|
1 |
| -Using asyncio with Elasticsearch |
2 |
| -================================ |
| 1 | +Async Elasticsearch API |
| 2 | +======================= |
3 | 3 |
|
4 | 4 | .. py:module:: elasticsearch
|
5 | 5 | :no-index:
|
6 | 6 |
|
7 |
| -The ``elasticsearch`` package supports async/await with |
8 |
| -`asyncio <https://docs.python.org/3/library/asyncio.html>`_ and `aiohttp <https://docs.aiohttp.org>`_. |
9 |
| -You can either install ``aiohttp`` directly or use the ``[async]`` extra: |
10 |
| - |
11 |
| - .. code-block:: bash |
12 |
| -
|
13 |
| - $ python -m pip install elasticsearch aiohttp |
14 |
| -
|
15 |
| - # - OR - |
16 |
| -
|
17 |
| - $ python -m pip install elasticsearch[async] |
18 |
| -
|
19 |
| -Getting Started with Async |
20 |
| --------------------------- |
21 |
| - |
22 |
| -After installation all async API endpoints are available via :class:`~elasticsearch.AsyncElasticsearch` |
23 |
| -and are used in the same way as other APIs, just with an extra ``await``: |
24 |
| - |
25 |
| - .. code-block:: python |
26 |
| -
|
27 |
| - import asyncio |
28 |
| - from elasticsearch import AsyncElasticsearch |
29 |
| -
|
30 |
| - client = AsyncElasticsearch() |
31 |
| -
|
32 |
| - async def main(): |
33 |
| - resp = await client.search( |
34 |
| - index="documents", |
35 |
| - body={"query": {"match_all": {}}}, |
36 |
| - size=20, |
37 |
| - ) |
38 |
| - print(resp) |
39 |
| -
|
40 |
| - loop = asyncio.get_event_loop() |
41 |
| - loop.run_until_complete(main()) |
42 |
| -
|
43 |
| -All APIs that are available under the sync client are also available under the async client. |
44 |
| - |
45 |
| -ASGI Applications and Elastic APM |
46 |
| ---------------------------------- |
47 |
| - |
48 |
| -`ASGI <https://asgi.readthedocs.io>`_ (Asynchronous Server Gateway Interface) is a new way to |
49 |
| -serve Python web applications making use of async I/O to achieve better performance. |
50 |
| -Some examples of ASGI frameworks include FastAPI, Django 3.0+, and Starlette. |
51 |
| -If you're using one of these frameworks along with Elasticsearch then you |
52 |
| -should be using :py:class:`~elasticsearch.AsyncElasticsearch` to avoid blocking |
53 |
| -the event loop with synchronous network calls for optimal performance. |
54 |
| - |
55 |
| -`Elastic APM <https://www.elastic.co/guide/en/apm/agent/python/current/index.html>`_ |
56 |
| -also supports tracing of async Elasticsearch queries just the same as |
57 |
| -synchronous queries. For an example on how to configure ``AsyncElasticsearch`` with |
58 |
| -a popular ASGI framework `FastAPI <https://fastapi.tiangolo.com/>`_ and APM tracing |
59 |
| -there is a `pre-built example <https://github.com/elastic/elasticsearch-py/tree/master/examples/fastapi-apm>`_ |
60 |
| -in the ``examples/fastapi-apm`` directory. |
61 |
| - |
62 |
| -Frequently Asked Questions |
63 |
| --------------------------- |
64 |
| - |
65 |
| -ValueError when initializing ``AsyncElasticsearch``? |
66 |
| -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
67 |
| - |
68 |
| -If when trying to use ``AsyncElasticsearch`` you receive ``ValueError: You must |
69 |
| -have 'aiohttp' installed to use AiohttpHttpNode`` you should ensure that you |
70 |
| -have ``aiohttp`` installed in your environment (check with ``$ python -m pip |
71 |
| -freeze | grep aiohttp``). Otherwise, async support won't be available. |
72 |
| - |
73 |
| -What about the ``elasticsearch-async`` package? |
74 |
| -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
75 |
| - |
76 |
| -Previously asyncio was supported separately via the `elasticsearch-async <https://github.com/elastic/elasticsearch-py-async>`_ |
77 |
| -package. The ``elasticsearch-async`` package has been deprecated in favor of |
78 |
| -``AsyncElasticsearch`` provided by the ``elasticsearch`` package |
79 |
| -in v7.8 and onwards. |
80 |
| - |
81 |
| -Receiving 'Unclosed client session / connector' warning? |
82 |
| -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
83 |
| - |
84 |
| -This warning is created by ``aiohttp`` when an open HTTP connection is |
85 |
| -garbage collected. You'll typically run into this when closing your application. |
86 |
| -To resolve the issue ensure that :meth:`~elasticsearch.AsyncElasticsearch.close` |
87 |
| -is called before the :py:class:`~elasticsearch.AsyncElasticsearch` instance is garbage collected. |
88 |
| - |
89 |
| -For example if using FastAPI that might look like this: |
90 |
| - |
91 |
| - .. code-block:: python |
92 |
| -
|
93 |
| - import os |
94 |
| - from contextlib import asynccontextmanager |
95 |
| -
|
96 |
| - from fastapi import FastAPI |
97 |
| - from elasticsearch import AsyncElasticsearch |
98 |
| -
|
99 |
| - ELASTICSEARCH_URL = os.environ["ELASTICSEARCH_URL"] |
100 |
| - client = None |
101 |
| -
|
102 |
| - @asynccontextmanager |
103 |
| - async def lifespan(app: FastAPI): |
104 |
| - global client |
105 |
| - client = AsyncElasticsearch(ELASTICSEARCH_URL) |
106 |
| - yield |
107 |
| - await client.close() |
108 |
| -
|
109 |
| - app = FastAPI(lifespan=lifespan) |
110 |
| -
|
111 |
| - @app.get("/") |
112 |
| - async def main(): |
113 |
| - return await client.info() |
114 |
| -
|
115 |
| -You can run this example by saving it to ``main.py`` and executing |
116 |
| -``ELASTICSEARCH_URL=http://localhost:9200 uvicorn main:app``. |
117 |
| - |
118 |
| - |
119 |
| -Async Helpers |
120 |
| -------------- |
121 |
| - |
122 |
| -Async variants of all helpers are available in ``elasticsearch.helpers`` |
123 |
| -and are all prefixed with ``async_*``. You'll notice that these APIs |
124 |
| -are identical to the ones in the sync :ref:`helpers` documentation. |
125 |
| - |
126 |
| -All async helpers that accept an iterator or generator also accept async iterators |
127 |
| -and async generators. |
128 |
| - |
129 |
| - .. py:module:: elasticsearch.helpers |
130 |
| - :no-index: |
131 |
| - |
132 |
| -Bulk and Streaming Bulk |
133 |
| -~~~~~~~~~~~~~~~~~~~~~~~ |
134 |
| - |
135 |
| - .. autofunction:: async_bulk |
136 |
| - |
137 |
| - .. code-block:: python |
138 |
| -
|
139 |
| - import asyncio |
140 |
| - from elasticsearch import AsyncElasticsearch |
141 |
| - from elasticsearch.helpers import async_bulk |
142 |
| -
|
143 |
| - client = AsyncElasticsearch() |
144 |
| -
|
145 |
| - async def gendata(): |
146 |
| - mywords = ['foo', 'bar', 'baz'] |
147 |
| - for word in mywords: |
148 |
| - yield { |
149 |
| - "_index": "mywords", |
150 |
| - "doc": {"word": word}, |
151 |
| - } |
152 |
| -
|
153 |
| - async def main(): |
154 |
| - await async_bulk(client, gendata()) |
155 |
| -
|
156 |
| - loop = asyncio.get_event_loop() |
157 |
| - loop.run_until_complete(main()) |
158 |
| -
|
159 |
| - .. autofunction:: async_streaming_bulk |
160 |
| - |
161 |
| - .. code-block:: python |
162 |
| -
|
163 |
| - import asyncio |
164 |
| - from elasticsearch import AsyncElasticsearch |
165 |
| - from elasticsearch.helpers import async_streaming_bulk |
166 |
| -
|
167 |
| - client = AsyncElasticsearch() |
168 |
| -
|
169 |
| - async def gendata(): |
170 |
| - mywords = ['foo', 'bar', 'baz'] |
171 |
| - for word in mywords: |
172 |
| - yield { |
173 |
| - "_index": "mywords", |
174 |
| - "word": word, |
175 |
| - } |
176 |
| -
|
177 |
| - async def main(): |
178 |
| - async for ok, result in async_streaming_bulk(client, gendata()): |
179 |
| - action, result = result.popitem() |
180 |
| - if not ok: |
181 |
| - print("failed to %s document %s" % ()) |
182 |
| -
|
183 |
| - loop = asyncio.get_event_loop() |
184 |
| - loop.run_until_complete(main()) |
185 |
| -
|
186 |
| -Scan |
187 |
| -~~~~ |
188 |
| - |
189 |
| - .. autofunction:: async_scan |
190 |
| - |
191 |
| - .. code-block:: python |
192 |
| -
|
193 |
| - import asyncio |
194 |
| - from elasticsearch import AsyncElasticsearch |
195 |
| - from elasticsearch.helpers import async_scan |
196 |
| -
|
197 |
| - client = AsyncElasticsearch() |
198 |
| -
|
199 |
| - async def main(): |
200 |
| - async for doc in async_scan( |
201 |
| - client=client, |
202 |
| - query={"query": {"match": {"title": "python"}}}, |
203 |
| - index="orders-*" |
204 |
| - ): |
205 |
| - print(doc) |
206 |
| -
|
207 |
| - loop = asyncio.get_event_loop() |
208 |
| - loop.run_until_complete(main()) |
209 |
| -
|
210 |
| -Reindex |
211 |
| -~~~~~~~ |
212 |
| - |
213 |
| - .. autofunction:: async_reindex |
214 |
| - |
215 |
| - |
216 |
| -API Reference |
217 |
| -------------- |
218 |
| - |
219 |
| - .. py:module:: elasticsearch |
220 |
| - :no-index: |
221 |
| - |
222 |
| -The API of :class:`~elasticsearch.AsyncElasticsearch` is nearly identical |
223 |
| -to the API of :class:`~elasticsearch.Elasticsearch` with the exception that |
224 |
| -every API call like :py:func:`~elasticsearch.AsyncElasticsearch.search` is |
225 |
| -an ``async`` function and requires an ``await`` to properly return the response |
226 |
| -body. |
227 |
| - |
228 |
| -AsyncElasticsearch |
229 |
| -~~~~~~~~~~~~~~~~~~ |
230 |
| - |
231 | 7 | .. note::
|
232 | 8 |
|
233 | 9 | To reference Elasticsearch APIs that are namespaced like ``.indices.create()``
|
234 | 10 | refer to the sync API reference. These APIs are identical between sync and async.
|
235 | 11 |
|
| 12 | +Elasticsearch |
| 13 | +------------- |
236 | 14 | .. autoclass:: AsyncElasticsearch
|
237 | 15 | :members:
|
0 commit comments