@jeromeetienne/openai-cache 1.0.2 → 1.0.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +22 -5
- package/dist/openai_cache.js +8 -6
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -7,11 +7,23 @@ You can use any Keyv storage backend (like Redis, filesystem, etc) to store the
|
|
|
7
7
|
See the [Keyv documentation](https://keyv.org/docs/) for more details on available storage options and how to set them up.
|
|
8
8
|
In the example below, we use a SQLite database to persist the cache.
|
|
9
9
|
|
|
10
|
+
# Installation
|
|
11
|
+
|
|
12
|
+
```bash
|
|
13
|
+
npm install @jeromeetienne/openai-cache
|
|
14
|
+
```
|
|
15
|
+
|
|
16
|
+
If you want to use the SQLite storage backend, you also need to install the `@keyv/sqlite` package:
|
|
17
|
+
|
|
18
|
+
```bash
|
|
19
|
+
npm install @keyv/sqlite
|
|
20
|
+
```
|
|
21
|
+
|
|
10
22
|
## Usage
|
|
11
23
|
|
|
12
24
|
```ts
|
|
13
25
|
import OpenAI from "openai";
|
|
14
|
-
import
|
|
26
|
+
import OpenAICache from "@jeromeetienne/openai-cache";
|
|
15
27
|
import KeyvSqlite from '@keyv/sqlite';
|
|
16
28
|
import { Cacheable } from "cacheable";
|
|
17
29
|
|
|
@@ -42,14 +54,14 @@ console.log(response.output_text);
|
|
|
42
54
|
- **PRO**: Reduces redundant API calls, saving time and costs.
|
|
43
55
|
data.
|
|
44
56
|
- **NOTE**: When `temperature === 0`, caching works optimally as responses are deterministic. However, with `temperature > 0`, caching may reduce variety across multiple calls since identical prompts will return cached results instead of generating new varied responses.
|
|
57
|
+
- **NOTE**: Only successful responses (`2xx`) are cached. Error responses (`4xx`/`5xx`) are returned normally but are not persisted.
|
|
45
58
|
|
|
46
59
|
|
|
47
60
|
## Possible improvements
|
|
48
61
|
- dont cache if temporature > 0 or top_p < 1, You’ll freeze randomness if cached
|
|
49
62
|
- NOTE: do that on options
|
|
50
|
-
-
|
|
51
|
-
|
|
52
|
-
- tools requests errors should not be cached
|
|
63
|
+
- add configurable cache policy for errors (for example, cache selected deterministic `4xx` while never caching `429`/`5xx`)
|
|
64
|
+
- tools requests errors should not be cached
|
|
53
65
|
|
|
54
66
|
## Developper Notes
|
|
55
67
|
|
|
@@ -60,4 +72,9 @@ A. Do the following steps:
|
|
|
60
72
|
npm run version:patch && npm run publish:all
|
|
61
73
|
```
|
|
62
74
|
|
|
63
|
-
Lots of trouble with the 2fa system
|
|
75
|
+
Lots of trouble with the 2fa system
|
|
76
|
+
|
|
77
|
+
Revevant Documentation:
|
|
78
|
+
- https://docs.npmjs.com/requiring-2fa-for-package-publishing-and-settings-modification
|
|
79
|
+
- https://docs.npmjs.com/trusted-publishers
|
|
80
|
+
- https://docs.npmjs.com/creating-and-viewing-access-tokens
|
package/dist/openai_cache.js
CHANGED
|
@@ -73,12 +73,14 @@ class OpenAICache {
|
|
|
73
73
|
// Collect headers and normalize them
|
|
74
74
|
const headers = Array.from(clonedResponse.headers.entries());
|
|
75
75
|
const normalizedHeaders = OpenAICache._normalizeHeaders(headers, responseBuffer.length);
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
|
|
76
|
+
if (response.ok) {
|
|
77
|
+
await this._cache.set(cacheKey, {
|
|
78
|
+
status: clonedResponse.status,
|
|
79
|
+
headers: normalizedHeaders,
|
|
80
|
+
body: responseBuffer.toString("base64"),
|
|
81
|
+
bodyEncoding: "base64",
|
|
82
|
+
});
|
|
83
|
+
}
|
|
82
84
|
// Return live response (body already buffered)
|
|
83
85
|
return new Response(responseBuffer, { status: response.status, headers: normalizedHeaders });
|
|
84
86
|
}
|