    import aiosolr

    client = aiosolr.Client(host=localhost, collection="example", port=8983, debug=True)

This sets the aiosolr logger to DEBUG level, and also sets the internally used HTTP session (provided by aiohttp) to the DEBUG level. This makes it easier to see the actual network requests going to Solr.