@vtstech/pi-ollama-sync 1.0.4 → 1.0.6

package/README.md

@@ -0,0 +1,40 @@
+# @vtstech/pi-ollama-sync
+
+Ollama sync extension for the [Pi Coding Agent](https://github.com/badlogic/pi-mono).
+
+Auto-populate `models.json` with all available Ollama models — works with local and remote instances.
+
+## Install
+
+```bash
+pi install "npm:@vtstech/pi-ollama-sync"
+```
+
+## Commands
+
+```bash
+/ollama-sync                          Sync from models.json URL (or localhost)
+/ollama-sync https://your-tunnel-url  Sync from a specific remote URL
+```
+
+## Features
+
+- Queries Ollama `/api/tags` for available models (local or remote)
+- Writes the actual Ollama URL back into `models.json` so other extensions pick it up
+- URL priority: CLI argument → existing `models.json` baseUrl → `OLLAMA_HOST` env → localhost
+- Preserves existing provider config (apiKey, compat settings)
+- Defaults to `openai-completions` API mode
+- Sorts models by size (smallest first)
+- Auto-detects reasoning-capable models (deepseek-r1, qwq, o1, o3, think, reason)
+- Merges with existing per-model settings
+- Per-model metadata in sync report (parameter size, quantization level, model family)
+- Registered as both `/ollama-sync` slash command and `ollama_sync` tool
+
+## Links
+
+- [Full Documentation](https://github.com/VTSTech/pi-coding-agent#ollama-sync-ollama-syncts)
+- [Changelog](https://github.com/VTSTech/pi-coding-agent/blob/main/CHANGELOG.md)
+
+## License
+
+MIT — [VTSTech](https://www.vts-tech.org)

package/ollama-sync.js

@@ -2,14 +2,15 @@
 import {
   MODELS_JSON_PATH as MODELS_FILE,
   fetchOllamaModels,
+  fetchContextLengthsBatched,
   readModelsJson,
   writeModelsJson,
   isReasoningModel,
   getOllamaBaseUrl
 } from "@vtstech/pi-shared/ollama";
-import { section, ok, warn, info } from "@vtstech/pi-shared/format";
+import { section, ok, warn, info, bytesHuman, estimateVram } from "@vtstech/pi-shared/format";
 var BRANDING = [
-  `  \u26A1 Pi Ollama Sync v1.0.3`,
+  `  \u26A1 Pi Ollama Sync v1.0.6`,
   `  Written by VTSTech`,
   `  GitHub: https://github.com/VTSTech`,
   `  Website: www.vts-tech.org`
@@ -27,13 +28,16 @@ function getProviderConfig(existing) {
     }
   };
 }
-function buildModelEntry(m) {
+function buildModelEntry(m, contextLength) {
+  const estimatedSize = estimateVram(m.details.parameter_size, m.details.quantization_level);
   return {
     id: m.name,
     reasoning: isReasoningModel(m.name),
     parameterSize: m.details.parameter_size,
     quantizationLevel: m.details.quantization_level,
-    modelFamily: m.details.family || m.details.families?.[0] || "unknown"
+    modelFamily: m.details.family || m.details.families?.[0] || "unknown",
+    contextLength,
+    estimatedSize
   };
 }
 function mergeModels(newModels, oldModels) {
@@ -74,7 +78,14 @@ function ollama_sync_temp_default(pi) {
           return;
         }
         const sorted = [...models].sort((a, b) => a.size - b.size);
-        const newModels = sorted.map(buildModelEntry);
+        ctx.ui.setStatus("ollama-sync", "Fetching model details...");
+        const contextMap = await fetchContextLengthsBatched(
+          ollamaBaseUrl,
+          sorted.map((m) => m.name)
+        );
+        const newModels = sorted.map(
+          (m) => buildModelEntry(m, contextMap.get(m.name))
+        );
         const oldIds = new Set(
           existing.providers["ollama"]?.models?.map((m) => m.id) ?? []
         );
@@ -96,8 +107,10 @@ function ollama_sync_temp_default(pi) {
         lines.push(section("Synced Models"));
         for (const m of newModels) {
           lines.push(ok(m.id));
+          const ctxStr = m.contextLength != null ? m.contextLength.toLocaleString() : "?";
+          const sizeStr = m.estimatedSize ? bytesHuman(m.estimatedSize) : "?";
           lines.push(
-            `       Params: ${m.parameterSize ?? "?"} \xB7 Quant: ${m.quantizationLevel ?? "?"} \xB7 Family: ${m.modelFamily ?? "?"}`
+            `       Params: ${m.parameterSize ?? "?"} \xB7 Quant: ${m.quantizationLevel ?? "?"} \xB7 Family: ${m.modelFamily ?? "?"} \xB7 Context: ${ctxStr} \xB7 VRAM: ~${sizeStr}`
           );
         }
         if (added.length > 0 || removed.length > 0) {
@@ -153,7 +166,13 @@ function ollama_sync_temp_default(pi) {
       try {
         const models = await fetchOllamaModels(ollamaBaseUrl);
         const sorted = [...models].sort((a, b) => a.size - b.size);
-        const newModels = sorted.map(buildModelEntry);
+        const contextMap = await fetchContextLengthsBatched(
+          ollamaBaseUrl,
+          sorted.map((m) => m.name)
+        );
+        const newModels = sorted.map(
+          (m) => buildModelEntry(m, contextMap.get(m.name))
+        );
         const mergedModels = mergeModels(
           newModels,
           existing.providers["ollama"]?.models ?? []
@@ -165,7 +184,11 @@ function ollama_sync_temp_default(pi) {
         };
         writeModelsJson(existing);
         const modelDetails = newModels.map(
-          (m) => `  \u2022 ${m.id} (${m.parameterSize}, ${m.quantizationLevel}, family: ${m.modelFamily})`
+          (m) => {
+            const ctxStr = m.contextLength ?? "?";
+            const sizeStr = m.estimatedSize ? bytesHuman(m.estimatedSize) : "?";
+            return `  \u2022 ${m.id} (${m.parameterSize}, ${m.quantizationLevel}, ctx: ${ctxStr}, ~${sizeStr})`;
+          }
         ).join("\n");
         return {
           content: [

package/package.json

@@ -1,9 +1,9 @@
 {
   "name": "@vtstech/pi-ollama-sync",
-  "version": "1.0.4",
+  "version": "1.0.6",
   "description": "Ollama model sync extension for Pi Coding Agent",
   "main": "ollama-sync.js",
-  "keywords": ["pi-package", "pi-extensions"],
+  "keywords": ["pi-extensions"],
   "license": "MIT",
   "access": "public",
   "type": "module",
@@ -14,7 +14,7 @@
     "url": "https://github.com/VTSTech/pi-coding-agent"
   },
   "dependencies": {
-    "@vtstech/pi-shared": "1.0.4"
+    "@vtstech/pi-shared": "1.0.6"
   },
   "peerDependencies": {
     "@mariozechner/pi-coding-agent": ">=0.66"