feat: add markdown link extraction and HF URL normalization in ComfyModelAnalyzer and ModelSearchService

2026-04-22 19:39:11 +02:00
parent 6116eb6d63
commit 60320e13a2
24 changed files with 1661 additions and 250 deletions
@@ -0,0 +1,54 @@
+package de.tki.comfymodels.service;
+
+import de.tki.comfymodels.domain.ModelInfo;
+import de.tki.comfymodels.service.impl.ComfyModelAnalyzer;
+import org.junit.jupiter.api.Test;
+import java.util.List;
+import static org.junit.jupiter.api.Assertions.*;
+
+public class ReproductionTest {
+
+    private final ComfyModelAnalyzer analyzer = new ComfyModelAnalyzer();
+
+    @Test
+    public void testDeduplicationAndUrlFiltering() {
+        String json = "{\n" +
+                "  \"nodes\": [\n" +
+                "    {\n" +
+                "      \"type\": \"MarkdownNote\",\n" +
+                "      \"widgets_values\": [\n" +
+                "        \"- [qwen_3_4b.safetensors](https://huggingface.co/Comfy-Org/z_image_turbo/resolve/main/split_files/text_encoders/qwen_3_4b.safetensors)\"\n" +
+                "      ]\n" +
+                "    },\n" +
+                "    {\n" +
+                "      \"type\": \"CLIPLoader\",\n" +
+                "      \"properties\": {\n" +
+                "        \"models\": [\n" +
+                "          {\n" +
+                "            \"name\": \"qwen_3_4b.safetensors\",\n" +
+                "            \"url\": \"https://huggingface.co/Comfy-Org/z_image_turbo/resolve/main/split_files/text_encoders/qwen_3_4b.safetensors\",\n" +
+                "            \"directory\": \"text_encoders\"\n" +
+                "          }\n" +
+                "        ]\n" +
+                "      }\n" +
+                "    }\n" +
+                "  ]\n" +
+                "}";
+
+        List<ModelInfo> models = analyzer.analyze(json, "test.json");
+
+        // Erwartung: Nur EIN Eintrag für qwen_3_4b.safetensors
+        // Der Name darf NICHT die URL sein.
+        long count = models.stream()
+                .filter(m -> m.getName().toLowerCase().contains("qwen"))
+                .count();
+
+        for (ModelInfo m : models) {
+            System.out.println("Found: " + m.getName() + " [" + m.getType() + "] URL: " + m.getUrl());
+        }
+
+        assertEquals(1, count, "Should only find one entry for qwen, not a second one with the URL as name");
+        assertEquals("qwen_3_4b.safetensors", models.get(0).getName());
+        assertEquals("text_encoders", models.get(0).getType());
+    }
+}