syft-objects 0.3.1

Share files with explicit mock vs private control

🔐 Syft Objects

Share files with explicit mock vs private control

Quick Start

import syft_objects as syo

# Create an object with demo and real content
obj = syo.syobj(
    name="AI Results",
    mock_contents="Model achieved good performance",
    private_contents="Accuracy: 94.5%, Cost: $127"
)

# Browse all your objects interactively
syo.objects

# Search for specific objects
syo.objects.search("financial")

What It Does

Mock vs Private Pattern: Every object has two versions:

• Mock: What everyone sees (demo/sample data)
• Private: What authorized users see (real data)

Example:

obj = syo.syobj(
    name="Customer Analysis", 
    mock_contents="Sample: 100 customers, avg age 42, 60% retention rate",
    private_contents="Full: 47,293 customers, avg age 41.7, 58.3% retention, avg LTV $1,247"
)

Interactive Object Browser

The syo.objects collection provides a beautiful interactive interface:

# Browse all objects with search and selection
syo.objects

# Search by name, email, description, or metadata
syo.objects.search("financial")
syo.objects.search("customer")

# Filter by email
syo.objects.filter_by_email("andrew")

# Get specific objects
selected = [syo.objects[i] for i in [0, 1, 5]]

# Refresh after creating new objects
syo.objects.refresh()

Interactive Features

• 🔍 Real-time search across names, emails, descriptions, and metadata
• ☑️ Multi-select with checkboxes
• 📋 Code generation - click "Generate Code" to get copy-paste Python
• 📊 10-column table with all object details
• 🔄 Auto-refresh - new objects appear immediately
• 📱 Responsive design with horizontal scrolling

Rich Object Display

Each object shows beautifully in Jupyter with:

• 📁 File information with availability status
• 🎯 Permission levels with color-coded badges
• 📋 Metadata including creation/update times
• 🖱️ Interactive buttons to view content inline
• 📄 Smart truncation for long content (first 1000 chars)

Permission Control

obj = syo.syobj(
    name="Financial Report",
    mock_contents="Q4 Summary: Revenue up 10%", 
    private_contents="Q4: $2.5M revenue, $400K profit, 23.7% margin",
    discovery_read=["public"],           # Who knows it exists
    mock_read=["[e-mail]"],  # Who sees demo
    private_read=["[e-mail]"]     # Who sees real data
)

File-Based Objects

# Use existing files
obj = syo.syobj(
    name="Dataset Analysis",
    mock_file="sample_100_rows.csv",      # Demo file
    private_file="full_50k_rows.csv"      # Real file  
)

Installation

pip install syft-objects

For SyftBox integration:

pip install syft-objects[syftbox]

Key Features

• 🎯 One function: syo.syobj() - simple and clean
• 🔒 Explicit control: You decide what goes in mock vs private
• 🎨 Beautiful display: Rich HTML widgets in Jupyter
• 🔍 Interactive browsing: Search and select objects easily
• 🆔 Unique filenames: No collisions with UID-based naming
• ⚡ Real-time updates: New objects appear immediately
• 📊 Comprehensive table: 10 columns with all object details
• 🖱️ Inline content: View file contents directly in notebook
• 🎯 Permission system: Fine-grained access control

Quick Reference

import syft_objects as syo

# Create objects
obj = syo.syobj(name="My Data", mock_contents="Demo", private_contents="Real")

# Browse interactively
syo.objects                    # Show interactive table
syo.objects[0]                 # Get first object  
syo.objects[:3]                # Get first 3 objects
len(syo.objects)               # Count objects

# Search and filter
syo.objects.search("keyword")           # Search everywhere
syo.objects.filter_by_email("user")    # Filter by email
syo.objects.get_by_indices([0,1,5])    # Get specific objects

# Utilities
syo.objects.list_unique_emails()       # List all emails
syo.objects.refresh()                  # Refresh collection

License

Apache License 2.0