code reviewed

backend/services/image_similarity.py

@@ -0,0 +1,48 @@
+import os
+import numpy as np
+import faiss
+import torch
+from PIL import Image
+from transformers import CLIPModel
+from torchvision import transforms
+
+# Device setup
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# Global variable for lazy loading
+_model = None
+_preprocess = None
+
+def get_model():
+    global _model
+    if _model is None:
+        _model = CLIPModel.from_pretrained('openai/clip-vit-base-patch16').to(device)
+        _model.eval()
+    return _model
+
+def get_preprocess():
+    global _preprocess
+    if _preprocess is None:
+        _preprocess = transforms.Compose([
+            transforms.Resize((224, 224)),
+            transforms.ToTensor(),
+            transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
+        ])
+    return _preprocess
+
+def transform_image(image):
+    model = get_model()
+    preprocess = get_preprocess()
+    image = preprocess(image).unsqueeze(0).to(device)
+    with torch.no_grad():
+        features = model.get_image_features(pixel_values=image)
+    return features.squeeze().cpu().numpy().astype(np.float32)
+
+def load_index(index_file):
+    return faiss.read_index(index_file)
+
+def search_similar_images(query_image, index, top_k=5):
+    query_embedding = transform_image(query_image)
+    query_embedding_normalized = query_embedding / np.linalg.norm(query_embedding)
+    distances, indices = index.search(np.array([query_embedding_normalized]), top_k)
+    return distances[0][:top_k], indices[0][:top_k] 

backend/services/image_utils.py

@@ -0,0 +1,16 @@
+import requests
+from PIL import Image
+from io import BytesIO
+
+
+def download_image(image_url: str, save_path: str) -> None:
+    """
+    Download an image from a URL and save it to a local path.
+    Args:
+        image_url (str): The URL of the image to download.
+        save_path (str): The local file path to save the image.
+    """
+    response = requests.get(image_url)
+    response.raise_for_status()
+    image = Image.open(BytesIO(response.content))
+    image.save(save_path) 

backend/services/result_aggregation.py

@@ -0,0 +1,58 @@
+import pandas as pd
+
+def aggregate_results(matched_names, expert_data, community_data, max_results=30):
+    """
+    Aggregate similar images, appraisal values, years, and status for matched tag names.
+    Args:
+        matched_names (list): List of tag names to match.
+        expert_data (pd.DataFrame): Expert dataset.
+        community_data (pd.DataFrame): Community dataset.
+        max_results (int): Maximum number of results to return.
+    Returns:
+        dict: Aggregated results with images, appraisal values, years, and status.
+    """
+    similar_data = []
+    
+    for title in matched_names:
+        # Handle expert data (no 'year' column)
+        community_items = community_data[community_data['brand_name'] == title]
+        expert_items = expert_data[expert_data['brand_name'] == title]
+        
+        # Process community data (has 'year' column)
+        if not community_items.empty:
+            # Use 'year' column if it exists, otherwise use 'year_start'
+            year_col = 'year' if 'year' in community_items.columns else 'year_start'
+            community_records = community_items[['front_tag', 'appraisal_value', 'key', 'status', year_col]].to_dict('records')
+            # Rename year column to 'year' for consistency
+            for record in community_records:
+                record['year'] = record.pop(year_col) if year_col in record else None
+            similar_data.extend(community_records)
+        
+        # Process expert data (no 'year' column)
+        if not expert_items.empty:
+            expert_records = expert_items[['front_tag', 'appraisal_value', 'key', 'status']].to_dict('records')
+            # Add None for year since expert data doesn't have it
+            for record in expert_records:
+                record['year'] = None
+            similar_data.extend(expert_records)
+    
+    # Remove duplicates by key, preserving order
+    seen_keys = set()
+    unique_data = []
+    for item in similar_data:
+        if item['key'] not in seen_keys:
+            seen_keys.add(item['key'])
+            unique_data.append(item)
+    
+    # Prepare results
+    similar_images = [item['front_tag'] for item in unique_data][:max_results]
+    appraisal_values = [item['appraisal_value'] for item in unique_data][:max_results]
+    years = [item.get('year') for item in unique_data][:max_results]
+    statuses = [item['status'] for item in unique_data][:max_results]
+    
+    return {
+        'similar_images': similar_images,
+        'appraisal_values': appraisal_values,
+        'years': years,
+        'statuses': statuses
+    }