Automated Image Validation System

A C#/.NET 8 console application that validates product images for a list of SKUs. For each SKU, the app fetches associated image URLs via a BigCommerce-like client (fake implementation included), downloads images, extracts metadata using a Node.js helper (Sharp), evaluates against quality criteria, and generates CSV reports.

Features

• Reads SKUs from an input CSV file (expects a header column named SKU)
• Fetches image URLs per SKU via IBigCommerceClient (fake client provided for deterministic testing)
• Downloads each image and extracts metadata using Node.js + Sharp (width, height, density/DPI when available)
• Validates each image against configurable thresholds (min width/height; optional DPI enforcement)
• Produces two reports:
  • Image-level report: validation per image
  • SKU-level report: rollup per SKU (flags SKUs with any failing images or no images)
• Bounded concurrency for faster processing

Prerequisites

• .NET 8 SDK
• Node.js + npm (Node 18+ recommended)

Setup

From the repository root:

1. Restore .NET packages:

   dotnet restore

2. Install Node dependencies for the metadata tool:

   cd tools/image-meta

   npm install

   cd ../..

Windows PowerShell note (npm execution policy)

If PowerShell blocks npm.ps1, run npm via the cmd shim:

& "C:\Program Files\nodejs\npm.cmd" install

Alternatively, run npm install from Command Prompt.

Configuration

Configuration is stored in src/ImageValidation.Cli/appsettings.json:

• Validation.MinWidthPx: minimum allowed width (px)
• Validation.MinHeightPx: minimum allowed height (px)
• Validation.MinDpi: minimum allowed DPI (if enforced)
• Validation.FailIfDpiMissing: if true, missing DPI is flagged; otherwise DPI is treated as N/A
• Concurrency.MaxConcurrency: max number of SKUs processed concurrently
• Node.ToolsDirectory: path to the Node helper directory (repo-root-relative)

Usage

Run with defaults (uses samples/input-skus.csv and writes to output/):

dotnet run --project .\src\ImageValidation.Cli\ImageValidation.Cli.csproj

Run with custom input/output:

dotnet run --project .\src\ImageValidation.Cli\ImageValidation.Cli.csproj -- <inputCsv> <outputImageReportCsv>

Example:

dotnet run --project .\src\ImageValidation.Cli\ImageValidation.Cli.csproj -- .\samples\input-skus.csv .\output\output-report.csv

Output

The app writes:

1. Image-level report (per-image)

   • Default: output/output-report.csv
   • Columns: Sku, ImageUrl, Width, Height, Dpi, DimensionResult, DpiResult, Status, Notes

2. SKU-level summary report (rollup)

   • Default: output/output-report-skus.csv
   • Columns: Sku, ImageCount, OkCount, FlagCount, Status, Notes

SKU status logic:

• FLAG if any image is flagged OR the SKU has no images
• otherwise OK

Project Structure

• src/

  • ImageValidation.Cli/ (Main .NET console application)

    • Models/ (Data models: Sku, Image, Metadata, ReportRow, SkuSummaryRow)
    • BigCommerce/ (IBigCommerceClient + FakeBigCommerceClient + placeholder real client)
    • Services/ (CSV, download, metadata extraction, validation, reporting, SKU summary)
    • Utilities/ (Process runner, temp file helpers, concurrency helpers)
    • appsettings.json (Configuration)
    • Program.cs (Entry point)

• tools/

  • image-meta/ (Node metadata extractor using Sharp)

    • package.json
    • image-meta.js

• samples/

  • input-skus.csv (Sample input file)

• output/ (Output report directory)

• README.md

How it Works

1. Read SKUs from the CSV input file

2. For each SKU:

   • Fetch image URLs from IBigCommerceClient
   • Download images to temporary files
   • Invoke node image-meta.js "<tempPath>" to extract metadata as JSON
   • Validate against configured thresholds

3. Write the image-level report CSV

4. Build and write the SKU-level summary CSV

5. Print a short console summary

Swapping FakeBigCommerceClient for a Real BigCommerce Client

1. Implement IBigCommerceClient using HttpClient (e.g., RealBigCommerceClient) to call BigCommerce endpoints:

   • Look up product by SKU (catalog/products)
   • Retrieve product images (catalog/products/{productId}/images)

2. Replace the client instantiation in Program.cs:

   // var bigCommerceClient = new FakeBigCommerceClient();
   var bigCommerceClient = new RealBigCommerceClient(httpClient, /* credentials/config */);

3. Provide credentials via configuration or environment variables for real integrations.

Additional Considerations for Real BigCommerce Implementation

When implementing a real BigCommerce client, we would need to consider the following complexities:

1. SKU Mapping Complexity: In real BigCommerce implementations, SKU mapping can be complex because:

   • SKUs can exist at both product and variant levels
   • A single SKU might map to multiple products/variants
   • The API might require different lookup paths depending on store configuration

2. API Response Structure: The BigCommerce API looks to return rich metadata with images that could probably be leveraged:

   • Image descriptions/alt text for accessibility validation
   • Image position/order for display validation
   • Image file names for naming convention validation

3. Performance Considerations:

   • I might consider adding caching for product lookups to avoid repeated API calls
   • Implement bulk operations if processing many SKUs (e.g., multiple SKUs per API call, safe parallelizing with Parallel.ForEachAsync...)
   • Handle rate limiting (429 responses) with appropriate backoff strategies (I've never implemented this but I know it's a thing)

4. Error Handling:

   • Handle authentication errors (401, 403)
   • Implement retry logic for transient failures
   • Log API errors for debugging and monitoring

Sample Input

samples/input-skus.csv includes:

• SKU001: multiple images (mixed results)
• SKU002: single passing image
• SKU003: no images
• SKU004: multiple images (mixed results)
• SKU005: unknown SKU (no images)