💬 Commit message: Update 2026-02-15 00:07:11, 20 files, 231 lines

📁 Files changed: 20
📝 Lines changed: 231

  • .coverage
  • models.db
  • screenshot.png
  • tensors-deployed.png
  • tensors-dropdown.png
  • tensors-final.png
  • tensors-fixed.png
  • tensors-reloaded.png
  • tensors-ui.png
  • civitai_routes.py
  • generate_routes.py
  • models_routes.py
  • sd_client.py
  • index-BQdjF_w0.css
  • index-CKJOpgtQ.js
  • index-DmOZ-7Sw.js
  • index.html
  • GenerateView.vue
  • app.ts
  • index.ts

tensors/server/civitai_routes.py

@@ -10,6 +10,7 @@ from fastapi import APIRouter, Query, Response
 from fastapi.responses import JSONResponse
 
 from tensors.config import CIVITAI_API_BASE, load_api_key
+from tensors.db import Database
 
 logger = logging.getLogger(__name__)
 
@@ -56,6 +57,18 @@ async def search_models(
             response = await client.get(url, params=params, headers=_get_headers(api_key))
             response.raise_for_status()
             result: dict[str, Any] = response.json()
+
+            # Cache all models from search results
+            items = result.get("items", [])
+            if items:
+                try:
+                    with Database() as db:
+                        db.init_schema()
+                        for model_data in items:
+                            db.cache_model(model_data)
+                except Exception as e:
+                    logger.warning("Failed to cache search results: %s", e)
+
             return result
     except httpx.HTTPStatusError as e:
         logger.error("CivitAI API error: %s", e.response.status_code)
@@ -67,7 +80,7 @@ async def search_models(
 
 @router.get("/model/{model_id}", response_model=None)
 async def get_model(model_id: int) -> dict[str, Any] | Response:
-    """Get model details from CivitAI."""
+    """Get model details from CivitAI and cache to database."""
     api_key = load_api_key()
     url = f"{CIVITAI_API_BASE}/models/{model_id}"
 
@@ -76,6 +89,15 @@ async def get_model(model_id: int) -> dict[str, Any] | Response:
             response = await client.get(url, headers=_get_headers(api_key))
             response.raise_for_status()
             result: dict[str, Any] = response.json()
+
+            # Cache the model data to database
+            try:
+                with Database() as db:
+                    db.init_schema()
+                    db.cache_model(result)
+            except Exception as e:
+                logger.warning("Failed to cache model %d: %s", model_id, e)
+
             return result
     except httpx.HTTPStatusError:
         return JSONResponse({"error": "Model not found"}, status_code=404)

tensors/server/generate_routes.py

@@ -129,7 +129,7 @@ def _process_image(
 # =============================================================================
 
 
-def create_generate_router() -> APIRouter:
+def create_generate_router() -> APIRouter:  # noqa: PLR0915
     """Build a router with /api/generate endpoint."""
     router = APIRouter(prefix="/api", tags=["generate"])
     gallery = Gallery()

tensors/server/models_routes.py

@@ -11,6 +11,7 @@ from fastapi import APIRouter, HTTPException, Request
 from pydantic import BaseModel
 
 from tensors.config import MODELS_DIR
+from tensors.db import Database
 from tensors.server.sd_client import get_sd_headers
 
 logger = logging.getLogger(__name__)
@@ -110,6 +111,56 @@ def scan_models(directory: Path, extensions: tuple[str, ...] = (".safetensors",
     return models
 
 
+def _enrich_with_metadata(models: list[dict[str, Any]]) -> list[dict[str, Any]]:
+    """Enrich model data with CivitAI metadata from database."""
+    try:
+        with Database() as db:
+            db.init_schema()
+
+            for model in models:
+                file_path = model.get("path", "")
+                file_info = db.get_local_file_by_path(file_path)
+
+                if file_info and file_info.get("civitai_model_id"):
+                    # Add human-readable name
+                    model["display_name"] = file_info.get("model_name") or model["name"]
+                    model["base_model"] = file_info.get("base_model")
+                    model["model_type"] = file_info.get("model_type")
+                    model["civitai_model_id"] = file_info.get("civitai_model_id")
+                    model["civitai_version_id"] = file_info.get("civitai_version_id")
+
+                    # Get thumbnail from version images
+                    version_id = file_info.get("civitai_version_id")
+                    if version_id:
+                        cur = db.conn.cursor()
+                        cur.execute(
+                            """
+                            SELECT url FROM version_images
+                            WHERE version_id = (SELECT id FROM model_versions WHERE civitai_id = ?)
+                            ORDER BY id LIMIT 1
+                            """,
+                            (version_id,),
+                        )
+                        row = cur.fetchone()
+                        if row:
+                            model["thumbnail_url"] = row[0]
+
+                    # Get trigger words
+                    triggers = db.get_triggers_by_version(version_id) if version_id else []
+                    model["triggers"] = triggers[:5]  # Limit to first 5
+                else:
+                    model["display_name"] = model["name"]
+
+    except Exception as e:
+        logger.warning("Failed to enrich models with metadata: %s", e)
+        # Fallback: just use filename as display name
+        for model in models:
+            if "display_name" not in model:
+                model["display_name"] = model["name"]
+
+    return models
+
+
 def scan_loras(directory: Path | None = None) -> list[dict[str, Any]]:
     """Scan for LoRA files."""
     lora_dir = directory or MODELS_DIR / "loras"
@@ -127,14 +178,15 @@ def scan_checkpoints(directory: Path | None = None) -> list[dict[str, Any]]:
 # =============================================================================
 
 
-def create_models_router() -> APIRouter:
+def create_models_router() -> APIRouter:  # noqa: PLR0915
     """Build a router with /api/models/* endpoints."""
     router = APIRouter(prefix="/api/models", tags=["models"])
 
     @router.get("")
     def list_models() -> dict[str, Any]:
-        """List available checkpoint models."""
+        """List available checkpoint models with metadata."""
         checkpoints = scan_checkpoints()
+        checkpoints = _enrich_with_metadata(checkpoints)
         return {
             "models": checkpoints,
             "total": len(checkpoints),
@@ -172,8 +224,9 @@ def create_models_router() -> APIRouter:
 
     @router.get("/loras")
     def list_loras() -> dict[str, Any]:
-        """List available LoRA files."""
+        """List available LoRA files with metadata."""
         loras = scan_loras()
+        loras = _enrich_with_metadata(loras)
         return {
             "loras": loras,
             "total": len(loras),
@@ -220,14 +273,14 @@ def create_models_router() -> APIRouter:
             stdout=asyncio.subprocess.PIPE,
             stderr=asyncio.subprocess.PIPE,
         )
-        _, stderr = await proc.communicate(new_content.encode())
+        _, tee_stderr = await proc.communicate(new_content.encode())
         if proc.returncode != 0:
-            raise HTTPException(status_code=500, detail=f"Failed to write env file: {stderr.decode()}")
+            raise HTTPException(status_code=500, detail=f"Failed to write env file: {tee_stderr.decode()}")
 
         # Restart sd-server
-        returncode, _stdout, stderr = await _run_command("sudo", "systemctl", "restart", "sd-server")
+        returncode, _stdout, restart_stderr = await _run_command("sudo", "systemctl", "restart", "sd-server")
         if returncode != 0:
-            raise HTTPException(status_code=500, detail=f"Failed to restart sd-server: {stderr}")
+            raise HTTPException(status_code=500, detail=f"Failed to restart sd-server: {restart_stderr}")
 
         logger.info(f"Switched model from {old_model} to {model_path}")
 

tensors/server/sd_client.py

@@ -2,10 +2,12 @@
 
 from __future__ import annotations
 
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import httpx
-from fastapi import Request
+
+if TYPE_CHECKING:
+    from fastapi import Request
 
 
 def get_sd_headers(request: Request) -> dict[str, str]:

tensors/server/static/index.html

@@ -5,8 +5,8 @@
     <link rel="icon" type="image/svg+xml" href="/vite.svg" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>Tensors</title>
-    <script type="module" crossorigin src="/assets/index-DmOZ-7Sw.js"></script>
-    <link rel="stylesheet" crossorigin href="/assets/index-CcuP2dTH.css">
+    <script type="module" crossorigin src="/assets/index-CKJOpgtQ.js"></script>
+    <link rel="stylesheet" crossorigin href="/assets/index-BQdjF_w0.css">
   </head>
   <body>
     <div id="app"></div>

tensors/server/ui/src/components/GenerateView.vue

@@ -30,12 +30,22 @@ interface ChatMessage {
 const messages = ref<ChatMessage[]>([])
 
 const modelItems = computed(() =>
-  store.models.map(m => ({ title: m.name, value: m.path }))
+  store.models.map(m => ({
+    title: m.display_name || m.name,
+    value: m.path,
+    thumbnail: m.thumbnail_url,
+    base_model: m.base_model,
+  }))
 )
 
 const loraItems = computed(() => [
-  { title: 'None', value: '' },
-  ...store.filteredLoras.map(l => ({ title: l.name, value: l.path }))
+  { title: 'None', value: '', thumbnail: null, triggers: [] },
+  ...store.filteredLoras.map(l => ({
+    title: l.display_name || l.name,
+    value: l.path,
+    thumbnail: l.thumbnail_url,
+    triggers: l.triggers || [],
+  }))
 ])
 
 
@@ -187,9 +197,33 @@ async function generate() {
             :disabled="store.switchingModel || generating"
             density="compact"
             hide-details
-            style="width: 200px"
+            style="width: 280px"
             @update:model-value="handleModelChange"
-          />
+          >
+            <template #selection="{ item }">
+              <div class="d-flex align-center ga-2">
+                <v-avatar v-if="item.raw.thumbnail" size="24" rounded="sm">
+                  <v-img :src="item.raw.thumbnail" cover />
+                </v-avatar>
+                <v-icon v-else size="24" color="grey">mdi-cube-outline</v-icon>
+                <span class="text-truncate">{{ item.title }}</span>
+              </div>
+            </template>
+            <template #item="{ item, props }">
+              <v-list-item v-bind="props" :title="undefined">
+                <template #prepend>
+                  <v-avatar v-if="item.raw.thumbnail" size="32" rounded="sm" class="mr-3">
+                    <v-img :src="item.raw.thumbnail" cover />
+                  </v-avatar>
+                  <v-icon v-else size="32" color="grey" class="mr-3">mdi-cube-outline</v-icon>
+                </template>
+                <v-list-item-title>{{ item.title }}</v-list-item-title>
+                <v-list-item-subtitle v-if="item.raw.base_model" class="text-caption">
+                  {{ item.raw.base_model }}
+                </v-list-item-subtitle>
+              </v-list-item>
+            </template>
+          </v-select>
         </div>
 
         <div class="d-flex align-center ga-2">
@@ -200,8 +234,32 @@ async function generate() {
             :disabled="generating"
             density="compact"
             hide-details
-            style="width: 150px"
-          />
+            style="width: 200px"
+          >
+            <template #selection="{ item }">
+              <div class="d-flex align-center ga-2">
+                <v-avatar v-if="item.raw.thumbnail" size="24" rounded="sm">
+                  <v-img :src="item.raw.thumbnail" cover />
+                </v-avatar>
+                <v-icon v-else size="24" color="grey">mdi-shimmer</v-icon>
+                <span class="text-truncate">{{ item.title }}</span>
+              </div>
+            </template>
+            <template #item="{ item, props }">
+              <v-list-item v-bind="props" :title="undefined">
+                <template #prepend>
+                  <v-avatar v-if="item.raw.thumbnail" size="32" rounded="sm" class="mr-3">
+                    <v-img :src="item.raw.thumbnail" cover />
+                  </v-avatar>
+                  <v-icon v-else size="32" color="grey" class="mr-3">mdi-shimmer</v-icon>
+                </template>
+                <v-list-item-title>{{ item.title }}</v-list-item-title>
+                <v-list-item-subtitle v-if="item.raw.triggers?.length" class="text-caption text-truncate">
+                  {{ item.raw.triggers.slice(0, 2).join(', ') }}
+                </v-list-item-subtitle>
+              </v-list-item>
+            </template>
+          </v-select>
           <v-text-field
             v-model.number="store.loraWeight"
             type="number"

tensors/server/ui/src/stores/app.ts

@@ -82,7 +82,13 @@ export const useAppStore = defineStore('app', () => {
       loras.value = lorasRes.loras
       activeModel.value = activeRes.model
       if (activeRes.model) {
-        selectedModel.value = activeRes.model
+        // Find full path for the active model (API returns model name without extension, v-select uses path)
+        const activeModelPath = modelsRes.models.find(m =>
+          m.name === activeRes.model ||
+          m.filename === activeRes.model ||
+          m.filename.startsWith(activeRes.model + '.')
+        )?.path
+        selectedModel.value = activeModelPath || activeRes.model
       }
     } catch (error) {
       console.error('Failed to load models:', error)

tensors/server/ui/src/types/index.ts

@@ -5,6 +5,14 @@ export interface Model {
   size_mb: number
   modified: number
   category: 'sd15' | 'large'
+  // Enriched from CivitAI metadata
+  display_name?: string
+  base_model?: string
+  model_type?: string
+  civitai_model_id?: number
+  civitai_version_id?: number
+  thumbnail_url?: string
+  triggers?: string[]
 }
 
 export interface LoRA {
@@ -14,6 +22,14 @@ export interface LoRA {
   size_mb: number
   modified: number
   category: 'sd15' | 'large'
+  // Enriched from CivitAI metadata
+  display_name?: string
+  base_model?: string
+  model_type?: string
+  civitai_model_id?: number
+  civitai_version_id?: number
+  thumbnail_url?: string
+  triggers?: string[]
 }
 
 export interface GeneratedImage {