summaryrefslogtreecommitdiff
path: root/internal/image
diff options
context:
space:
mode:
authorPaul Buetow <paul@buetow.org>2025-07-15 23:42:32 +0300
committerPaul Buetow <paul@buetow.org>2025-07-15 23:42:32 +0300
commitb105333c061ea165b3b79317415cbb8b9cfb7c75 (patch)
treec2682cc156c372d85ab52d514df4316ceda9071d /internal/image
parent61529facc2c5321de9f0ab9123cb1de25bcab62c (diff)
feat: add English translations and detailed attribution files
- Automatic Bulgarian to English translation for all words - Save translations to separate _translation.txt files - Include translations in Anki CSV export - Add detailed attribution files for audio and images: - Audio: model, voice, speed, instructions, processed text - Image: model, size, quality, style, full prompt used - Expand image styles to 42 different options (including superhero comic, yoga, etc.) - Improve image prompts to strongly avoid text generation - Fix image overwrite issue - now overwrites existing files instead of failing 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
Diffstat (limited to 'internal/image')
-rw-r--r--internal/image/openai.go62
1 files changed, 57 insertions, 5 deletions
diff --git a/internal/image/openai.go b/internal/image/openai.go
index ee05c80..c4b2e9d 100644
--- a/internal/image/openai.go
+++ b/internal/image/openai.go
@@ -11,6 +11,7 @@ import (
"os"
"path/filepath"
"strings"
+ "time"
"github.com/sashabaranov/go-openai"
)
@@ -25,6 +26,7 @@ type OpenAIClient struct {
style string // natural or vivid (dall-e-3 only)
cacheDir string
enableCache bool
+ lastPrompt string // Store the last used prompt for attribution
}
// OpenAIConfig holds configuration for the OpenAI image provider
@@ -124,6 +126,9 @@ func (c *OpenAIClient) Search(ctx context.Context, opts *SearchOptions) ([]Searc
// Create educational prompt
prompt := c.createEducationalPrompt(opts.Query, translatedWord)
+ // Store the prompt for attribution
+ c.lastPrompt = prompt
+
// Log the prompt to stdout for debugging
fmt.Printf("OpenAI Image Generation Prompt: %s\n", prompt)
fmt.Printf("OpenAI Image Generation: Using model '%s' with size '%s'\n", c.model, c.size)
@@ -221,7 +226,16 @@ func (c *OpenAIClient) Download(ctx context.Context, url string) (io.ReadCloser,
// GetAttribution returns the required attribution text
func (c *OpenAIClient) GetAttribution(result *SearchResult) string {
- return "Image generated by OpenAI DALL-E"
+ attribution := fmt.Sprintf("Image generated by OpenAI DALL-E\n\n")
+ attribution += fmt.Sprintf("Model: %s\n", c.model)
+ attribution += fmt.Sprintf("Size: %s\n", c.size)
+ if c.model == "dall-e-3" {
+ attribution += fmt.Sprintf("Quality: %s\n", c.quality)
+ attribution += fmt.Sprintf("Style: %s\n", c.style)
+ }
+ attribution += fmt.Sprintf("\nPrompt used:\n%s\n", c.lastPrompt)
+ attribution += fmt.Sprintf("\nGenerated at: %s\n", time.Now().Format("2006-01-02 15:04:05"))
+ return attribution
}
// Name returns the name of the provider
@@ -231,8 +245,9 @@ func (c *OpenAIClient) Name() string {
// createEducationalPrompt generates a prompt optimized for language learning
func (c *OpenAIClient) createEducationalPrompt(bulgarianWord, englishTranslation string) string {
- // Define different art styles for variety
+ // Define different art styles for variety (42 styles total)
styles := []string{
+ // Original styles (1-10)
"photorealistic, high quality photograph",
"detailed digital illustration, clean vector art style",
"vibrant cartoon style, animated movie quality",
@@ -243,9 +258,46 @@ func (c *OpenAIClient) createEducationalPrompt(bulgarianWord, englishTranslation
"oil painting, classical art style",
"paper cut art, layered craft style",
"isometric illustration, technical drawing style",
- "superhero comic book style, dynamic action pose, bold colors",
- "yoga/wellness illustration style, peaceful zen aesthetic",
+
+ // Requested styles (11-13)
+ "superhero comic book style, dynamic action pose, bold colors, Marvel/DC aesthetic",
+ "super-realistic person practicing yoga, serene wellness photography",
"cute illustration with cats interacting with the subject, whimsical cat-themed",
+
+ // Additional artistic styles (14-25)
+ "impressionist painting style, Monet-inspired brushstrokes",
+ "art nouveau style, decorative organic forms, Mucha-inspired",
+ "pop art style, Andy Warhol inspired, bold contrasting colors",
+ "Japanese ukiyo-e woodblock print style, traditional aesthetic",
+ "steampunk illustration, Victorian era mechanical elements",
+ "cyberpunk neon aesthetic, futuristic digital art",
+ "medieval illuminated manuscript style, gold leaf details",
+ "graffiti street art style, urban spray paint aesthetic",
+ "stained glass window art, cathedral-inspired design",
+ "mosaic tile art style, Byzantine-inspired patterns",
+ "art deco style, geometric patterns, 1920s aesthetic",
+ "surrealist style, Salvador Dali inspired dreamlike quality",
+
+ // Photography styles (26-32)
+ "macro photography style, extreme close-up detail",
+ "vintage polaroid photograph, retro instant camera aesthetic",
+ "film noir style, dramatic black and white photography",
+ "golden hour photography, warm sunset lighting",
+ "underwater photography style, ethereal aquatic atmosphere",
+ "aerial drone photography, bird's eye view perspective",
+ "long exposure photography, motion blur effects",
+
+ // Modern digital styles (33-42)
+ "vaporwave aesthetic, 80s-90s retro digital art",
+ "low poly 3D art style, geometric simplified forms",
+ "pixel art style, 8-bit retro video game aesthetic",
+ "glitch art style, digital distortion effects",
+ "double exposure photography, artistic overlay effect",
+ "tilt-shift photography, miniature world effect",
+ "infrared photography style, otherworldly color palette",
+ "holographic iridescent style, rainbow prismatic effects",
+ "origami paper folding art style, geometric paper craft",
+ "chalk art style, sidewalk drawing aesthetic",
}
// Select a random style
@@ -258,7 +310,7 @@ func (c *OpenAIClient) createEducationalPrompt(bulgarianWord, englishTranslation
"This is for the Bulgarian word '%s' which means %s. "+
"The image should be educational and suitable for language learning flashcards. "+
"Requirements: single main subject, plain background, clear and recognizable. "+
- "No text, labels, or writing in the image.",
+ "IMPORTANT: No text whatsoever. Do not include any words, letters, typography, labels, captions, or writing of any kind. Image only, without any text elements.",
selectedStyle, englishTranslation, bulgarianWord, englishTranslation,
)
}