@umituz/react-native-ai-generation-content 1.26.2 → 1.26.4

package/src/infrastructure/executors/image-executor.ts

@@ -0,0 +1,133 @@
+/**
+ * Image Generation Executor
+ * Generic executor for all image generation features
+ */
+
+import type {
+  GenerationExecutor,
+  ImageGenerationInput,
+  ImageGenerationOutput,
+  GenerationOptions,
+  GenerationResult,
+} from "../../domains/generation/domain/generation.types";
+import { providerRegistry } from "../services/provider-registry.service";
+
+declare const __DEV__: boolean;
+
+export class ImageExecutor
+  implements GenerationExecutor<ImageGenerationInput, ImageGenerationOutput>
+{
+  async generate(
+    model: string,
+    input: ImageGenerationInput,
+    options?: GenerationOptions,
+  ): Promise<GenerationResult<ImageGenerationOutput>> {
+    try {
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[ImageExecutor] Starting generation", {
+          model,
+          hasImages: !!input.imageUrls,
+          imageCount: input.imageUrls?.length || 0,
+          promptLength: input.prompt?.length || 0,
+        });
+      }
+
+      const provider = providerRegistry.getActiveProvider();
+
+      if (!provider?.isInitialized()) {
+        if (typeof __DEV__ !== "undefined" && __DEV__) {
+          console.error("[ImageExecutor] Provider not initialized");
+        }
+        return { success: false, error: "AI provider not initialized" };
+      }
+
+      options?.onProgress?.(5);
+
+      const modelInput = this.buildModelInput(input);
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[ImageExecutor] Model input prepared", {
+          imageCount: modelInput.image_urls?.length || 0,
+          aspectRatio: modelInput.aspect_ratio,
+          outputFormat: modelInput.output_format,
+        });
+      }
+
+      options?.onProgress?.(10);
+
+      const result = await provider.subscribe(model, modelInput, {
+        timeoutMs: options?.timeoutMs ?? 300000,
+        onQueueUpdate: (status) => {
+          if (typeof __DEV__ !== "undefined" && __DEV__) {
+            console.log("[ImageExecutor] Queue status:", status.status);
+          }
+
+          if (status.status === "IN_QUEUE") {
+            options?.onProgress?.(20);
+          } else if (status.status === "IN_PROGRESS") {
+            options?.onProgress?.(50);
+          }
+        },
+      });
+
+      options?.onProgress?.(90);
+
+      const imageUrl = this.extractImageUrl(result);
+
+      if (!imageUrl) {
+        if (typeof __DEV__ !== "undefined" && __DEV__) {
+          console.error("[ImageExecutor] No image URL in response");
+        }
+        return { success: false, error: "No image generated" };
+      }
+
+      options?.onProgress?.(100);
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[ImageExecutor] Generation successful", {
+          imageUrl: imageUrl.slice(0, 100),
+        });
+      }
+
+      return { success: true, data: { imageUrl } };
+    } catch (error) {
+      const message =
+        error instanceof Error ? error.message : "Generation failed";
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.error("[ImageExecutor] Generation error:", {
+          message,
+          error,
+        });
+      }
+
+      return { success: false, error: message };
+    }
+  }
+
+  private buildModelInput(input: ImageGenerationInput) {
+    const { imageUrls, prompt, aspectRatio, outputFormat } = input;
+
+    const formattedUrls = imageUrls?.map((url) =>
+      url.startsWith("data:") ? url : `data:image/jpeg;base64,${url}`,
+    );
+
+    return {
+      ...(formattedUrls && formattedUrls.length > 0
+        ? { image_urls: formattedUrls }
+        : {}),
+      prompt,
+      aspect_ratio: aspectRatio ?? "4:3",
+      output_format: outputFormat ?? "jpeg",
+      num_images: 1,
+    };
+  }
+
+  private extractImageUrl(result: unknown): string | undefined {
+    const rawResult = result as Record<string, unknown>;
+    const data = (rawResult?.data ?? rawResult) as {
+      images?: Array<{ url: string }>;
+    };
+    return data?.images?.[0]?.url;
+  }
+}

package/src/infrastructure/executors/video-executor.ts

@@ -0,0 +1,130 @@
+/**
+ * Video Generation Executor
+ * Generic executor for all video generation features
+ */
+
+import type {
+  GenerationExecutor,
+  VideoGenerationInput,
+  VideoGenerationOutput,
+  GenerationOptions,
+  GenerationResult,
+} from "../../domains/generation/domain/generation.types";
+import { providerRegistry } from "../services/provider-registry.service";
+
+declare const __DEV__: boolean;
+
+export class VideoExecutor
+  implements GenerationExecutor<VideoGenerationInput, VideoGenerationOutput>
+{
+  async generate(
+    model: string,
+    input: VideoGenerationInput,
+    options?: GenerationOptions,
+  ): Promise<GenerationResult<VideoGenerationOutput>> {
+    try {
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[VideoExecutor] Starting generation", {
+          model,
+          hasSourceImage: !!input.sourceImageBase64,
+          hasTargetImage: !!input.targetImageBase64,
+          hasPrompt: !!input.prompt,
+        });
+      }
+
+      const provider = providerRegistry.getActiveProvider();
+
+      if (!provider?.isInitialized()) {
+        if (typeof __DEV__ !== "undefined" && __DEV__) {
+          console.error("[VideoExecutor] Provider not initialized");
+        }
+        return { success: false, error: "AI provider not initialized" };
+      }
+
+      options?.onProgress?.(5);
+
+      const modelInput = this.buildModelInput(input);
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[VideoExecutor] Model input prepared");
+      }
+
+      options?.onProgress?.(10);
+
+      const result = await provider.subscribe(model, modelInput, {
+        timeoutMs: options?.timeoutMs ?? 300000,
+        onQueueUpdate: (status) => {
+          if (typeof __DEV__ !== "undefined" && __DEV__) {
+            console.log("[VideoExecutor] Queue status:", status.status);
+          }
+
+          if (status.status === "IN_QUEUE") {
+            options?.onProgress?.(20);
+          } else if (status.status === "IN_PROGRESS") {
+            options?.onProgress?.(50);
+          }
+        },
+      });
+
+      options?.onProgress?.(90);
+
+      const videoUrl = this.extractVideoUrl(result);
+
+      if (!videoUrl) {
+        if (typeof __DEV__ !== "undefined" && __DEV__) {
+          console.error("[VideoExecutor] No video URL in response");
+        }
+        return { success: false, error: "No video generated" };
+      }
+
+      options?.onProgress?.(100);
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.log("[VideoExecutor] Generation successful", {
+          videoUrl: videoUrl.slice(0, 100),
+        });
+      }
+
+      return { success: true, data: { videoUrl } };
+    } catch (error) {
+      const message =
+        error instanceof Error ? error.message : "Generation failed";
+
+      if (typeof __DEV__ !== "undefined" && __DEV__) {
+        console.error("[VideoExecutor] Generation error:", {
+          message,
+          error,
+        });
+      }
+
+      return { success: false, error: message };
+    }
+  }
+
+  private buildModelInput(input: VideoGenerationInput) {
+    const { sourceImageBase64, targetImageBase64, prompt } = input;
+
+    const formatBase64 = (base64: string): string => {
+      if (!base64) return "";
+      return base64.startsWith("data:")
+        ? base64
+        : `data:image/jpeg;base64,${base64}`;
+    };
+
+    return {
+      image_url: formatBase64(sourceImageBase64),
+      ...(targetImageBase64
+        ? { driver_image_url: formatBase64(targetImageBase64) }
+        : {}),
+      ...(prompt ? { prompt } : {}),
+    };
+  }
+
+  private extractVideoUrl(result: unknown): string | undefined {
+    const rawResult = result as Record<string, unknown>;
+    const data = (rawResult?.data ?? rawResult) as {
+      video?: { url: string };
+    };
+    return data?.video?.url;
+  }
+}

package/src/features/ai-hug/README.md

@@ -1,438 +0,0 @@
-# AI Hug Feature
-
-Generate AI-powered hug images between two people.
-
-## 📍 Import Path
-
-```typescript
-import { useAIHugFeature } from '@umituz/react-native-ai-generation-content';
-```
-
-**Location**: `src/features/ai-hug/`
-
-## 🎯 Feature Purpose
-
-Create realistic hug images by combining two photos using AI. Automatically detects poses, positions arms naturally, and generates high-quality facial matching for various hug types.
-
----
-
-## 📋 Usage Strategy
-
-### When to Use This Feature
-
-✅ **Use Cases:**
-- Creating couple photos
-- Generating family moments
-- Fun friendship images
-- Social media content
-- Creative personal projects
-
-❌ **When NOT to Use:**
-- Non-consensual image generation
-- Misleading or deceptive content
-- Harassment or bullying
-- Commercial use without permissions
-
-### Implementation Strategy
-
-1. **Select TWO photos** (person 1 and person 2)
-2. **Choose hug type** (romantic, friendly, family, cute)
-3. **Validate both photos** before generation
-4. **Generate AI hug** with progress tracking
-5. **Preview result** and offer regeneration
-6. **Save or share** the final image
-
----
-
-## ⚠️ Critical Rules (MUST FOLLOW)
-
-### 1. Image Requirements
-- **MUST** provide TWO distinct images (person 1 + person 2)
-- **MUST** contain at least one visible person in each image
-- **MUST** use high-quality images (min 512x512 recommended)
-- **MUST** ensure faces are clearly visible
-- **MUST NOT** use images with no detectable people
-
-### 2. Configuration
-- **MUST** provide valid `userId` for tracking
-- **MUST** specify `hugType` (romantic, friendly, family, cute)
-- **MUST** implement `onError` callback
-- **MUST** implement `onSelectPerson1` and `onSelectPerson2` callbacks
-- **MUST** handle both images being selected before processing
-
-### 3. State Management
-- **MUST** check `isReady` before enabling generate button
-- **MUST** verify both images are selected
-- **MUST** handle `isProcessing` state to prevent duplicate requests
-- **MUST** display `error` state to users
-- **MUST** implement proper cleanup on unmount
-
-### 4. Performance
-- **MUST** limit image size (<10MB each)
-- **MUST** compress images before processing
-- **MUST** implement loading indicators during processing
-- **MUST** cache results locally
-- **MUST NOT** generate multiple hugs simultaneously
-
-### 5. Ethics & Privacy
-- **MUST** obtain consent from people whose photos are used
-- **MUST** provide clear usage terms
-- **MUST** implement content moderation
-- **MUST** prevent malicious use cases
-- **MUST** log processing for audit trail
-
----
-
-## 🚫 Prohibitions (MUST AVOID)
-
-### Strictly Forbidden
-
-❌ **NEVER** do the following:
-
-1. **No Single Image**
-   - Always requires TWO images (person 1 + person 2)
-   - Never attempt with missing image
-
-2. **No Non-Consensual Generation**
-   - Always obtain permission from subjects
-   - Never generate hugs without consent
-   - Implement age verification for minors
-
-3. **No Malicious Use**
-   - Never use for harassment or bullying
-   - Never create misleading content
-   - Never use for deception
-
-4. **No Unhandled Errors**
-   - Never ignore generation failures
-   - Always handle detection errors gracefully
-   - Provide clear error messages
-
-5. **No Memory Leaks**
-   - Never store large images in state unnecessarily
-   - Always cleanup image references on unmount
-   - Implement proper image disposal
-
-6. **No Blocked UI**
-   - Never process without user confirmation
-   - Always show progress indicator
-   - Never block main thread with image processing
-
-7. **No Missing Context**
-   - Never confuse which person is which
-   - Always provide clear UI labels
-   - Show preview before processing
-
----
-
-## 🤖 AI Agent Directions
-
-### For AI Code Generation Tools
-
-When using this feature with AI code generation tools, follow these guidelines:
-
-#### Prompt Template for AI Agents
-
-```
-You are implementing an AI hug generation feature using @umituz/react-native-ai-generation-content.
-
-REQUIREMENTS:
-1. Import from: @umituz/react-native-ai-generation-content
-2. Use the useAIHugFeature hook
-3. Require TWO images (person 1 + person 2)
-4. Implement dual image selection UI
-5. Select hug type (romantic, friendly, family, cute)
-6. Validate both images before generation
-7. Add confirmation dialog before generation
-8. Handle long processing times with progress
-9. Implement proper error handling
-10. Implement cleanup on unmount
-
-CRITICAL RULES:
-- MUST obtain consent from subjects
-- MUST validate both images before processing
-- MUST provide clear UI for person 1 vs person 2
-- MUST handle generation errors gracefully
-- MUST prevent malicious use cases
-- MUST implement content moderation
-- NEVER process without user confirmation
-
-CONFIGURATION:
-- Provide valid userId (string)
-- Set hugType: 'romantic' | 'friendly' | 'family' | 'cute'
-- Implement onSelectPerson1 callback
-- Implement onSelectPerson2 callback
-- Implement onSaveResult callback
-- Configure callbacks: onProcessingStart, onProcessingComplete, onError
-
-OPTIONS:
-- hugType: Select hug style
-- preserveFaces: boolean (maintain facial features)
-- enhanceQuality: boolean (enhance output quality)
-
-STRICTLY FORBIDDEN:
-- No single image processing
-- No non-consensual generation
-- No malicious use
-- No unhandled errors
-- No memory leaks
-- No missing UI context
-
-ETHICS CHECKLIST:
-- [ ] Consent mechanism implemented
-- [ ] Age verification for minors
-- [ ] Content moderation in place
-- [ ] Usage terms provided
-- [ ] Audit trail logging
-- [ ] Report/flag functionality
-```
-
-#### AI Implementation Checklist
-
-Use this checklist when generating code:
-
-- [ ] Feature imported from correct path
-- [ ] Dual image selection implemented
-- [ ] Person 1/Person 2 labels clear
-- [ ] Hug type selector added
-- [ ] Both images validated before processing
-- [ ] Confirmation dialog added
-- [ ] Progress indicator during processing
-- [ ] Error display with user-friendly message
-- [ ] Result preview before saving
-- [ ] Consent mechanism in place
-- [ ] Cleanup on unmount
-- [ ] Content moderation configured
-
----
-
-## 🛠️ Configuration Strategy
-
-### Essential Configuration
-
-```typescript
-// Required fields
-{
-  userId: string
-  hugType: 'romantic' | 'friendly' | 'family' | 'cute'
-  onSelectPerson1: () => Promise<string | null>
-  onSelectPerson2: () => Promise<string | null>
-}
-
-// Optional callbacks
-{
-  onProcessingStart?: () => void
-  onProcessingComplete?: (result) => void
-  onError?: (error: string) => void
-}
-```
-
-### Recommended Settings
-
-1. **Hug Types**
-   - Romantic: Intimate couple hugs
-   - Friendly: Casual friendship hugs
-   - Family: Warm family embraces
-   - Cute: Adorable, playful hugs
-
-2. **Image Quality**
-   - Minimum: 512x512 resolution
-   - Recommended: 1024x1024 or higher
-   - Format: JPEG or PNG
-   - Max size: 10MB per image
-
-3. **Performance Settings**
-   - Compress images before upload
-   - Show progress for long operations
-   - Implement timeout (120s default)
-   - Enable result caching
-
----
-
-## 📊 State Management
-
-### Feature States
-
-**isReady**: boolean
-- Both images selected and validated
-- Check before enabling generate button
-
-**isProcessing**: boolean
-- AI hug generation in progress
-- Show loading/progress indicator
-- Disable generate button
-
-**progress**: number (0-100)
-- Generation progress percentage
-- Update progress bar
-
-**error**: string | null
-- Error message if generation failed
-- Common errors: "No person found in image", "Generation failed"
-
-**result**: {
-  imageUrl: string
-  hugType?: string
-  metadata?: any
-}
-
----
-
-## 🔐 Ethics & Privacy
-
-### Consent Requirements
-
-- **MUST** obtain explicit consent from all subjects
-- **MUST** provide clear explanation of how images will be used
-- **MUST** implement age verification
-- **MUST** allow subjects to opt-out
-
-### Content Moderation
-
-- **MUST** filter inappropriate content
-- **MUST** prevent malicious use cases
-- **MUST** implement reporting mechanism
-- **MUST** review flagged content
-
-### Usage Guidelines
-
-- **MUST** provide terms of service
-- **MUST** clearly label AI-generated content
-- **MUST** prevent deception/misrepresentation
-- **MUST** comply with deepfake regulations
-
----
-
-## 🎨 Best Practices
-
-### Image Selection
-
-1. **Photo Quality**
-   - Good: Clear, well-lit, high-resolution photos
-   - Bad: Blurry, dark, low-resolution images
-
-2. **Pose Clarity**
-   - Good: Full-body or waist-up shots
-   - Bad: Extreme close-ups, obscured poses
-
-3. **Face Visibility**
-   - Good: Clear, frontal face shots
-   - Bad: Occluded or profile faces
-
-4. **Lighting**
-   - Similar lighting conditions work best
-   - Front-facing well-lit photos ideal
-
-### User Experience
-
-1. **Clear UI**
-   - Label person 1 vs person 2 clearly
-   - Show preview before processing
-   - Add confirmation dialog
-
-2. **Error Handling**
-   - Explain "no person found" errors
-   - Provide troubleshooting tips
-   - Offer retry option
-
-3. **Performance**
-   - Compress images before upload
-   - Show progress for long operations
-   - Cache results for re-download
-
----
-
-## 🐛 Common Pitfalls
-
-### Detection Issues
-
-❌ **Problem**: "No person found" error
-✅ **Solution**: Ensure people are clearly visible, well-lit, frontal
-
-### Quality Issues
-
-❌ **Problem**: Poor quality generation
-✅ **Solution**: Use higher resolution images, better lighting, try different hug type
-
-### UX Confusion
-
-❌ **Problem**: Users confused about which person is which
-✅ **Solution**: Clear labels, visual indicators, preview
-
-### Privacy Concerns
-
-❌ **Problem**: Non-consensual image generation
-✅ **Solution**: Implement consent mechanisms, age verification
-
----
-
-## 📦 Related Components
-
-Use these components from the library:
-
-- **DualImagePicker**: Select two images
-- **HugTypeSelector**: Choose hug style
-- **ResultImageCard**: Display result with actions
-- **ConfirmationDialog**: Confirm before processing
-
-Located at: `src/presentation/components/`
-
----
-
-## 🔄 Migration Strategy
-
-If migrating from previous implementation:
-
-1. **Update imports** to new path
-2. **Add dual image selection** (person 1 + person 2)
-3. **Implement consent mechanism**
-4. **Add hug type selector**
-5. **Update state handling** for both images
-6. **Test all error cases**
-
----
-
-## ⚖️ Legal Considerations
-
-### Compliance
-
-- **Deepfake Regulations**: Comply with local laws
-- **Privacy Laws**: GDPR, CCPA compliance
-- **Consent Requirements**: Explicit permission needed
-- **Age Restrictions**: Verify adult subjects
-- **Content Labeling**: Mark as AI-generated
-
-### Best Practices
-
-- Provide attribution for source images
-- Allow content reporting/flagging
-- Implement audit trail logging
-- Cooperate with takedown requests
-
----
-
-## 📚 Additional Resources
-
-- Main documentation: `/docs/`
-- API reference: `/docs/api/`
-- Examples: `/docs/examples/basic/ai-hug/`
-- Ethics guidelines: `/docs/ethics.md`
-
----
-
-**Last Updated**: 2025-01-08
-**Version**: 2.0.0 (Strategy-based Documentation)
-
----
-
-## 📝 Changelog
-
-### v2.0.0 - 2025-01-08
-- **BREAKING**: Documentation format changed to strategy-based
-- Removed extensive code examples
-- Added ethics and privacy guidelines
-- Added rules, prohibitions, and AI agent directions
-- Focus on responsible AI usage
-
-### v1.0.0 - Initial Release
-- Initial feature documentation

package/src/features/ai-hug/domain/index.ts

@@ -1,5 +0,0 @@
-/**
- * AI Hug Domain Index
- */
-
-export * from "./types";