Skip to content

Commit 4adebcf

Browse files
davidkyleJoanFM
davidkyle
and
JoanFM
authored
[docs] Add documentation for JinaAI service (elastic#118782) (elastic#119693)
# Conflicts: # docs/reference/inference/put-inference.asciidoc Co-authored-by: Joan Fontanals <[email protected]>
1 parent 5952eb8 commit 4adebcf

File tree

3 files changed

+257
-0
lines changed

3 files changed

+257
-0
lines changed

‎docs/reference/inference/inference-apis.asciidoc‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -149,6 +149,7 @@ include::service-elser.asciidoc[]
149149
include::service-google-ai-studio.asciidoc[]
150150
include::service-google-vertex-ai.asciidoc[]
151151
include::service-hugging-face.asciidoc[]
152+
include::service-jinaai.asciidoc[]
152153
include::service-mistral.asciidoc[]
153154
include::service-openai.asciidoc[]
154155
include::service-watsonx-ai.asciidoc[]

‎docs/reference/inference/put-inference.asciidoc‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -78,6 +78,7 @@ Click the links to review the configuration details of the services:
7878
* <<infer-service-mistral,Mistral>> (`text_embedding`)
7979
* <<infer-service-openai,OpenAI>> (`completion`, `text_embedding`)
8080
* <<infer-service-watsonx-ai>> (`text_embedding`)
81+
* <<infer-service-jinaai,JinaAI>> (`text_embedding`, `rerank`)
8182

8283
The {es} and ELSER services run on a {ml} node in your {es} cluster. The rest of
8384
the services connect to external providers.

‎docs/reference/inference/service-jinaai.asciidoc‎

Lines changed: 255 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,255 @@
1+
[[infer-service-jinaai]]
2+
=== JinaAI {infer} service
3+
4+
Creates an {infer} endpoint to perform an {infer} task with the `jinaai` service.
5+
6+
7+
[discrete]
8+
[[infer-service-jinaai-api-request]]
9+
==== {api-request-title}
10+
11+
`PUT /_inference/<task_type>/<inference_id>`
12+
13+
[discrete]
14+
[[infer-service-jinaai-api-path-params]]
15+
==== {api-path-parms-title}
16+
17+
`<inference_id>`::
18+
(Required, string)
19+
include::inference-shared.asciidoc[tag=inference-id]
20+
21+
`<task_type>`::
22+
(Required, string)
23+
include::inference-shared.asciidoc[tag=task-type]
24+
+
25+
--
26+
Available task types:
27+
28+
* `text_embedding`,
29+
* `rerank`.
30+
--
31+
32+
[discrete]
33+
[[infer-service-jinaai-api-request-body]]
34+
==== {api-request-body-title}
35+
36+
`chunking_settings`::
37+
(Optional, object)
38+
include::inference-shared.asciidoc[tag=chunking-settings]
39+
40+
`max_chunking_size`:::
41+
(Optional, integer)
42+
include::inference-shared.asciidoc[tag=chunking-settings-max-chunking-size]
43+
44+
`overlap`:::
45+
(Optional, integer)
46+
include::inference-shared.asciidoc[tag=chunking-settings-overlap]
47+
48+
`sentence_overlap`:::
49+
(Optional, integer)
50+
include::inference-shared.asciidoc[tag=chunking-settings-sentence-overlap]
51+
52+
`strategy`:::
53+
(Optional, string)
54+
include::inference-shared.asciidoc[tag=chunking-settings-strategy]
55+
56+
`service`::
57+
(Required, string)
58+
The type of service supported for the specified task type. In this case,
59+
`jinaai`.
60+
61+
`service_settings`::
62+
(Required, object)
63+
include::inference-shared.asciidoc[tag=service-settings]
64+
+
65+
--
66+
These settings are specific to the `jinaai` service.
67+
--
68+
69+
`api_key`:::
70+
(Required, string)
71+
A valid API key for your JinaAI account.
72+
You can find it at https://jina.ai/embeddings/.
73+
+
74+
--
75+
include::inference-shared.asciidoc[tag=api-key-admonition]
76+
--
77+
78+
`rate_limit`:::
79+
(Optional, object)
80+
The default rate limit for the `jinaai` service is 2000 requests per minute for all task types.
81+
You can modify this using the `requests_per_minute` setting in your service settings:
82+
+
83+
--
84+
include::inference-shared.asciidoc[tag=request-per-minute-example]
85+
86+
More information about JinaAI's rate limits can be found in https://jina.ai/contact-sales/#rate-limit.
87+
--
88+
+
89+
.`service_settings` for the `rerank` task type
90+
[%collapsible%closed]
91+
=====
92+
`model_id`::
93+
(Required, string)
94+
The name of the model to use for the {infer} task.
95+
To review the available `rerank` compatible models, refer to https://jina.ai/reranker.
96+
=====
97+
+
98+
.`service_settings` for the `text_embedding` task type
99+
[%collapsible%closed]
100+
=====
101+
`model_id`:::
102+
(Optional, string)
103+
The name of the model to use for the {infer} task.
104+
To review the available `text_embedding` models, refer to the
105+
https://jina.ai/embeddings/.
106+
107+
`similarity`:::
108+
(Optional, string)
109+
Similarity measure. One of `cosine`, `dot_product`, `l2_norm`.
110+
Defaults based on the `embedding_type` (`float` -> `dot_product`, `int8/byte` -> `cosine`).
111+
=====
112+
113+
114+
115+
`task_settings`::
116+
(Optional, object)
117+
include::inference-shared.asciidoc[tag=task-settings]
118+
+
119+
.`task_settings` for the `rerank` task type
120+
[%collapsible%closed]
121+
=====
122+
`return_documents`::
123+
(Optional, boolean)
124+
Specify whether to return doc text within the results.
125+
126+
`top_n`::
127+
(Optional, integer)
128+
The number of most relevant documents to return, defaults to the number of the documents.
129+
If this {infer} endpoint is used in a `text_similarity_reranker` retriever query and `top_n` is set, it must be greater than or equal to `rank_window_size` in the query.
130+
=====
131+
+
132+
.`task_settings` for the `text_embedding` task type
133+
[%collapsible%closed]
134+
=====
135+
`task`:::
136+
(Optional, string)
137+
Specifies the task passed to the model.
138+
Valid values are:
139+
* `classification`: use it for embeddings passed through a text classifier.
140+
* `clustering`: use it for the embeddings run through a clustering algorithm.
141+
* `ingest`: use it for storing document embeddings in a vector database.
142+
* `search`: use it for storing embeddings of search queries run against a vector database to find relevant documents.
143+
=====
144+
145+
146+
[discrete]
147+
[[inference-example-jinaai]]
148+
==== JinaAI service examples
149+
150+
The following examples demonstrate how to create {infer} endpoints for `text_embeddings` and `rerank` tasks using the JinaAI service and use them in search requests.
151+
152+
First, we create the `embeddings` service:
153+
154+
[source,console]
155+
------------------------------------------------------------
156+
PUT _inference/text_embedding/jinaai-embeddings
157+
{
158+
"service": "jinaai",
159+
"service_settings": {
160+
"model_id": "jina-embeddings-v3",
161+
"api_key": "<api_key>"
162+
}
163+
}
164+
------------------------------------------------------------
165+
// TEST[skip:uses ML]
166+
167+
Then, we create the `rerank` service:
168+
[source,console]
169+
------------------------------------------------------------
170+
PUT _inference/rerank/jinaai-rerank
171+
{
172+
"service": "jinaai",
173+
"service_settings": {
174+
"api_key": "<api_key>",
175+
"model_id": "jina-reranker-v2-base-multilingual"
176+
},
177+
"task_settings": {
178+
"top_n": 10,
179+
"return_documents": true
180+
}
181+
}
182+
------------------------------------------------------------
183+
// TEST[skip:uses ML]
184+
185+
Now we can create an index that will use `jinaai-embeddings` service to index the documents.
186+
187+
[source,console]
188+
------------------------------------------------------------
189+
PUT jinaai-index
190+
{
191+
"mappings": {
192+
"properties": {
193+
"content": {
194+
"type": "semantic_text",
195+
"inference_id": "jinaai-embeddings"
196+
}
197+
}
198+
}
199+
}
200+
------------------------------------------------------------
201+
// TEST[skip:uses ML]
202+
203+
[source,console]
204+
------------------------------------------------------------
205+
PUT jinaai-index/_bulk
206+
{ "index" : { "_index" : "jinaai-index", "_id" : "1" } }
207+
{"content": "Sarah Johnson is a talented marine biologist working at the Oceanographic Institute. Her groundbreaking research on coral reef ecosystems has garnered international attention and numerous accolades."}
208+
{ "index" : { "_index" : "jinaai-index", "_id" : "2" } }
209+
{"content": "She spends months at a time diving in remote locations, meticulously documenting the intricate relationships between various marine species. "}
210+
{ "index" : { "_index" : "jinaai-index", "_id" : "3" } }
211+
{"content": "Her dedication to preserving these delicate underwater environments has inspired a new generation of conservationists."}
212+
------------------------------------------------------------
213+
// TEST[skip:uses ML]
214+
215+
Now, with the index created, we can search with and without the reranker service.
216+
217+
[source,console]
218+
------------------------------------------------------------
219+
GET jinaai-index/_search
220+
{
221+
"query": {
222+
"semantic": {
223+
"field": "content",
224+
"query": "who inspired taking care of the sea?"
225+
}
226+
}
227+
}
228+
------------------------------------------------------------
229+
// TEST[skip:uses ML]
230+
231+
[source,console]
232+
------------------------------------------------------------
233+
POST jinaai-index/_search
234+
{
235+
"retriever": {
236+
"text_similarity_reranker": {
237+
"retriever": {
238+
"standard": {
239+
"query": {
240+
"semantic": {
241+
"field": "content",
242+
"query": "who inspired taking care of the sea?"
243+
}
244+
}
245+
}
246+
},
247+
"field": "content",
248+
"rank_window_size": 100,
249+
"inference_id": "jinaai-rerank",
250+
"inference_text": "who inspired taking care of the sea?"
251+
}
252+
}
253+
}
254+
------------------------------------------------------------
255+
// TEST[skip:uses ML]

0 commit comments

Comments
 (0)