legal_summariser 0.3.0 → 0.3.1

data/lib/legal_summariser/pdf_annotator.rb

@@ -0,0 +1,601 @@
+require 'json'
+require 'fileutils'
+require 'digest'
+
+module LegalSummariser
+  # Advanced PDF annotation system for legal document analysis output
+  class PDFAnnotator
+    class AnnotationError < StandardError; end
+    class PDFError < StandardError; end
+    class UnsupportedFormatError < StandardError; end
+
+    # Annotation types supported
+    ANNOTATION_TYPES = {
+      highlight: {
+        color: '#FFFF00',
+        opacity: 0.3,
+        description: 'Highlighted text'
+      },
+      note: {
+        color: '#FFA500',
+        opacity: 0.8,
+        description: 'Sticky note annotation'
+      },
+      warning: {
+        color: '#FF6B6B',
+        opacity: 0.5,
+        description: 'Warning or risk indicator'
+      },
+      important: {
+        color: '#4ECDC4',
+        opacity: 0.4,
+        description: 'Important clause or section'
+      },
+      question: {
+        color: '#95E1D3',
+        opacity: 0.4,
+        description: 'Question or clarification needed'
+      },
+      summary: {
+        color: '#A8E6CF',
+        opacity: 0.3,
+        description: 'Summary or key point'
+      }
+    }.freeze
+
+    # Risk level color coding
+    RISK_COLORS = {
+      high: '#FF4757',
+      medium: '#FFA502',
+      low: '#2ED573',
+      info: '#3742FA'
+    }.freeze
+
+    attr_reader :config, :logger, :annotations_dir
+
+    def initialize(config = nil)
+      @config = config || LegalSummariser.configuration
+      @logger = @config.logger
+      @annotations_dir = File.join(@config.cache_dir, 'pdf_annotations')
+      
+      setup_directories
+    end
+
+    # Create annotated PDF from analysis results
+    def create_annotated_pdf(pdf_path, analysis_results, output_path, options = {})
+      validate_pdf_path(pdf_path)
+      
+      @logger&.info("Creating annotated PDF from #{pdf_path}")
+      
+      begin
+        # Extract text positions from PDF
+        text_positions = extract_text_positions(pdf_path)
+        
+        # Generate annotations from analysis results
+        annotations = generate_annotations_from_analysis(analysis_results, text_positions, options)
+        
+        # Create annotated PDF
+        annotated_pdf_data = create_pdf_with_annotations(pdf_path, annotations, options)
+        
+        # Save annotated PDF
+        File.write(output_path, annotated_pdf_data, mode: 'wb')
+        
+        # Save annotation metadata
+        save_annotation_metadata(output_path, annotations, analysis_results)
+        
+        @logger&.info("Annotated PDF created: #{output_path}")
+        
+        {
+          input_pdf: pdf_path,
+          output_pdf: output_path,
+          annotations_count: annotations.length,
+          annotation_types: annotations.group_by { |a| a[:type] }.transform_values(&:count),
+          metadata_file: get_metadata_path(output_path)
+        }
+        
+      rescue => e
+        @logger&.error("PDF annotation failed: #{e.message}")
+        raise AnnotationError, "Failed to create annotated PDF: #{e.message}"
+      end
+    end
+
+    # Add custom annotations to PDF
+    def add_custom_annotations(pdf_path, custom_annotations, output_path, options = {})
+      validate_pdf_path(pdf_path)
+      validate_annotations(custom_annotations)
+      
+      @logger&.info("Adding #{custom_annotations.length} custom annotations to PDF")
+      
+      begin
+        # Process custom annotations
+        processed_annotations = process_custom_annotations(custom_annotations, options)
+        
+        # Create annotated PDF
+        annotated_pdf_data = create_pdf_with_annotations(pdf_path, processed_annotations, options)
+        
+        # Save result
+        File.write(output_path, annotated_pdf_data, mode: 'wb')
+        save_annotation_metadata(output_path, processed_annotations, { custom: true })
+        
+        {
+          input_pdf: pdf_path,
+          output_pdf: output_path,
+          custom_annotations: processed_annotations.length
+        }
+        
+      rescue => e
+        @logger&.error("Custom annotation failed: #{e.message}")
+        raise AnnotationError, "Failed to add custom annotations: #{e.message}"
+      end
+    end
+
+    # Extract annotations from an annotated PDF
+    def extract_annotations(pdf_path)
+      validate_pdf_path(pdf_path)
+      
+      @logger&.info("Extracting annotations from #{pdf_path}")
+      
+      begin
+        # Check for metadata file first
+        metadata_path = get_metadata_path(pdf_path)
+        
+        if File.exist?(metadata_path)
+          metadata = JSON.parse(File.read(metadata_path))
+          return metadata['annotations'] || []
+        end
+        
+        # Fallback: try to extract from PDF directly
+        extract_annotations_from_pdf(pdf_path)
+        
+      rescue => e
+        @logger&.error("Annotation extraction failed: #{e.message}")
+        raise AnnotationError, "Failed to extract annotations: #{e.message}"
+      end
+    end
+
+    # Generate annotation report
+    def generate_annotation_report(pdf_path, format = :json)
+      annotations = extract_annotations(pdf_path)
+      
+      case format
+      when :json
+        generate_json_report(annotations)
+      when :markdown
+        generate_markdown_report(annotations)
+      when :html
+        generate_html_report(annotations)
+      else
+        raise UnsupportedFormatError, "Unsupported report format: #{format}"
+      end
+    end
+
+    # Merge multiple annotated PDFs
+    def merge_annotated_pdfs(pdf_paths, output_path, options = {})
+      @logger&.info("Merging #{pdf_paths.length} annotated PDFs")
+      
+      begin
+        merged_annotations = []
+        page_offset = 0
+        
+        pdf_paths.each_with_index do |pdf_path, index|
+          validate_pdf_path(pdf_path)
+          
+          # Extract annotations and adjust page numbers
+          annotations = extract_annotations(pdf_path)
+          annotations.each do |annotation|
+            annotation[:page] += page_offset if annotation[:page]
+            annotation[:source_pdf] = File.basename(pdf_path)
+            merged_annotations << annotation
+          end
+          
+          # Get page count for offset calculation
+          page_count = get_pdf_page_count(pdf_path)
+          page_offset += page_count
+        end
+        
+        # Create merged PDF (placeholder implementation)
+        create_merged_pdf(pdf_paths, output_path, merged_annotations, options)
+        
+        {
+          merged_pdf: output_path,
+          source_pdfs: pdf_paths.length,
+          total_annotations: merged_annotations.length
+        }
+        
+      rescue => e
+        @logger&.error("PDF merging failed: #{e.message}")
+        raise AnnotationError, "Failed to merge annotated PDFs: #{e.message}"
+      end
+    end
+
+    # Get annotation statistics
+    def get_annotation_statistics(pdf_path)
+      annotations = extract_annotations(pdf_path)
+      
+      {
+        total_annotations: annotations.length,
+        by_type: annotations.group_by { |a| a[:type] }.transform_values(&:count),
+        by_risk_level: annotations.select { |a| a[:risk_level] }
+                                 .group_by { |a| a[:risk_level] }
+                                 .transform_values(&:count),
+        by_page: annotations.group_by { |a| a[:page] }.transform_values(&:count),
+        coverage: calculate_annotation_coverage(annotations),
+        summary: generate_annotation_summary(annotations)
+      }
+    end
+
+    private
+
+    def setup_directories
+      FileUtils.mkdir_p(@annotations_dir) unless Dir.exist?(@annotations_dir)
+    end
+
+    def validate_pdf_path(pdf_path)
+      raise PDFError, "PDF file not found: #{pdf_path}" unless File.exist?(pdf_path)
+      raise PDFError, "Invalid PDF file: #{pdf_path}" unless pdf_path.downcase.end_with?('.pdf')
+    end
+
+    def validate_annotations(annotations)
+      raise AnnotationError, "Annotations must be an array" unless annotations.is_a?(Array)
+      
+      annotations.each_with_index do |annotation, index|
+        unless annotation.is_a?(Hash)
+          raise AnnotationError, "Annotation #{index} must be a hash"
+        end
+        
+        required_keys = [:type, :text, :page]
+        missing_keys = required_keys - annotation.keys
+        
+        unless missing_keys.empty?
+          raise AnnotationError, "Annotation #{index} missing keys: #{missing_keys.join(', ')}"
+        end
+      end
+    end
+
+    def extract_text_positions(pdf_path)
+      # Placeholder implementation for text position extraction
+      # In a real implementation, you would use a PDF library like PDF::Reader
+      # to extract text positions and coordinates
+      
+      @logger&.info("Extracting text positions from PDF (placeholder)")
+      
+      # Simulated text positions
+      {
+        pages: [
+          {
+            page_number: 1,
+            width: 612,
+            height: 792,
+            text_blocks: [
+              {
+                text: "Sample contract text",
+                x: 72,
+                y: 720,
+                width: 200,
+                height: 20
+              }
+            ]
+          }
+        ]
+      }
+    end
+
+    def generate_annotations_from_analysis(analysis_results, text_positions, options = {})
+      annotations = []
+      
+      # Generate annotations from summary
+      if analysis_results[:summary]
+        summary_annotations = create_summary_annotations(analysis_results[:summary], text_positions)
+        annotations.concat(summary_annotations)
+      end
+      
+      # Generate annotations from risks
+      if analysis_results[:risks]
+        risk_annotations = create_risk_annotations(analysis_results[:risks], text_positions)
+        annotations.concat(risk_annotations)
+      end
+      
+      # Generate annotations from clauses
+      if analysis_results[:clauses]
+        clause_annotations = create_clause_annotations(analysis_results[:clauses], text_positions)
+        annotations.concat(clause_annotations)
+      end
+      
+      # Generate annotations from plain language suggestions
+      if analysis_results[:plain_language]
+        plain_language_annotations = create_plain_language_annotations(analysis_results[:plain_language], text_positions)
+        annotations.concat(plain_language_annotations)
+      end
+      
+      annotations
+    end
+
+    def create_summary_annotations(summary_data, text_positions)
+      annotations = []
+      
+      if summary_data[:key_points]
+        summary_data[:key_points].each_with_index do |point, index|
+          annotations << {
+            type: :summary,
+            text: point,
+            note: "Key Point #{index + 1}",
+            page: 1, # Simplified - would need actual text matching
+            color: ANNOTATION_TYPES[:summary][:color],
+            opacity: ANNOTATION_TYPES[:summary][:opacity]
+          }
+        end
+      end
+      
+      annotations
+    end
+
+    def create_risk_annotations(risks_data, text_positions)
+      annotations = []
+      
+      [:high_risks, :medium_risks, :low_risks].each do |risk_level|
+        next unless risks_data[risk_level]
+        
+        level = risk_level.to_s.split('_').first.to_sym
+        
+        risks_data[risk_level].each do |risk|
+          annotations << {
+            type: :warning,
+            text: risk[:text] || risk,
+            note: "#{level.capitalize} Risk: #{risk[:description] || risk}",
+            risk_level: level,
+            page: 1, # Simplified
+            color: RISK_COLORS[level],
+            opacity: 0.6
+          }
+        end
+      end
+      
+      annotations
+    end
+
+    def create_clause_annotations(clauses_data, text_positions)
+      annotations = []
+      
+      clauses_data.each do |clause_type, clauses|
+        next unless clauses.is_a?(Array)
+        
+        clauses.each do |clause|
+          annotations << {
+            type: :important,
+            text: clause[:text] || clause,
+            note: "#{clause_type.to_s.humanize} Clause",
+            clause_type: clause_type,
+            page: 1, # Simplified
+            color: ANNOTATION_TYPES[:important][:color],
+            opacity: ANNOTATION_TYPES[:important][:opacity]
+          }
+        end
+      end
+      
+      annotations
+    end
+
+    def create_plain_language_annotations(plain_language_data, text_positions)
+      annotations = []
+      
+      if plain_language_data[:simplified_text]
+        # Create annotations for complex terms that were simplified
+        annotations << {
+          type: :note,
+          text: "Plain language version available",
+          note: "This document has been converted to plain English. See attached simplified version.",
+          page: 1,
+          color: ANNOTATION_TYPES[:note][:color],
+          opacity: ANNOTATION_TYPES[:note][:opacity]
+        }
+      end
+      
+      annotations
+    end
+
+    def process_custom_annotations(custom_annotations, options = {})
+      processed = []
+      
+      custom_annotations.each do |annotation|
+        processed_annotation = {
+          type: annotation[:type] || :note,
+          text: annotation[:text],
+          note: annotation[:note] || annotation[:comment],
+          page: annotation[:page] || 1,
+          color: annotation[:color] || ANNOTATION_TYPES[annotation[:type] || :note][:color],
+          opacity: annotation[:opacity] || ANNOTATION_TYPES[annotation[:type] || :note][:opacity]
+        }
+        
+        # Add position if provided
+        if annotation[:position]
+          processed_annotation[:position] = annotation[:position]
+        end
+        
+        processed << processed_annotation
+      end
+      
+      processed
+    end
+
+    def create_pdf_with_annotations(pdf_path, annotations, options = {})
+      # Placeholder implementation for PDF annotation
+      # In a real implementation, you would use a PDF library like Prawn or HexaPDF
+      # to add actual annotations to the PDF
+      
+      @logger&.info("Creating PDF with #{annotations.length} annotations (placeholder)")
+      
+      # For now, just copy the original PDF
+      # In practice, this would create a new PDF with annotations overlaid
+      File.read(pdf_path, mode: 'rb')
+    end
+
+    def save_annotation_metadata(pdf_path, annotations, analysis_results)
+      metadata = {
+        pdf_file: File.basename(pdf_path),
+        created_at: Time.now.iso8601,
+        annotations: annotations,
+        analysis_results: analysis_results,
+        annotation_statistics: {
+          total: annotations.length,
+          by_type: annotations.group_by { |a| a[:type] }.transform_values(&:count)
+        }
+      }
+      
+      metadata_path = get_metadata_path(pdf_path)
+      File.write(metadata_path, JSON.pretty_generate(metadata))
+    end
+
+    def get_metadata_path(pdf_path)
+      base_name = File.basename(pdf_path, '.pdf')
+      File.join(@annotations_dir, "#{base_name}_annotations.json")
+    end
+
+    def extract_annotations_from_pdf(pdf_path)
+      # Placeholder for extracting annotations directly from PDF
+      # This would use a PDF library to read existing annotations
+      []
+    end
+
+    def generate_json_report(annotations)
+      {
+        report_type: 'annotation_report',
+        generated_at: Time.now.iso8601,
+        total_annotations: annotations.length,
+        annotations: annotations,
+        statistics: calculate_annotation_statistics(annotations)
+      }.to_json
+    end
+
+    def generate_markdown_report(annotations)
+      report = "# PDF Annotation Report\n\n"
+      report += "Generated: #{Time.now.strftime('%Y-%m-%d %H:%M:%S')}\n\n"
+      report += "Total Annotations: #{annotations.length}\n\n"
+      
+      # Group by type
+      annotations.group_by { |a| a[:type] }.each do |type, type_annotations|
+        report += "## #{type.to_s.capitalize} Annotations (#{type_annotations.length})\n\n"
+        
+        type_annotations.each_with_index do |annotation, index|
+          report += "### #{index + 1}. Page #{annotation[:page]}\n"
+          report += "**Text:** #{annotation[:text]}\n\n"
+          report += "**Note:** #{annotation[:note]}\n\n" if annotation[:note]
+          report += "---\n\n"
+        end
+      end
+      
+      report
+    end
+
+    def generate_html_report(annotations)
+      html = <<~HTML
+        <!DOCTYPE html>
+        <html>
+        <head>
+          <title>PDF Annotation Report</title>
+          <style>
+            body { font-family: Arial, sans-serif; margin: 20px; }
+            .annotation { border: 1px solid #ccc; margin: 10px 0; padding: 10px; }
+            .type-highlight { border-left: 4px solid #FFFF00; }
+            .type-warning { border-left: 4px solid #FF6B6B; }
+            .type-note { border-left: 4px solid #FFA500; }
+            .type-important { border-left: 4px solid #4ECDC4; }
+          </style>
+        </head>
+        <body>
+          <h1>PDF Annotation Report</h1>
+          <p>Generated: #{Time.now.strftime('%Y-%m-%d %H:%M:%S')}</p>
+          <p>Total Annotations: #{annotations.length}</p>
+      HTML
+      
+      annotations.each_with_index do |annotation, index|
+        html += <<~HTML
+          <div class="annotation type-#{annotation[:type]}">
+            <h3>Annotation #{index + 1} - Page #{annotation[:page]}</h3>
+            <p><strong>Type:</strong> #{annotation[:type]}</p>
+            <p><strong>Text:</strong> #{annotation[:text]}</p>
+        HTML
+        
+        if annotation[:note]
+          html += "<p><strong>Note:</strong> #{annotation[:note]}</p>"
+        end
+        
+        html += "</div>"
+      end
+      
+      html += "</body></html>"
+      html
+    end
+
+    def create_merged_pdf(pdf_paths, output_path, annotations, options = {})
+      # Placeholder for PDF merging
+      # In practice, this would use a PDF library to merge PDFs and preserve annotations
+      
+      @logger&.info("Merging PDFs (placeholder implementation)")
+      
+      # For now, just copy the first PDF
+      if pdf_paths.any?
+        FileUtils.cp(pdf_paths.first, output_path)
+      end
+      
+      # Save merged annotations metadata
+      save_annotation_metadata(output_path, annotations, { merged: true, source_pdfs: pdf_paths })
+    end
+
+    def get_pdf_page_count(pdf_path)
+      # Placeholder for getting PDF page count
+      # In practice, this would use a PDF library to count pages
+      10 # Default assumption
+    end
+
+    def calculate_annotation_coverage(annotations)
+      return 0 if annotations.empty?
+      
+      pages_with_annotations = annotations.map { |a| a[:page] }.uniq.length
+      total_pages = annotations.map { |a| a[:page] }.max || 1
+      
+      (pages_with_annotations.to_f / total_pages * 100).round(1)
+    end
+
+    def generate_annotation_summary(annotations)
+      summary = {
+        most_annotated_page: nil,
+        most_common_type: nil,
+        risk_distribution: {},
+        recommendations: []
+      }
+      
+      # Most annotated page
+      page_counts = annotations.group_by { |a| a[:page] }.transform_values(&:count)
+      summary[:most_annotated_page] = page_counts.max_by { |_, count| count }&.first
+      
+      # Most common annotation type
+      type_counts = annotations.group_by { |a| a[:type] }.transform_values(&:count)
+      summary[:most_common_type] = type_counts.max_by { |_, count| count }&.first
+      
+      # Risk distribution
+      risk_annotations = annotations.select { |a| a[:risk_level] }
+      summary[:risk_distribution] = risk_annotations.group_by { |a| a[:risk_level] }
+                                                   .transform_values(&:count)
+      
+      # Generate recommendations
+      if summary[:risk_distribution][:high]&.> 0
+        summary[:recommendations] << "High-risk items require immediate attention"
+      end
+      
+      if type_counts[:warning]&.> 5
+        summary[:recommendations] << "Multiple warnings detected - consider legal review"
+      end
+      
+      summary
+    end
+
+    def calculate_annotation_statistics(annotations)
+      {
+        total: annotations.length,
+        by_type: annotations.group_by { |a| a[:type] }.transform_values(&:count),
+        by_page: annotations.group_by { |a| a[:page] }.transform_values(&:count),
+        with_risks: annotations.count { |a| a[:risk_level] }
+      }
+    end
+  end
+end