Advanced NN Tutorial: Mastering File Handling in Workflows

Key Points:

• Binary View shows additional information about files (name, directory, type, size)
• When dealing only with files, the Table, JSON, and Schema views remain empty
• You can preview compatible files directly in the interface
• Files can be downloaded to your local machine directly from the Binary View

My Take:

The Binary View is a game-changer for workflows that process visual assets or documents. Being able to preview files directly within the interface saves significant time compared to downloading and opening them externally each time.

Key Points:

• Compress/Decompress nodes for handling ZIP archives
• Read/Write File nodes for disk operations (especially useful for self-hosting)
• Conversion nodes for transforming between JSON objects and file formats

My Take:

The file handling capabilities in NN are quite robust compared to some other automation platforms. If you’re self-hosting NN, the disk operation nodes become particularly valuable for creating workflows that interact with your local file system.

Key Points:

• HTTP Request nodes can retrieve files when “File” is selected as the expected response format
• Retrieved files are stored in specified fields (e.g., “logo” in the example)
• Binary items append to each other rather than overwriting in sequential nodes
• Multiple binary files can be collected in a single workflow execution

// Example of HTTP Request node configuration for file retrieval
{
  "url": "https://domain.com/nn-logo.svg",
  "method": "GET",
  "responseFormat": "file",
  "responsePropertyName": "logo"
}
    

My Take:

The ability to append binary items is incredibly useful for building workflows that gather multiple files from different sources before processing them together. Just remember that you need to keep track of your field naming when working with multiple files.

Key Points:

• The Compress Files node handles both compression and decompression
• You must specify which binary fields to include in the archive
• Set the output format to ZIP and provide a filename for the archive
• The resulting compressed file can be downloaded directly from the node output

My Take:

This compression feature is particularly valuable when building workflows that need to deliver multiple files to users or other systems as a single package. For example, bundling generated reports with supporting images or data files.

Key Points:

• Use HTTP Request to fetch a ZIP file from a server
• The Decompress node extracts all files contained in the archive
• Extracted files appear as a single item with multiple binary files
• Each extracted file includes metadata (name, type, size)

My Take:

This decompression capability opens up interesting workflow possibilities – you could create systems that process uploaded ZIP archives from users, extracting and processing multiple files in a single operation. Just be mindful that at this stage, all files are still part of a single workflow item.

Key Points:

• After decompression, all extracted files exist within a single workflow item
• Use a Code node to split binary files into separate items
• Template code is available for this common operation
• After splitting, each file becomes an individual binary item that can be processed independently

// Example code structure for splitting binary files
// (Check NN templates for the exact implementation)
return items.flatMap(item => {
  const results = [];
  for (const binaryKey of Object.keys(item.binary)) {
    for (const fileKey of Object.keys(item.binary[binaryKey])) {
      const newItem = {
        json: {},
        binary: {}
      };
      newItem.binary[binaryKey] = item.binary[binaryKey][fileKey];
      results.push(newItem);
    }
  }
  return results;
});
    

My Take:

This splitting technique is critical for workflows that need to process each file differently. For example, if you’re extracting images from a ZIP archive and need to resize each one separately or run them through different analysis tools, you’ll need this pattern.