Skip to content

Commit c81a295

Browse files
paoloredispaoloredis
committed
Port redisvl documentation
1 parent 29f35d9 commit c81a295

22 files changed

+8427
-0
lines changed

‎content/integrate/redisvl/api/_index.md‎

Lines changed: 54 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,54 @@
1+
---
2+
linkTitle: RedisVL API
3+
title: RedisVL API
4+
type: integration
5+
---
6+
7+
8+
Reference documentation for the RedisVL API.
9+
10+
11+
12+
* [Schema](schema/)
13+
* [IndexSchema](schema/#indexschema)
14+
* [Defining Fields](schema/#defining-fields)
15+
* [Supported Field Types and Attributes](schema/#supported-field-types-and-attributes)
16+
* [Search Index Classes](searchindex/)
17+
* [SearchIndex](searchindex/#searchindex)
18+
* [AsyncSearchIndex](searchindex/#asyncsearchindex)
19+
* [Query](query/)
20+
* [VectorQuery](query/#vectorquery)
21+
* [VectorRangeQuery](query/#vectorrangequery)
22+
* [FilterQuery](query/#filterquery)
23+
* [CountQuery](query/#countquery)
24+
* [Filter](filter/)
25+
* [FilterExpression](filter/#filterexpression)
26+
* [Tag](filter/#tag)
27+
* [Text](filter/#text)
28+
* [Num](filter/#num)
29+
* [Geo](filter/#geo)
30+
* [GeoRadius](filter/#georadius)
31+
* [Vectorizers](vectorizer/)
32+
* [HFTextVectorizer](vectorizer/#hftextvectorizer)
33+
* [OpenAITextVectorizer](vectorizer/#openaitextvectorizer)
34+
* [AzureOpenAITextVectorizer](vectorizer/#azureopenaitextvectorizer)
35+
* [VertexAITextVectorizer](vectorizer/#vertexaitextvectorizer)
36+
* [CohereTextVectorizer](vectorizer/#coheretextvectorizer)
37+
* [BedrockTextVectorizer](vectorizer/#bedrocktextvectorizer)
38+
* [CustomTextVectorizer](vectorizer/#customtextvectorizer)
39+
* [VoyageAITextVectorizer](vectorizer/#voyageaitextvectorizer)
40+
* [Rerankers](reranker/)
41+
* [CohereReranker](reranker/#coherereranker)
42+
* [HFCrossEncoderReranker](reranker/#hfcrossencoderreranker)
43+
* [VoyageAIReranker](reranker/#voyageaireranker)
44+
* [LLM Cache](cache/)
45+
* [SemanticCache](cache/#semanticcache)
46+
* [LLM Session Manager](session_manager/)
47+
* [SemanticSessionManager](session_manager/#semanticsessionmanager)
48+
* [StandardSessionManager](session_manager/#standardsessionmanager)
49+
* [Semantic Router](router/)
50+
* [Semantic Router](router/#semantic-router-api)
51+
* [Routing Config](router/#routing-config)
52+
* [Route](router/#route)
53+
* [Route Match](router/#route-match)
54+
* [Distance Aggregation Method](router/#distance-aggregation-method)

‎content/integrate/redisvl/api/cache.md‎

Lines changed: 320 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,320 @@
1+
---
2+
linkTitle: LLM Cache
3+
title: LLM Cache
4+
type: integration
5+
---
6+
7+
8+
## SemanticCache
9+
10+
<a id="semantic-cache-api"></a>
11+
12+
### *class* SemanticCache(name='llmcache', distance_threshold=0.1, ttl=None, vectorizer=None, filterable_fields=None, redis_client=None, redis_url='redis://localhost:6379', connection_kwargs={}, overwrite=False, \*\*kwargs)
13+
14+
Bases: `BaseLLMCache`
15+
16+
Semantic Cache for Large Language Models.
17+
18+
Semantic Cache for Large Language Models.
19+
20+
* **Parameters:**
21+
* **name** (*str* *,* *optional*) – The name of the semantic cache search index.
22+
Defaults to “llmcache”.
23+
* **distance_threshold** (*float* *,* *optional*) – Semantic threshold for the
24+
cache. Defaults to 0.1.
25+
* **ttl** (*Optional* *[**int* *]* *,* *optional*) – The time-to-live for records cached
26+
in Redis. Defaults to None.
27+
* **vectorizer** (*Optional* *[**BaseVectorizer* *]* *,* *optional*) – The vectorizer for the cache.
28+
Defaults to HFTextVectorizer.
29+
* **filterable_fields** (*Optional* *[**List* *[**Dict* *[**str* *,* *Any* *]* *]* *]*) – An optional list of RedisVL fields
30+
that can be used to customize cache retrieval with filters.
31+
* **redis_client** (*Optional* *[**Redis* *]* *,* *optional*) – A redis client connection instance.
32+
Defaults to None.
33+
* **redis_url** (*str* *,* *optional*) – The redis url. Defaults to redis://localhost:6379.
34+
* **connection_kwargs** (*Dict* *[**str* *,* *Any* *]*) – The connection arguments
35+
for the redis client. Defaults to empty {}.
36+
* **overwrite** (*bool*) – Whether or not to force overwrite the schema for
37+
the semantic cache index. Defaults to false.
38+
* **Raises:**
39+
* **TypeError** – If an invalid vectorizer is provided.
40+
* **TypeError** – If the TTL value is not an int.
41+
* **ValueError** – If the threshold is not between 0 and 1.
42+
* **ValueError** – If existing schema does not match new schema and overwrite is False.
43+
44+
#### *async* acheck(prompt=None, vector=None, num_results=1, return_fields=None, filter_expression=None, distance_threshold=None)
45+
46+
Async check the semantic cache for results similar to the specified prompt
47+
or vector.
48+
49+
This method searches the cache using vector similarity with
50+
either a raw text prompt (converted to a vector) or a provided vector as
51+
input. It checks for semantically similar prompts and fetches the cached
52+
LLM responses.
53+
54+
* **Parameters:**
55+
* **prompt** (*Optional* *[**str* *]* *,* *optional*) – The text prompt to search for in
56+
the cache.
57+
* **vector** (*Optional* *[**List* *[**float* *]* *]* *,* *optional*) – The vector representation
58+
of the prompt to search for in the cache.
59+
* **num_results** (*int* *,* *optional*) – The number of cached results to return.
60+
Defaults to 1.
61+
* **return_fields** (*Optional* *[**List* *[**str* *]* *]* *,* *optional*) – The fields to include
62+
in each returned result. If None, defaults to all available
63+
fields in the cached entry.
64+
* **filter_expression** (*Optional* *[*[*FilterExpression*](filter.md#redisvl.query.filter.FilterExpression) *]*) – Optional filter expression
65+
that can be used to filter cache results. Defaults to None and
66+
the full cache will be searched.
67+
* **distance_threshold** (*Optional* *[**float* *]*) – The threshold for semantic
68+
vector distance.
69+
* **Returns:**
70+
A list of dicts containing the requested
71+
: return fields for each similar cached response.
72+
* **Return type:**
73+
List[Dict[str, Any]]
74+
* **Raises:**
75+
* **ValueError** – If neither a prompt nor a vector is specified.
76+
* **ValueError** – if ‘vector’ has incorrect dimensions.
77+
* **TypeError** – If return_fields is not a list when provided.
78+
79+
```python
80+
response = await cache.acheck(
81+
prompt="What is the captial city of France?"
82+
)
83+
```
84+
85+
#### *async* adrop(ids=None, keys=None)
86+
87+
Async expire specific entries from the cache by id or specific
88+
Redis key.
89+
90+
* **Parameters:**
91+
* **ids** (*Optional* *[**str* *]*) – The document ID or IDs to remove from the cache.
92+
* **keys** (*Optional* *[**str* *]*) – The Redis keys to remove from the cache.
93+
* **Return type:**
94+
None
95+
96+
#### *async* astore(prompt, response, vector=None, metadata=None, filters=None, ttl=None)
97+
98+
Async stores the specified key-value pair in the cache along with metadata.
99+
100+
* **Parameters:**
101+
* **prompt** (*str*) – The user prompt to cache.
102+
* **response** (*str*) – The LLM response to cache.
103+
* **vector** (*Optional* *[**List* *[**float* *]* *]* *,* *optional*) – The prompt vector to
104+
cache. Defaults to None, and the prompt vector is generated on
105+
demand.
106+
* **metadata** (*Optional* *[**Dict* *[**str* *,* *Any* *]* *]* *,* *optional*) – The optional metadata to cache
107+
alongside the prompt and response. Defaults to None.
108+
* **filters** (*Optional* *[**Dict* *[**str* *,* *Any* *]* *]*) – The optional tag to assign to the cache entry.
109+
Defaults to None.
110+
* **ttl** (*Optional* *[**int* *]*) – The optional TTL override to use on this individual cache
111+
entry. Defaults to the global TTL setting.
112+
* **Returns:**
113+
The Redis key for the entries added to the semantic cache.
114+
* **Return type:**
115+
str
116+
* **Raises:**
117+
* **ValueError** – If neither prompt nor vector is specified.
118+
* **ValueError** – if vector has incorrect dimensions.
119+
* **TypeError** – If provided metadata is not a dictionary.
120+
121+
```python
122+
key = await cache.astore(
123+
prompt="What is the captial city of France?",
124+
response="Paris",
125+
metadata={"city": "Paris", "country": "France"}
126+
)
127+
```
128+
129+
#### *async* aupdate(key, \*\*kwargs)
130+
131+
Async update specific fields within an existing cache entry. If no fields
132+
are passed, then only the document TTL is refreshed.
133+
134+
* **Parameters:**
135+
**key** (*str*) – the key of the document to update using kwargs.
136+
* **Raises:**
137+
* **ValueError if an incorrect mapping is provided as a kwarg.** –
138+
* **TypeError if metadata is provided and not** **of** **type dict.** –
139+
* **Return type:**
140+
None
141+
142+
```python
143+
key = await cache.astore('this is a prompt', 'this is a response')
144+
await cache.aupdate(
145+
key,
146+
metadata={"hit_count": 1, "model_name": "Llama-2-7b"}
147+
)
148+
```
149+
150+
#### check(prompt=None, vector=None, num_results=1, return_fields=None, filter_expression=None, distance_threshold=None)
151+
152+
Checks the semantic cache for results similar to the specified prompt
153+
or vector.
154+
155+
This method searches the cache using vector similarity with
156+
either a raw text prompt (converted to a vector) or a provided vector as
157+
input. It checks for semantically similar prompts and fetches the cached
158+
LLM responses.
159+
160+
* **Parameters:**
161+
* **prompt** (*Optional* *[**str* *]* *,* *optional*) – The text prompt to search for in
162+
the cache.
163+
* **vector** (*Optional* *[**List* *[**float* *]* *]* *,* *optional*) – The vector representation
164+
of the prompt to search for in the cache.
165+
* **num_results** (*int* *,* *optional*) – The number of cached results to return.
166+
Defaults to 1.
167+
* **return_fields** (*Optional* *[**List* *[**str* *]* *]* *,* *optional*) – The fields to include
168+
in each returned result. If None, defaults to all available
169+
fields in the cached entry.
170+
* **filter_expression** (*Optional* *[*[*FilterExpression*](filter.md#redisvl.query.filter.FilterExpression) *]*) – Optional filter expression
171+
that can be used to filter cache results. Defaults to None and
172+
the full cache will be searched.
173+
* **distance_threshold** (*Optional* *[**float* *]*) – The threshold for semantic
174+
vector distance.
175+
* **Returns:**
176+
A list of dicts containing the requested
177+
: return fields for each similar cached response.
178+
* **Return type:**
179+
List[Dict[str, Any]]
180+
* **Raises:**
181+
* **ValueError** – If neither a prompt nor a vector is specified.
182+
* **ValueError** – if ‘vector’ has incorrect dimensions.
183+
* **TypeError** – If return_fields is not a list when provided.
184+
185+
```python
186+
response = cache.check(
187+
prompt="What is the captial city of France?"
188+
)
189+
```
190+
191+
#### clear()
192+
193+
Clear the cache of all keys while preserving the index.
194+
195+
* **Return type:**
196+
None
197+
198+
#### delete()
199+
200+
Clear the semantic cache of all keys and remove the underlying search
201+
index.
202+
203+
* **Return type:**
204+
None
205+
206+
#### drop(ids=None, keys=None)
207+
208+
Manually expire specific entries from the cache by id or specific
209+
Redis key.
210+
211+
* **Parameters:**
212+
* **ids** (*Optional* *[**str* *]*) – The document ID or IDs to remove from the cache.
213+
* **keys** (*Optional* *[**str* *]*) – The Redis keys to remove from the cache.
214+
* **Return type:**
215+
None
216+
217+
#### set_threshold(distance_threshold)
218+
219+
Sets the semantic distance threshold for the cache.
220+
221+
* **Parameters:**
222+
**distance_threshold** (*float*) – The semantic distance threshold for
223+
the cache.
224+
* **Raises:**
225+
**ValueError** – If the threshold is not between 0 and 1.
226+
* **Return type:**
227+
None
228+
229+
#### set_ttl(ttl=None)
230+
231+
Set the default TTL, in seconds, for entries in the cache.
232+
233+
* **Parameters:**
234+
**ttl** (*Optional* *[**int* *]* *,* *optional*) – The optional time-to-live expiration
235+
for the cache, in seconds.
236+
* **Raises:**
237+
**ValueError** – If the time-to-live value is not an integer.
238+
239+
#### store(prompt, response, vector=None, metadata=None, filters=None, ttl=None)
240+
241+
Stores the specified key-value pair in the cache along with metadata.
242+
243+
* **Parameters:**
244+
* **prompt** (*str*) – The user prompt to cache.
245+
* **response** (*str*) – The LLM response to cache.
246+
* **vector** (*Optional* *[**List* *[**float* *]* *]* *,* *optional*) – The prompt vector to
247+
cache. Defaults to None, and the prompt vector is generated on
248+
demand.
249+
* **metadata** (*Optional* *[**Dict* *[**str* *,* *Any* *]* *]* *,* *optional*) – The optional metadata to cache
250+
alongside the prompt and response. Defaults to None.
251+
* **filters** (*Optional* *[**Dict* *[**str* *,* *Any* *]* *]*) – The optional tag to assign to the cache entry.
252+
Defaults to None.
253+
* **ttl** (*Optional* *[**int* *]*) – The optional TTL override to use on this individual cache
254+
entry. Defaults to the global TTL setting.
255+
* **Returns:**
256+
The Redis key for the entries added to the semantic cache.
257+
* **Return type:**
258+
str
259+
* **Raises:**
260+
* **ValueError** – If neither prompt nor vector is specified.
261+
* **ValueError** – if vector has incorrect dimensions.
262+
* **TypeError** – If provided metadata is not a dictionary.
263+
264+
```python
265+
key = cache.store(
266+
prompt="What is the captial city of France?",
267+
response="Paris",
268+
metadata={"city": "Paris", "country": "France"}
269+
)
270+
```
271+
272+
#### update(key, \*\*kwargs)
273+
274+
Update specific fields within an existing cache entry. If no fields
275+
are passed, then only the document TTL is refreshed.
276+
277+
* **Parameters:**
278+
**key** (*str*) – the key of the document to update using kwargs.
279+
* **Raises:**
280+
* **ValueError if an incorrect mapping is provided as a kwarg.** –
281+
* **TypeError if metadata is provided and not** **of** **type dict.** –
282+
* **Return type:**
283+
None
284+
285+
```python
286+
key = cache.store('this is a prompt', 'this is a response')
287+
cache.update(key, metadata={"hit_count": 1, "model_name": "Llama-2-7b"})
288+
)
289+
```
290+
291+
#### *property* aindex *: [AsyncSearchIndex](searchindex.md#redisvl.index.AsyncSearchIndex) | None*
292+
293+
The underlying AsyncSearchIndex for the cache.
294+
295+
* **Returns:**
296+
The async search index.
297+
* **Return type:**
298+
[AsyncSearchIndex](searchindex.md#redisvl.index.AsyncSearchIndex)
299+
300+
#### *property* distance_threshold *: float*
301+
302+
The semantic distance threshold for the cache.
303+
304+
* **Returns:**
305+
The semantic distance threshold.
306+
* **Return type:**
307+
float
308+
309+
#### *property* index *: [SearchIndex](searchindex.md#redisvl.index.SearchIndex)*
310+
311+
The underlying SearchIndex for the cache.
312+
313+
* **Returns:**
314+
The search index.
315+
* **Return type:**
316+
[SearchIndex](searchindex.md#redisvl.index.SearchIndex)
317+
318+
#### *property* ttl *: int | None*
319+
320+
The default TTL, in seconds, for entries in the cache.

0 commit comments

Comments
 (0)