@renoise/video-maker 0.1.3 → 0.2.1

package/skills/tiktok-content-maker/scripts/analyze-images.ts

@@ -1,122 +0,0 @@
-#!/usr/bin/env npx tsx
-/**
- * Analyze product + model images using Gemini for ecommerce video scripting.
- * Usage: npx tsx analyze-images.ts <product_image> [model_image]
- * Model image is optional — if omitted, only product analysis is returned.
- */
-
-import fs from 'fs/promises'
-import path from 'path'
-import { getGeminiClient, fileToInlinePart } from '../../../lib/gemini.js'
-
-const PROMPT_WITH_MODEL = `You are an expert ecommerce video director and product analyst. Analyze these two images:
-
-Image 1: Product photo
-Image 2: Model/person photo
-
-Return a JSON object with exactly this structure (no markdown, no code fences, pure JSON only):
-
-{
-  "product": {
-    "type": "product category (e.g. dress, sneaker, lipstick)",
-    "color": "main colors",
-    "material": "fabric/material description",
-    "highlights": "key visual features and design details",
-    "brand_tone": "brand positioning vibe (e.g. luxury, streetwear, minimalist)"
-  },
-  "model": {
-    "gender": "male/female",
-    "age_range": "estimated age range",
-    "hair": "hair style and color description",
-    "outfit": "what the model is wearing",
-    "vibe": "overall style/aesthetic (e.g. elegant, sporty, casual chic)"
-  },
-  "scene_suggestions": [
-    "scene 1 description with specific location and lighting",
-    "scene 2 description with specific location and lighting",
-    "scene 3 description with specific location and lighting"
-  ],
-  "selling_points": [
-    "selling point 1 — specific and compelling",
-    "selling point 2 — specific and compelling",
-    "selling point 3 — specific and compelling"
-  ]
-}`
-
-const PROMPT_PRODUCT_ONLY = `You are an expert ecommerce video director and product analyst. Analyze this product image.
-
-Return a JSON object with exactly this structure (no markdown, no code fences, pure JSON only):
-
-{
-  "product": {
-    "type": "product category (e.g. dress, sneaker, lipstick)",
-    "color": "main colors",
-    "material": "fabric/material description",
-    "highlights": "key visual features and design details",
-    "brand_tone": "brand positioning vibe (e.g. luxury, streetwear, minimalist)"
-  },
-  "scene_suggestions": [
-    "scene 1 description with specific location and lighting",
-    "scene 2 description with specific location and lighting",
-    "scene 3 description with specific location and lighting"
-  ],
-  "selling_points": [
-    "selling point 1 — specific and compelling",
-    "selling point 2 — specific and compelling",
-    "selling point 3 — specific and compelling"
-  ]
-}`
-
-async function main() {
-  const [productImage, modelImage] = process.argv.slice(2)
-  if (!productImage) {
-    console.error('Usage: analyze-images.ts <product_image> [model_image]')
-    process.exit(1)
-  }
-
-  // Validate files exist
-  await fs.access(productImage).catch(() => {
-    console.error(`ERROR: Product image not found: ${productImage}`)
-    process.exit(1)
-  })
-  if (modelImage) {
-    await fs.access(modelImage).catch(() => {
-      console.error(`ERROR: Model image not found: ${modelImage}`)
-      process.exit(1)
-    })
-  }
-
-  const genAI = getGeminiClient()
-  const model = genAI.getGenerativeModel({
-    model: process.env.GEMINI_MODEL ?? 'gemini-3.1-flash',
-    generationConfig: {
-      temperature: 0.3,
-      maxOutputTokens: 2048,
-    },
-  })
-
-  const productPart = await fileToInlinePart(productImage)
-
-  const parts: Array<{ text: string } | Awaited<ReturnType<typeof fileToInlinePart>>> = modelImage
-    ? [{ text: PROMPT_WITH_MODEL }, productPart, await fileToInlinePart(modelImage)]
-    : [{ text: PROMPT_PRODUCT_ONLY }, productPart]
-
-  const result = await model.generateContent(parts)
-
-  const text = result.response.text()
-
-  // Try to parse and pretty-print as JSON
-  const cleaned = text.replace(/^```json\n?/, '').replace(/\n?```$/, '').trim()
-  try {
-    const parsed = JSON.parse(cleaned)
-    console.log(JSON.stringify(parsed, null, 2))
-  } catch {
-    // If not valid JSON, output raw text
-    console.log(text)
-  }
-}
-
-main().catch((err) => {
-  console.error('ERROR:', err.message)
-  process.exit(1)
-})