pycdn 1.1.5

Revolutionary Package CDN with Natural Import System - Stream Python packages instantly without local installation!

PyCDN - Revolutionary Package CDN with Natural Import System 🚀

The Netflix of Python packages - Stream packages instantly without local installation!

PyCDN revolutionizes Python package management by serving packages through CDN networks with lazy loading and natural import syntax. Say goodbye to dependency hell and pip install delays!

🌟 Revolutionary Import System

PyCDN now supports natural Python import syntax using an advanced meta path import hook system. This means you can import packages directly from CDN servers as if they were installed locally!

Classic Usage (Still Works)

import pycdn

# Connect to CDN server
cdn = pycdn.pkg("http://localhost:8000")

# Use packages via attribute access
result = cdn.math.sqrt(16)  # Returns 4.0
data = cdn.numpy.array([1, 2, 3, 4, 5])
model = cdn.sklearn.LinearRegression()

🎯 NEW: Natural Import Syntax

import pycdn

# Connect and register CDN
cdn = pycdn.pkg("http://localhost:8000")  # Registers 'cdn' prefix

# Now use natural Python imports!
from cdn.openai import OpenAI
from cdn.numpy import array, mean
from cdn.pandas import DataFrame
from cdn.sklearn.ensemble import RandomForestClassifier

# Use exactly like local packages
client = OpenAI(api_key="your-key")
data = array([1, 2, 3, 4, 5])
avg = mean(data)
df = DataFrame({"col1": [1, 2, 3]})
model = RandomForestClassifier()

🔧 Custom Import Prefixes

# Use custom prefixes for different CDN servers
ml_cdn = pycdn.pkg("http://ml-cdn:8000", prefix="ml")
data_cdn = pycdn.pkg("http://data-cdn:8000", prefix="data")

# Import from different CDNs
from ml.tensorflow import keras
from ml.pytorch import nn
from data.pandas import DataFrame
from data.dask import dataframe as dd

🏢 Multiple CDN Support

# Connect to multiple CDN environments
prod = pycdn.pkg("http://prod-cdn:8000", prefix="prod")
dev = pycdn.pkg("http://dev-cdn:8000", prefix="dev")
test = pycdn.pkg("http://test-cdn:8000", prefix="test")

# Import from specific environments
from prod.stable_package import ProductionClass
from dev.beta_package import ExperimentalFeature
from test.mock_package import TestDouble

🎯 Key Features

• 🔥 Natural Import Syntax: Use from cdn.package import Class - feels exactly like local imports
• ⚡ Instant Access: No pip install required - packages execute remotely
• 🌍 Global CDN: Packages served from edge locations worldwide
• 🔒 Zero Dependencies: No local installation or dependency conflicts
• 💾 Smart Caching: Intelligent caching with hybrid in-memory + disk storage
• 🛡️ Security: Sandboxed execution with runtime security scanning
• 📊 Analytics: Usage tracking and performance monitoring
• 🔄 Multi-CDN: Connect to multiple CDN servers simultaneously
• 🧰 Development Mode: Local fallback and enhanced debugging

🚀 Installation

pip install pycdn

📖 Quick Start

1. Basic Usage

import pycdn

# Connect to CDN
cdn = pycdn.pkg("http://localhost:8000")

# Classic usage
result = cdn.math.sqrt(16)
print(result)  # 4.0

# Natural imports (NEW!)
from cdn.math import sqrt, pow
print(sqrt(25))    # 5.0
print(pow(2, 3))   # 8.0

2. Working with Classes

import pycdn

cdn = pycdn.pkg("http://localhost:8000")

# Import and use classes naturally
from cdn.openai import OpenAI
from cdn.sklearn.ensemble import RandomForestClassifier

# Create instances and call methods
client = OpenAI(api_key="your-key")
response = client.chat.completions.create(
    model="gpt-3.5-turbo",
    messages=[{"role": "user", "content": "Hello!"}]
)

model = RandomForestClassifier(n_estimators=100)
model.fit(X_train, y_train)
predictions = model.predict(X_test)

3. Data Science Workflow

import pycdn

# Connect to data science CDN
ds_cdn = pycdn.pkg("http://ds-cdn:8000", prefix="ds")

# Natural imports for entire data pipeline
from ds.pandas import DataFrame, read_csv
from ds.numpy import array, mean, std
from ds.matplotlib.pyplot as plt
from ds.sklearn.model_selection import train_test_split
from ds.sklearn.ensemble import RandomForestRegressor
from ds.sklearn.metrics import mean_squared_error

# Use exactly like local packages
df = read_csv("data.csv")
X = df[["feature1", "feature2"]]
y = df["target"]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
model = RandomForestRegressor()
model.fit(X_train, y_train)

predictions = model.predict(X_test)
mse = mean_squared_error(y_test, predictions)
print(f"MSE: {mse}")

🛠️ Advanced Features

Dynamic Prefix Management

cdn = pycdn.pkg("http://localhost:8000", prefix="initial")

# Change prefix dynamically
cdn.set_prefix("dynamic")

# Enable specific package imports
cdn.enable_imports(["tensorflow", "pytorch", "scikit-learn"])

# Now these work:
from dynamic.tensorflow import keras
from dynamic.pytorch import nn

Error Handling

from pycdn import PyCDNRemoteError

try:
    from cdn.nonexistent import something
except PyCDNRemoteError as e:
    print(f"CDN Error: {e.message}")
    print(f"Package: {e.package_name}")
    if e.remote_traceback:
        print(f"Remote traceback: {e.remote_traceback}")

Development Mode

# Configure for development
pycdn.configure(debug=True, timeout=60)

cdn = pycdn.pkg("http://localhost:8000", prefix="dev")
# Enables local fallback, enhanced debugging, mock mode

CDN Management

# View active CDN mappings
mappings = pycdn.get_cdn_mappings()
print(mappings)  # {'cdn': 'http://localhost:8000', 'ml': 'http://ml-cdn:8000'}

# Clear all mappings
pycdn.clear_cdn_mappings()

# Register/unregister specific prefixes
pycdn.register_cdn_client("custom", cdn_client)
pycdn.unregister_cdn_client("custom")

🏗️ Server Setup

Deploy packages to your CDN server:

from pycdn.server import CDNServer

# Create CDN server
server = CDNServer(port=8000)

# Deploy packages
server.deploy_package("math", version="1.0.0")
server.deploy_package("numpy", version="1.24.0")
server.deploy_package("openai", version="1.0.0")

# Start server
server.start()

🌐 Meta Path Import System

PyCDN uses Python's sys.meta_path to intercept imports and resolve them from CDN servers:

import sys
import pycdn

# When you create a CDN connection
cdn = pycdn.pkg("http://localhost:8000", prefix="cdn")

# PyCDN automatically:
# 1. Registers a MetaPathFinder in sys.meta_path
# 2. Maps the 'cdn' prefix to your CDN client
# 3. Intercepts imports starting with 'cdn.'
# 4. Creates proxy objects that execute remotely
# 5. Handles classes, functions, modules transparently

# All of this happens automatically!
from cdn.package import Class  # Intercepted and resolved

🔧 Configuration

# Global configuration
pycdn.configure(
    debug=True,
    timeout=30,
    cache_size="100MB",
    max_retries=3
)

# Per-connection configuration
cdn = pycdn.pkg(
    "http://localhost:8000",
    prefix="cdn",
    timeout=60,
    api_key="your-api-key",
    region="us-east-1",
    cache_size="500MB",
    max_retries=5,
    debug=True
)

📊 Performance

• 🚀 First call: ~50-100ms (network + execution)
• ⚡ Cached calls: ~1-5ms (local cache hit)
• 💾 Memory usage: Minimal (only proxy objects stored locally)
• 🌍 Global reach: CDN edge servers reduce latency worldwide
• 📈 Scalability: Automatic scaling based on demand

🛡️ Security

• 🔒 Sandboxed execution: Each package runs in isolated environment
• 🛡️ Runtime scanning: Real-time security vulnerability detection
• 🚫 Package allowlists: Control which packages can be imported
• 🔐 API authentication: Secure CDN access with API keys
• 📝 Audit logs: Complete execution history and monitoring

🧪 Examples

Check out our comprehensive examples:

• examples/quick_import_start.py - Basic import system usage
• examples/client/advanced_import_demo.py - Advanced features showcase
• examples/client/basic_usage.py - Classic PyCDN usage
• examples/server/ - Server deployment examples

🤝 Contributing

We welcome contributions! Check out our Contributing Guide for details.

📄 License

Apache-2.0 License - see LICENSE for details.

🌟 Why PyCDN?

Traditional package management is broken:

• ❌ Long installation times
• ❌ Dependency conflicts
• ❌ Storage space waste
• ❌ Environment inconsistencies
• ❌ Version management complexity

PyCDN fixes all of this:

• ✅ Instant access - no installation needed
• ✅ Zero conflicts - packages run remotely
• ✅ Minimal storage - only proxy objects locally
• ✅ Consistent environments - CDN guarantees consistency
• ✅ Automatic updates - always use latest versions
• ✅ Natural syntax - import exactly like local packages

---

PyCDN: The Netflix of Python packages 🎬
Stream packages instantly, anywhere, anytime!