@cltec/pi-splash-stats 0.1.0 → 0.1.1

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/CHANGELOG.md CHANGED
@@ -3,6 +3,13 @@
3
3
  Notable changes to this project. Entries are grouped by release; the latest is listed first.
4
4
  This repository follows [Keep a Changelog](https://keepachangelog.com/) for entry style.
5
5
 
6
+ ## [0.1.1]
7
+
8
+ ### Changed
9
+
10
+ - README: credit `pi-llama-cpp-stats` as the inspiration for this extension.
11
+ - README: link the [Splash](https://github.com/incoai/splash) server repository where Splash is mentioned.
12
+
6
13
  ## [0.1.0]
7
14
 
8
15
  Initial release of the `pi-splash-stats` Pi extension.
package/README.md CHANGED
@@ -1,6 +1,6 @@
1
1
  # Pi Splash Stats
2
2
 
3
- A Pi extension that shows live inference statistics for OpenAI-compatible **Splash** servers in the
3
+ A Pi extension that shows live inference statistics for OpenAI-compatible [**Splash**](https://github.com/incoai/splash) servers in the
4
4
  interactive terminal UI: a **cache-aware prefill progress bar** while the prompt is being processed,
5
5
  and a **live generation token count** while the model streams its answer.
6
6
 
@@ -8,6 +8,8 @@ It is a passive observer. It enriches and watches the inference traffic Pi alrea
8
8
  verified Splash backend, then renders two transient status lines in the TUI. It changes no prompts,
9
9
  no tools, and no responses, and it stays completely silent in every non-interactive mode.
10
10
 
11
+ This extension is inspired by [pi-llama-cpp-stats](https://www.npmjs.com/package/pi-llama-cpp-stats).
12
+
11
13
  ```
12
14
  ── ⠸ Prefilling... ██████████░░░░░░░░░░ 50% · 2s · 20.0 tok/s ─────────────────────
13
15
  Generating... ≈5 tok
@@ -71,7 +73,7 @@ response completes or is aborted.
71
73
 
72
74
  Two launch workflows both work and produce equivalent statistics for the same Splash-served model:
73
75
 
74
- - **`splash pi ...`** — the Splash launcher starts Pi already pointed at a running Splash model. The
76
+ - **`splash pi ...`** — the [Splash](https://github.com/incoai/splash) launcher starts Pi already pointed at a running Splash model. The
75
77
  extension observes that model's requests exactly as it would for a directly launched Pi; it does not
76
78
  inspect the launcher, parent process, or a particular provider name. If you switch to a non-Splash
77
79
  (e.g. cloud) model, the stats clear immediately and subsequent requests are left untouched.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@cltec/pi-splash-stats",
3
- "version": "0.1.0",
3
+ "version": "0.1.1",
4
4
  "private": false,
5
5
  "description": "Pi extension that displays live inference statistics for OpenAI-compatible Splash servers.",
6
6
  "type": "module",