You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+7-6Lines changed: 7 additions & 6 deletions
Original file line number
Diff line number
Diff line change
@@ -37,7 +37,7 @@ $ composer require rubix/server
37
37
38
38
---
39
39
### Servers
40
-
Rubix model servers are stand-alone processes that wrap an estimator in an API that can be queried over a network connection. Since servers implements their own networking stack, they can be run directly from the PHP command line interface (CLI) without the need for an intermediary server such as Nginx or Apache. By utilizing concurrency, each server instance is able to handle thousands of connections at the same time. Need more inference throughput? Model servers scale by adding more instances behind a load balancer.
40
+
Rubix model servers are stand-alone processes that wrap an estimator in an API that can be queried over a network connection. Since servers implements their own networking stack, they can be run directly from the PHP command line interface (CLI) without the need for an intermediary server such as Nginx or Apache. By utilizing concurrency, each server instance is able to handle thousands of connections at the same time.
41
41
42
42
To boot up a server, pass a trained estimator instance to the `serve()` method:
A JSON over HTTP server exposing Representational State Transfer (REST) and GraphQL APIs. The HTTP Server operates with ubiquitous standards making it compatible with a wide range of systems. In addition, it provides its own web-based user interface for real-time server monitoring.
96
+
A JSON over HTTP server exposing Representational State Transfer (REST) and GraphQL APIs. The HTTP Server operates using ubiquitous standards making it compatible with a wide range of systems. In addition, it provides its own web-based user interface for dataset visualization and real-time server monitoring.
| 2 | port | 8000 | int | The network port to run the HTTP services on. |
105
105
| 3 | cert | null | string | The path to the certificate used to authenticate and encrypt the HTTP channel. |
106
106
| 4 | middlewares |[]| array | The stack of server middleware to run on each request/response. |
107
-
| 5 | max concurrent requests |20| int | The maximum number of requests that can be handled concurrently. |
107
+
| 5 | max concurrent requests |10| int | The maximum number of requests that can be handled concurrently. |
108
108
| 6 | sse reconnect buffer | 50 | int | The maximum number of events to store in the server-sent events (SSE) reconnect buffer. |
109
109
110
110
**Example**
@@ -138,6 +138,7 @@ The HTTP server exposes the following resources and their methods.
138
138
| GET | /server/dashboard | Return the properties of the server dashboard. |
139
139
| GET | /server/dashboard/events | Subscribe to the dashboard events stream. |
140
140
| GET | /ui/server | The server dashboard interface. |
141
+
| GET | /ui/visualizer/bubble | Dataset bubble chart visualizer. |
141
142
142
143
#### Web Interface
143
144
The HTTP server provides its own high-level user interface to the GraphQL API it exposes under the hood. To access the on-demand web UI, navigate to `http://hostname:port` (or `https://hostname:port` if using a secure socket connection) using your web browser. The example below is a screen capture of the server dashboard in dark mode.
@@ -334,8 +335,8 @@ Adds the necessary authorization headers to the request using the Basic scheme.
334
335
#### Parameters
335
336
| # | Param | Default | Type | Description |
336
337
|---|---|---|---|---|
337
-
| 1 | username || string | The username. |
338
-
| 2 | password || string | The password. |
338
+
| 1 | username || string | The user's name. |
339
+
| 2 | password || string | The user's password. |
339
340
340
341
**Example**
341
342
@@ -398,7 +399,7 @@ Yes, model server are designed to coexist with other web servers (including othe
398
399
Since model servers are inference-only (i.e. they only support queries), they scale horizontally by adding more instances behind a load balancer such as [Nginx](http://nginx.org).
399
400
400
401
#### Do servers support compression?
401
-
Yes, the HTTP Server supports both Gzip and Deflate compression schemes applied to the request bodies. However, since response bodies tend to be considerably smaller compared to requests for this application, they do not support outbound message compression.
402
+
Yes, the HTTP Server supports both Gzip and Deflate compression schemes applied to the request bodies and to the response bodies of requests for static assets.
402
403
403
404
## License
404
405
The code is licensed [MIT](LICENSE.md) and the documentation is licensed [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/).
0 commit comments