$
docker run -d --name redis-stack -p 6379:6379 -p 8001:8001 redis/redis-stack:latest

$
!pip install redisvl==0.1.0
$
!pip install cohere==4.45
$
!pip install jsonlines

1
from redis import Redis
2
from redisvl.index import SearchIndex
3
from redisvl.schema import IndexSchema
4
from redisvl.utils.vectorize import CohereTextVectorizer
5
from redisvl.query import VectorQuery
6
from redisvl.query.filter import Tag, Text, Num
7
import jsonlines

1
version: "0.1.0"
2
index:
3
  name: semantic_search_demo
4
  prefix: rvl
5
  storage_type: hash
6
7
fields:
8
  - name: url
9
    type: text
10
  - name: title
11
    type: tag
12
  - name: text
13
    type: text
14
  - name: wiki_id
15
    type: numeric
16
  - name: paragraph_id
17
    type: numeric
18
  - name: id
19
    type: numeric
20
  - name: views
21
    type: numeric
22
  - name: langs
23
    type: numeric
24
  - name: embedding
25
    type: vector
26
    attrs:
27
      algorithm: flat
28
      dims: 1024
29
      distance_metric: cosine
30
      datatype: float32

1
# create a vectorizer
2
api_key = "{Insert your cohere API Key}"
3
4
cohere_vectorizer = CohereTextVectorizer(
5
    model="embed-english-v3.0",
6
    api_config={"api_key": api_key},
7
)

1
# construct a search index from the schema - this schema is called "semantic_search_demo"
2
schema = IndexSchema.from_yaml("./schema.yaml")
3
client = Redis.from_url("redis://localhost:6379")
4
index = SearchIndex(schema, client)
5
6
# create the index (no data yet)
7
index.create(overwrite=True)

1
!rvl index listall

15:39:22 [RedisVL] INFO   Indices:
15:39:22 [RedisVL] INFO   1. semantic_search_demo

1
!rvl index info -i semantic_search_demo

1
# read in your documents
2
jsonl_file_path = "data/redis_guide_data.jsonl"
3
4
corpus = []
5
text_to_embed = []
6
7
with jsonlines.open(jsonl_file_path, mode="r") as reader:
8
    for line in reader:
9
        corpus.append(line)
10
        # we want to store the embeddings of the field called `text`
11
        text_to_embed.append(line["text"])
12
13
# call embed_many which returns an array
14
# hash data structures get serialized as a string and thus we store the embeddings in hashes as a byte string (handled by numpy)
15
res = cohere_vectorizer.embed_many(
16
    text_to_embed, input_type="search_document", as_buffer=True
17
)

1
# contruct the data payload to be uploaded to your index
2
data = [
3
    {
4
        "url": row["url"],
5
        "title": row["title"],
6
        "text": row["text"],
7
        "wiki_id": row["wiki_id"],
8
        "paragraph_id": row["paragraph_id"],
9
        "id": row["id"],
10
        "views": row["views"],
11
        "langs": row["langs"],
12
        "embedding": v,
13
    }
14
    for row, v in zip(corpus, res)
15
]
16
17
# load the data into your index
18
index.load(data)

1
# use the Cohere vectorizer again to create a query embedding
2
query_embedding = cohere_vectorizer.embed(
3
    "What did Microsoft release in 2015?",
4
    input_type="search_query",
5
    as_buffer=True,
6
)
7
8
9
query = VectorQuery(
10
    vector=query_embedding,
11
    vector_field_name="embedding",
12
    return_fields=[
13
        "url",
14
        "wiki_id",
15
        "paragraph_id",
16
        "id",
17
        "views",
18
        "langs",
19
        "title",
20
        "text",
21
    ],
22
    num_results=5,
23
)
24
25
results = index.query(query)
26
27
for doc in results:
28
    print(
29
        f"Title:{doc['title']}\nText:{doc['text']}\nDistance {doc['vector_distance']}\n\n"
30
    )

1
# Initialize a tag filter
2
tag_filter = Tag("title") == "Microsoft Office"
3
4
# set the tag filter on our existing query
5
query.set_filter(tag_filter)
6
7
results = index.query(query)
8
9
for doc in results:
10
    print(
11
        f"Title:{doc['title']}\nText:{doc['text']}\nDistance {doc['vector_distance']}\n"
12
    )

1
# define a tag match on the title, text match on the text field, and numeric filter on the views field
2
filter_data = (
3
    (Tag("title") == "Elizabeth II")
4
    & (Text("text") % "born")
5
    & (Num("views") > 4500)
6
)
7
8
query_embedding = co.embed(
9
    "When was she born?", input_type="search_query", as_buffer=True
10
)
11
12
# reinitialize the query with the filter expression
13
query = VectorQuery(
14
    vector=query_embedding,
15
    vector_field_name="embedding",
16
    return_fields=[
17
        "url",
18
        "wiki_id",
19
        "paragraph_id",
20
        "id",
21
        "views",
22
        "langs",
23
        "title",
24
        "text",
25
    ],
26
    num_results=5,
27
    filter_expression=filter_data,
28
)
29
30
results = index.query(query)
31
print(results)
32
33
for doc in results:
34
    print(
35
        f"Title:{doc['title']}\nText:{doc['text']}\nDistance {doc['vector_distance']}\nView {doc['views']}"
36
    )