aidp 0.27.0 → 0.28.0

data/lib/aidp/planning/analyzers/feedback_analyzer.rb

@@ -0,0 +1,365 @@
+# frozen_string_literal: true
+
+require_relative "../../logger"
+
+module Aidp
+  module Planning
+    module Analyzers
+      # Analyzes user feedback using AI to extract insights and recommendations
+      # Uses Zero Framework Cognition (ZFC) - NO regex, heuristics, or keyword matching
+      # All semantic analysis delegated to AI Decision Engine
+      class FeedbackAnalyzer
+        def initialize(ai_decision_engine:, config: nil)
+          @ai_decision_engine = ai_decision_engine
+          @config = config || Aidp::Config.agile_config
+        end
+
+        # Analyze feedback data and generate insights
+        # @param feedback_data [Hash] Parsed feedback data from FeedbackDataParser
+        # @return [Hash] Analysis results with findings and recommendations
+        def analyze(feedback_data)
+          Aidp.log_debug("feedback_analyzer", "analyze",
+            response_count: feedback_data[:response_count],
+            format: feedback_data[:format])
+
+          # Use AI to perform semantic analysis
+          analysis = analyze_with_ai(feedback_data)
+
+          {
+            summary: analysis[:summary],
+            findings: analysis[:findings],
+            trends: analysis[:trends],
+            insights: analysis[:insights],
+            sentiment_breakdown: analysis[:sentiment_breakdown],
+            feature_feedback: analysis[:feature_feedback],
+            recommendations: analysis[:recommendations],
+            priority_issues: analysis[:priority_issues],
+            positive_highlights: analysis[:positive_highlights],
+            metadata: {
+              generated_at: Time.now.iso8601,
+              responses_analyzed: feedback_data[:response_count],
+              source_file: feedback_data[:source_file],
+              source_format: feedback_data[:format]
+            }
+          }
+        end
+
+        # Format feedback analysis as markdown
+        # @param analysis [Hash] Feedback analysis structure
+        # @return [String] Markdown formatted analysis
+        def format_as_markdown(analysis)
+          Aidp.log_debug("feedback_analyzer", "format_as_markdown")
+
+          output = ["# User Feedback Analysis", ""]
+          output << "**Generated:** #{analysis[:metadata][:generated_at]}"
+          output << "**Responses Analyzed:** #{analysis[:metadata][:responses_analyzed]}"
+          output << "**Source:** #{analysis[:metadata][:source_file]}"
+          output << ""
+
+          output << "## Executive Summary"
+          output << ""
+          output << analysis[:summary]
+          output << ""
+
+          output << "## Sentiment Breakdown"
+          output << ""
+          output << "| Sentiment | Count | Percentage |"
+          output << "|-----------|-------|------------|"
+          analysis[:sentiment_breakdown].each do |sentiment|
+            output << "| #{sentiment[:type]} | #{sentiment[:count]} | #{sentiment[:percentage]}% |"
+          end
+          output << ""
+
+          output << "## Key Findings"
+          output << ""
+          analysis[:findings].each_with_index do |finding, idx|
+            output << "### #{idx + 1}. #{finding[:title]}"
+            output << ""
+            output << finding[:description]
+            output << ""
+            output << "**Evidence:**"
+            finding[:evidence].each do |evidence|
+              output << "- #{evidence}"
+            end
+            output << ""
+            output << "**Impact:** #{finding[:impact]}"
+            output << ""
+          end
+
+          output << "## Trends and Patterns"
+          output << ""
+          analysis[:trends].each_with_index do |trend, idx|
+            output << "### #{idx + 1}. #{trend[:title]}"
+            output << ""
+            output << trend[:description]
+            output << ""
+            output << "**Frequency:** #{trend[:frequency]}"
+            output << ""
+            output << "**Implication:** #{trend[:implication]}"
+            output << ""
+          end
+
+          output << "## Insights"
+          output << ""
+          analysis[:insights].each do |insight|
+            output << "- **#{insight[:category]}:** #{insight[:description]}"
+            output << ""
+          end
+
+          output << "## Feature-Specific Feedback"
+          output << ""
+          analysis[:feature_feedback].each do |feedback|
+            output << "### #{feedback[:feature_name]}"
+            output << ""
+            output << "**Overall Sentiment:** #{feedback[:sentiment]}"
+            output << ""
+            output << "**Positive Feedback:**"
+            feedback[:positive].each do |pos|
+              output << "- #{pos}"
+            end
+            output << ""
+            output << "**Negative Feedback:**"
+            feedback[:negative].each do |neg|
+              output << "- #{neg}"
+            end
+            output << ""
+            output << "**Suggested Improvements:**"
+            feedback[:improvements].each do |imp|
+              output << "- #{imp}"
+            end
+            output << ""
+          end
+
+          output << "## Priority Issues"
+          output << ""
+          output << "Issues requiring immediate attention:"
+          output << ""
+          analysis[:priority_issues].each_with_index do |issue, idx|
+            output << "#{idx + 1}. **#{issue[:title]}** (Priority: #{issue[:priority]})"
+            output << "   - Impact: #{issue[:impact]}"
+            output << "   - Affected Users: #{issue[:affected_users]}"
+            output << "   - Recommended Action: #{issue[:action]}"
+            output << ""
+          end
+
+          output << "## Positive Highlights"
+          output << ""
+          output << "What users loved:"
+          output << ""
+          analysis[:positive_highlights].each do |highlight|
+            output << "- #{highlight}"
+          end
+          output << ""
+
+          output << "## Recommendations"
+          output << ""
+          analysis[:recommendations].each_with_index do |rec, idx|
+            output << "### #{idx + 1}. #{rec[:title]}"
+            output << ""
+            output << rec[:description]
+            output << ""
+            output << "**Rationale:** #{rec[:rationale]}"
+            output << ""
+            output << "**Effort:** #{rec[:effort]}"
+            output << ""
+            output << "**Expected Impact:** #{rec[:impact]}"
+            output << ""
+          end
+
+          output.join("\n")
+        end
+
+        private
+
+        def analyze_with_ai(feedback_data)
+          Aidp.log_debug("feedback_analyzer", "analyze_with_ai")
+
+          prompt = build_analysis_prompt(feedback_data)
+
+          schema = {
+            type: "object",
+            properties: {
+              summary: {type: "string"},
+              findings: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    title: {type: "string"},
+                    description: {type: "string"},
+                    evidence: {type: "array", items: {type: "string"}},
+                    impact: {type: "string"}
+                  }
+                }
+              },
+              trends: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    title: {type: "string"},
+                    description: {type: "string"},
+                    frequency: {type: "string"},
+                    implication: {type: "string"}
+                  }
+                }
+              },
+              insights: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    category: {type: "string"},
+                    description: {type: "string"}
+                  }
+                }
+              },
+              sentiment_breakdown: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    type: {type: "string"},
+                    count: {type: "integer"},
+                    percentage: {type: "number"}
+                  }
+                }
+              },
+              feature_feedback: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    feature_name: {type: "string"},
+                    sentiment: {type: "string"},
+                    positive: {type: "array", items: {type: "string"}},
+                    negative: {type: "array", items: {type: "string"}},
+                    improvements: {type: "array", items: {type: "string"}}
+                  }
+                }
+              },
+              recommendations: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    title: {type: "string"},
+                    description: {type: "string"},
+                    rationale: {type: "string"},
+                    effort: {type: "string"},
+                    impact: {type: "string"}
+                  }
+                }
+              },
+              priority_issues: {
+                type: "array",
+                items: {
+                  type: "object",
+                  properties: {
+                    title: {type: "string"},
+                    priority: {type: "string"},
+                    impact: {type: "string"},
+                    affected_users: {type: "string"},
+                    action: {type: "string"}
+                  }
+                }
+              },
+              positive_highlights: {
+                type: "array",
+                items: {type: "string"}
+              }
+            },
+            required: ["summary", "findings", "recommendations"]
+          }
+
+          decision = @ai_decision_engine.decide(
+            context: "feedback_analysis",
+            prompt: prompt,
+            data: feedback_data,
+            schema: schema
+          )
+
+          Aidp.log_debug("feedback_analyzer", "ai_analysis_complete",
+            findings: decision[:findings]&.size || 0,
+            recommendations: decision[:recommendations]&.size || 0)
+
+          decision
+        end
+
+        def build_analysis_prompt(feedback_data)
+          # Format response data for AI analysis
+          responses_text = feedback_data[:responses]&.map&.with_index { |r, i|
+            <<~RESPONSE
+              Response #{i + 1}:
+              - ID: #{r[:respondent_id]}
+              - Timestamp: #{r[:timestamp]}
+              - Rating: #{r[:rating]}
+              - Feature: #{r[:feature]}
+              - Feedback: #{r[:feedback_text]}
+              - Sentiment: #{r[:sentiment]}
+              - Tags: #{r[:tags]&.join(", ")}
+            RESPONSE
+          }&.join("\n") || "No responses"
+
+          <<~PROMPT
+            Analyze the following user feedback data and extract insights, trends, and recommendations.
+
+            FEEDBACK DATA:
+            Total Responses: #{feedback_data[:response_count]}
+            Source: #{feedback_data[:source_file]}
+            Format: #{feedback_data[:format]}
+
+            RESPONSES:
+            #{responses_text}
+
+            TASK:
+            Perform comprehensive feedback analysis:
+
+            1. SUMMARY
+               - High-level overview of feedback (2-3 paragraphs)
+               - Overall sentiment and key themes
+
+            2. KEY FINDINGS (3-5 findings)
+               - Important discoveries from the data
+               - Evidence supporting each finding
+               - Impact assessment (high/medium/low)
+
+            3. TRENDS AND PATTERNS
+               - Recurring themes
+               - Frequency of occurrence
+               - Implications for product development
+
+            4. INSIGHTS
+               - Categorized insights (usability, features, performance, etc.)
+               - Actionable observations
+
+            5. SENTIMENT BREAKDOWN
+               - Distribution of positive, negative, neutral sentiment
+               - Counts and percentages
+
+            6. FEATURE-SPECIFIC FEEDBACK
+               - For each mentioned feature: sentiment, positive/negative feedback, improvements
+
+            7. PRIORITY ISSUES
+               - Critical issues requiring immediate attention
+               - Priority level (critical/high/medium)
+               - Affected user count
+               - Recommended action
+
+            8. POSITIVE HIGHLIGHTS
+               - What users loved
+               - Strengths to maintain or amplify
+
+            9. RECOMMENDATIONS (4-6 recommendations)
+               - Specific, actionable recommendations
+               - Rationale based on feedback
+               - Effort estimate (low/medium/high)
+               - Expected impact
+
+            Use semantic analysis to understand context and meaning. Look beyond keywords to understand user intent and emotion.
+          PROMPT
+        end
+      end
+    end
+  end
+end