Refactor localization strings to use constant for text domain; add image context limit feature

- Updated localization strings in various classes to use the constant `GROQ_AI_PRODUCT_TEXT_DOMAIN` instead of hardcoded text domain.
- Introduced an image context limit setting in the settings manager and adjusted related methods to accommodate this new feature.
- Modified prompt builder to handle image context limit when building product context blocks and retrieving image payloads.
- Enhanced error handling and response structures to include new fields related to image context.
- Added support for title suggestions in the structured response from the AI.

assets/css/admin.css

@@ -123,6 +123,36 @@
 	width: 100%;
 }
 
+.groq-ai-title-suggestions {
+	border: 1px dashed #dcdcde;
+	border-radius: 4px;
+	padding: 8px;
+	background: #fefefe;
+	margin-top: 8px;
+}
+
+.groq-ai-title-suggestions__options {
+	display: flex;
+	flex-direction: column;
+	gap: 6px;
+	margin-top: 6px;
+}
+
+.groq-ai-title-suggestions__option {
+	display: flex;
+	align-items: flex-start;
+	gap: 8px;
+}
+
+.groq-ai-title-suggestions__option input[type='radio'] {
+	margin-top: 3px;
+}
+
+.groq-ai-title-suggestions__hint {
+	margin-top: 6px;
+	margin-bottom: 0;
+}
+
 .groq-ai-modal__raw {
 	margin-top: 16px;
 }