Spaces:
Sleeping
Sleeping
| from typing import List, Dict, Tuple, Any | |
| from utils import get_confidence_color, get_confidence_bg_color | |
| # Method colors (consistent across themes) | |
| METHOD_COLORS = { | |
| "base": "#f39c12", # Orange | |
| "voyage": "#3498db", # Blue | |
| "chicory": "#9b59b6", # Purple | |
| "openai": "#2ecc71", # Green | |
| "expanded": "#e74c3c", # Red | |
| "hybrid": "#1abc9c", # Turquoise | |
| "categories": "#1abc9c", # Same as hybrid | |
| "ingredients": "#f39c12" # Same as base | |
| } | |
| # Method display names | |
| METHOD_NAMES = { | |
| "base": "Base Embeddings", | |
| "voyage": "Voyage AI Reranker", | |
| "chicory": "Chicory Parser", | |
| "openai": "OpenAI Reranker", | |
| "expanded": "Expanded Description", | |
| "hybrid": "Hybrid Matching", | |
| "categories": "Category Matches", | |
| "ingredients": "Ingredient Matches" | |
| } | |
| def parse_result_item(item): | |
| """Helper function to parse result items into display text and score""" | |
| display_text = str(item) # Default | |
| score = 0.0 | |
| # Handle tuple formats | |
| if isinstance(item, tuple): | |
| if len(item) == 2: | |
| match, score = item | |
| display_text = str(match) | |
| elif len(item) == 3: | |
| id_val, text, score = item | |
| display_text = f"<strong>{id_val}</strong>: {text}" if text else str(id_val) | |
| # Handle dictionary formats | |
| elif isinstance(item, dict): | |
| if "name" in item and "score" in item: | |
| display_text = item["name"] | |
| score = item["score"] | |
| elif "category" in item and "confidence" in item: | |
| display_text = item["category"] | |
| score = item["confidence"] | |
| elif "ingredient" in item and "relevance_score" in item: | |
| display_text = item["ingredient"] | |
| score = item["relevance_score"] | |
| # Ensure score is a float | |
| try: | |
| score = float(score) | |
| except (ValueError, TypeError): | |
| score = 0.0 | |
| return display_text, score | |
| def filter_results_by_threshold(results, confidence_threshold=0.0): | |
| """Helper function to filter results by confidence threshold""" | |
| filtered_results = [] | |
| for item in results: | |
| _, score = parse_result_item(item) | |
| if score >= confidence_threshold: | |
| filtered_results.append(item) | |
| return filtered_results | |
| def format_confidence_badge(score): | |
| """Formats the confidence score as a styled span (badge)""" | |
| confidence_percent = int(score * 100) | |
| confidence_color = get_confidence_color(score) | |
| bg_color = get_confidence_bg_color(score) | |
| # Determine text color based on background lightness for better contrast | |
| # Simple heuristic: if bg_color is light (e.g., yellow, light green), use black text, otherwise use white/light text. | |
| # This is approximate. A proper solution would involve calculating luminance. | |
| light_bgs = ["#ffffcc", "#ccffcc", "#cceeff"] # Add more light hex codes if needed | |
| text_color = "#000000" if bg_color.lower() in light_bgs else "var(--text-color)" # Default to theme text color otherwise | |
| return ( | |
| f"<span style='background-color: {bg_color}; border: 1px solid {confidence_color}; color: {text_color}; " | |
| f"font-weight: 600; padding: 3px 8px; border-radius: 5px; font-size: 0.9em; " | |
| f"min-width: 80px; text-align: center; display: inline-block; margin-left: 10px;'>" | |
| f"{confidence_percent}%</span>" | |
| ) | |
| def format_result_list_html(results, confidence_threshold=0.0): | |
| """Formats a list of results (matches) into an HTML list.""" | |
| filtered_results = filter_results_by_threshold(results, confidence_threshold) | |
| if not filtered_results: | |
| return "<p style='color: grey; font-style: italic; margin: 10px 0;'>No matches found above confidence threshold.</p>" | |
| html = "<ul style='list-style-type: none; padding-left: 0; margin-top: 10px;'>" | |
| for item in filtered_results: | |
| display_text, score = parse_result_item(item) | |
| badge = format_confidence_badge(score) | |
| # Reintroduce subtle background and border using theme variables | |
| html += ( | |
| f"<li style='display: flex; justify-content: space-between; align-items: center; " | |
| f"margin-bottom: 8px; padding: 8px; border-radius: 4px; " | |
| f"background-color: var(--secondary-background-color); " # Use secondary background | |
| f"border: 1px solid rgba(128, 128, 128, 0.2);'>" # Subtle grey border | |
| f"<span style='flex-grow: 1; margin-right: 10px; word-wrap: break-word; color: var(--text-color);'>{display_text}</span>" # Ensure text uses theme color | |
| f"{badge}" | |
| f"</li>" | |
| ) | |
| html += "</ul>" | |
| return html | |
| def format_result_card(title, content_html): | |
| """Creates a basic card structure for a single product result.""" | |
| # Use secondary background for the card, slightly stronger border | |
| return ( | |
| f"<div style='margin-bottom: 20px; border: 1px solid rgba(128, 128, 128, 0.3); border-radius: 8px; padding: 15px; background-color: var(--secondary-background-color);'>" | |
| f"<h3 style='margin-top: 0; margin-bottom: 15px; font-size: 1.1em; border-bottom: 1px solid rgba(128, 128, 128, 0.2); padding-bottom: 10px; color: var(--text-color);'>{title}</h3>" # Ensure header uses theme text color | |
| f"{content_html}" | |
| f"</div>" | |
| ) | |
| def format_info_panel(title, text): | |
| """Formats an informational panel (e.g., for expanded description).""" | |
| # Use a slightly different background, maybe derived from primary color with transparency | |
| # Or stick to secondary background for consistency | |
| return ( | |
| f"<div style='border-left: 4px solid var(--primary-color); padding: 10px 15px; margin-bottom: 15px; border-radius: 4px; background-color: var(--secondary-background-color);'>" | |
| f"<h4 style='margin-top: 0; margin-bottom: 8px; font-size: 1em; color: var(--text-color);'>{title}</h4>" # Use theme text color | |
| f"<p style='margin-bottom: 0; font-size: 0.95em; color: var(--text-color);'>{text}</p>" # Use theme text color | |
| f"</div>" | |
| ) | |
| def format_method_results_section(method_key, results, confidence_threshold=0.0): | |
| """Formats results for a specific method within a comparison.""" | |
| method_name = METHOD_NAMES.get(method_key, method_key.replace('_', ' ').title()) | |
| color_hex = METHOD_COLORS.get(method_key, "var(--text-color)") # Fallback to theme text color | |
| results_html = format_result_list_html(results, confidence_threshold) | |
| # Keep the method-specific color for the header border | |
| return ( | |
| f"<div style='margin-bottom: 15px; padding-bottom: 15px; border-bottom: 1px dashed rgba(128, 128, 128, 0.2);'>" # Subtle dashed border between methods | |
| f"<h4 style='margin-top: 0; margin-bottom: 10px; color: {color_hex}; border-bottom: 2px solid {color_hex}; padding-bottom: 5px; display: inline-block;'>{method_name}</h4>" | |
| f"{results_html}" | |
| f"</div>" | |
| ) | |
| # --- Main Formatting Functions Called by UI Tabs --- | |
| def format_comparison_html(product, method_results, expanded_description="", confidence_threshold=0.5): | |
| """Formats the comparison results for multiple methods for one product.""" | |
| content_html = "" | |
| if expanded_description: | |
| content_html += format_info_panel("Expanded Description", expanded_description) | |
| method_order = ["base", "voyage", "chicory", "openai"] | |
| sections_html = [] | |
| for method_key in method_order: | |
| if method_key in method_results and method_results[method_key]: | |
| sections_html.append(format_method_results_section( | |
| method_key=method_key, | |
| results=method_results.get(method_key, []), | |
| confidence_threshold=confidence_threshold | |
| )) | |
| # Join sections, remove last border if sections exist | |
| if sections_html: | |
| # Remove the border-bottom style from the last section's div | |
| last_section = sections_html[-1] | |
| if "border-bottom: 1px dashed rgba(128, 128, 128, 0.2);" in last_section: | |
| sections_html[-1] = last_section.replace("border-bottom: 1px dashed rgba(128, 128, 128, 0.2);", "") | |
| content_html += "".join(sections_html) | |
| return format_result_card(title=product, content_html=content_html) | |
| def format_reranking_results_html(results, match_type="ingredients", show_scores=True, include_explanation=False, | |
| method="voyage", confidence_threshold=0.0): | |
| """Formats results from reranking methods (Voyage, OpenAI).""" | |
| if not results: | |
| return f"<p style='color: grey; font-style: italic;'>No {match_type.lower()} matches found.</p>" | |
| html_elements = [] | |
| for result in results: | |
| product_name = result.get("product_name", "Unknown Product") | |
| matching_items = result.get("matching_items", []) | |
| item_scores = result.get("item_scores", []) | |
| explanation = result.get("explanation", "") if include_explanation else "" | |
| if len(item_scores) != len(matching_items): | |
| item_scores = [result.get("confidence", 0.0)] * len(matching_items) | |
| formatted_matches = [] | |
| for i, item in enumerate(matching_items): | |
| if ":" in str(item) and match_type == "categories": | |
| parts = str(item).split(":", 1) | |
| id_val = parts[0].strip() | |
| text = parts[1].strip() if len(parts) > 1 else "" | |
| formatted_matches.append((id_val, text, item_scores[i])) | |
| else: | |
| formatted_matches.append((str(item), item_scores[i])) | |
| content_html = "" | |
| if explanation: | |
| content_html += format_info_panel("Expanded Description", explanation) | |
| list_html = format_result_list_html(formatted_matches, confidence_threshold) | |
| content_html += list_html | |
| # Only add card if there's content to show (explanation or non-empty list html) | |
| if explanation or "<li" in list_html: # Check if list contains items | |
| html_elements.append(format_result_card(title=product_name, content_html=content_html)) | |
| return "".join(html_elements) | |
| def format_categories_html(product, categories, chicory_result=None, explanation="", match_type="categories", confidence_threshold=0.0): | |
| """Formats category or ingredient matching results (non-reranked).""" | |
| content_html = "" | |
| has_content = False | |
| if explanation: | |
| content_html += format_info_panel("Expanded Description", explanation) | |
| has_content = True | |
| # Add Chicory results if available and applicable | |
| if chicory_result and match_type == "ingredients": | |
| chicory_title = METHOD_NAMES.get("chicory", "Chicory Parser") | |
| content_html += f"<h4 style='margin-top: 15px; margin-bottom: 10px; color: {METHOD_COLORS.get('chicory', '#9b59b6')};'>{chicory_title}</h4>" | |
| if isinstance(chicory_result, dict): | |
| ingredient = chicory_result.get("ingredient", "Not found") | |
| score = chicory_result.get("confidence", 0) | |
| badge = format_confidence_badge(score) | |
| content_html += ( | |
| f"<div style='display: flex; justify-content: space-between; align-items: center; padding: 8px; border-radius: 4px; margin-bottom: 15px; background-color: var(--secondary-background-color); border: 1px solid rgba(128, 128, 128, 0.2);'>" # Use theme bg/border | |
| f"<span style='flex-grow: 1; margin-right: 10px; color: var(--text-color);'>{ingredient}</span>" # Use theme text color | |
| f"{badge}" | |
| f"</div>" | |
| ) | |
| has_content = True | |
| else: | |
| content_html += "<p style='color: grey; font-style: italic;'>No Chicory results available.</p>" | |
| # Add the main category/ingredient results | |
| match_title = METHOD_NAMES.get(match_type, match_type.capitalize()) | |
| color_hex = METHOD_COLORS.get(match_type, "var(--text-color)") # Fallback to theme text color | |
| content_html += f"<h4 style='margin-top: 15px; margin-bottom: 10px; color: {color_hex};'>{match_title}</h4>" | |
| list_html = format_result_list_html(categories, confidence_threshold) | |
| content_html += list_html | |
| if "<li" in list_html: # Check if list contains items | |
| has_content = True | |
| # Only return card if there's content to show | |
| if has_content: | |
| return format_result_card(title=product, content_html=content_html) | |
| else: | |
| return "" # Return empty string if nothing to show | |