1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
|
# Licensed to Elasticsearch B.V. under one or more contributor
# license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright
# ownership. Elasticsearch B.V. licenses this file to you under
# the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
"""
# Semantic Text example
Requirements:
$ pip install "elasticsearch[async]" tqdm
Before running this example, an ELSER inference endpoint must be created in the
Elasticsearch cluster. This can be done manually from Kibana, or with the
following curl command from a terminal:
curl -X PUT \
"$ELASTICSEARCH_URL/_inference/sparse_embedding/my-elser-endpoint" \
-H "Content-Type: application/json" \
-d '{"service":"elser","service_settings":{"num_allocations":1,"num_threads":1}}'
To run the example:
$ python semantic_text.py "text to search"
The index will be created automatically if it does not exist. Add
`--recreate-index` to the command to regenerate it.
The example dataset includes a selection of workplace documents. The
following are good example queries to try out with this dataset:
$ python semantic_text.py "work from home"
$ python semantic_text.py "vacation time"
$ python semantic_text.py "can I bring a bird to work?"
When the index is created, the inference service will split the documents into
short passages, and for each passage a sparse embedding will be generated using
Elastic's ELSER v2 model.
"""
import argparse
import asyncio
import json
import os
from datetime import datetime
from typing import Any, Optional
from urllib.request import urlopen
from tqdm import tqdm
from elasticsearch import dsl
DATASET_URL = "https://raw.githubusercontent.com/elastic/elasticsearch-labs/main/datasets/workplace-documents.json"
class WorkplaceDoc(dsl.AsyncDocument):
class Index:
name = "workplace_documents_semantic"
name: str
summary: str
content: Any = dsl.mapped_field(
dsl.field.SemanticText(inference_id="my-elser-endpoint")
)
created: datetime
updated: Optional[datetime]
url: str = dsl.mapped_field(dsl.Keyword())
category: str = dsl.mapped_field(dsl.Keyword())
async def create() -> None:
# create the index
await WorkplaceDoc._index.delete(ignore_unavailable=True)
await WorkplaceDoc.init()
# download the data
dataset = json.loads(urlopen(DATASET_URL).read())
# import the dataset
for data in tqdm(dataset, desc="Indexing documents..."):
doc = WorkplaceDoc(
name=data["name"],
summary=data["summary"],
content=data["content"],
created=data.get("created_on"),
updated=data.get("updated_at"),
url=data["url"],
category=data["category"],
)
await doc.save()
# refresh the index
await WorkplaceDoc._index.refresh()
async def search(query: str) -> dsl.AsyncSearch[WorkplaceDoc]:
search = WorkplaceDoc.search()
search = search[:5]
return search.query(dsl.query.Semantic(field=WorkplaceDoc.content, query=query))
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(description="Vector database with Elasticsearch")
parser.add_argument(
"--recreate-index", action="store_true", help="Recreate and populate the index"
)
parser.add_argument("query", action="store", help="The search query")
return parser.parse_args()
async def main() -> None:
args = parse_args()
# initiate the default connection to elasticsearch
dsl.async_connections.create_connection(hosts=[os.environ["ELASTICSEARCH_URL"]])
if args.recreate_index or not await WorkplaceDoc._index.exists():
await create()
results = await search(args.query)
async for hit in results:
print(
f"Document: {hit.name} [Category: {hit.category}] [Score: {hit.meta.score}]"
)
print(f"Content: {hit.content.text}")
print("--------------------\n")
# close the connection
await dsl.async_connections.get_connection().close()
if __name__ == "__main__":
asyncio.run(main())
|
