Redisearch Python simple example

Question

The documentation for Redisearch module is not the best at the moment. Can someone please give short examples that will answer these questions? Thank you

1. How to use Redisearch module with redis-py?
2. How to sort by multiple fields?
3. How to paginate results?

Answer 1

Posting the answer. I spent quite some time to figure all these out, so hopefully it will save time for some of you. Redisearch is truly an amazing technology that deserves to be used by many.

First and foremost you have to create an index to use redisearch:

from redis.commands.search.query import Query
from redis.commands.search.field import TextField, NumericField
from redis.commands.search.indexDefinition import IndexDefinition, IndexType
from redis.commands.search.aggregation import AggregateRequest, Asc, Desc

import redis

client = redis.Redis(
        host=redis_host,
        port=redis_port,
        password=redis_password,
        decode_responses=True,
    )

index_name = 'idx:flowers'
index = client.ft(index_name)

#  check if index exists if not then create
try:
    index.info()
except:
    schema = (
        TextField('name', weight=0.7),  #  textfields can also be sortable
        TextField('type'),
        NumericField('length', sortable=True),  #  numericfields cannot have weights
        NumericField('rating', sortable=True),
    )
    #  index creation. 
    #  using IndexDefinition(prefix= will index automagically all redis keys that have this prefix
    client.ft(index_name).create_index(schema, definition=IndexDefinition(prefix=['flower:']))

    #  since the index is new we need to load the data (if it does not exist on redis already)

    flowers = [
        {'name': 'gladiola', 'type': 'cactus', 'length': 22, 'rating': 3.57, 'color': 'violet'},
        {'name': 'eleanor', 'type': 'cactus,pine', 'length': 15, 'rating': 4.34, 'color': 'brown'},
        {'name': 'loveme', 'type': 'berry', 'length': 13, 'rating': 5, 'color': 'blue'},
    ]

    for i, flower in enumerate(flowers):
        pipe = client.pipeline()
        pipe.hset(name='flower:' + str(i), key='name', value=flower['name'])
        pipe.hset(name='flower:' + str(i), key='type', value=flower['type'])
        pipe.hset(name='flower:' + str(i), key='length', value=flower['length'])
        pipe.hset(name='flower:' + str(i), key='rating', value=flower['rating'])
        pipe.hset(name='flower:' + str(i), key='color', value=flower['color'])
        pipe.execute()

all these hashes will be indexed because their key has the prefix 'flower:'

searching:

#  first we need to get the index we are going to work with
index = client.ft(index_name)

#  wildcard search from both sides on all searchable fields
results = index.search(Query('*actu*'))  
results = results.docs
for result in results:
    print(result.__dict__)

#  sort by a sortable field in ascending order (you can sort by a non-sortable field but it will be slower)
results = index.search(Query('*actu*').sort_by(field='length', asc=True))

#  pagination. 'num' is size of the page
results = index.search(Query('*actu*').paging(offset=0, num=10))

#  search a specific field
results = index.search(Query('@name:*actu*'))

#  search specific fields - OR operation
results = index.search(Query('@name|type:*actu*'))

#  search specific fields - AND operation
results = index.search(Query('@name:*actu* @type:*actu*'))

#  sort by multiple fields - length and rating. We need to use Aggregation to achieve this
request = AggregateRequest(
    query='@name|type:*actu*'
    ).sort_by(
        Asc('@length'), 
        Desc('@rating'), 
    ).limit(  #  pagination
        0, 10
    ).load(  #  we need to indicate the fields we want to get in results
        '@name', 
        '@type', 
        '@length', 
        '@rating', 
        '@color', 
    )
#  perform aggregate search
results = index.aggregate(request)
print(results.rows)

Point out mistakes and improvements, I'll gladly edit the answer.