{

"cells": [

{

"type": "markdown",

"id": "md_01",

"content": "# Data Storage Example\n\nThis notebook demonstrates how to use the explicit storage system to persist data across notebook sessions.\n\n## Key Concepts\n\n1. **`storage.set(key, value)`** - Store data with explicit serialization\n2. **`storage.get(key)`** - Retrieve stored data\n3. **`storage.has(key)`** - Check if data exists\n4. **`storage.keys()`** - List all stored keys\n5. **User Controls Serialization** - You decide how to serialize/deserialize your data\n\nThis approach gives you full control over data persistence and avoids automatic overriding issues."

},

{

"type": "code",

"id": "code_01",

"code": "// Create a DataFrame and store it explicitly\nconst dfd = require('danfojs');\n\n// Create sample data\nconst data = {\n name: ['Alice', 'Bob', 'Charlie', 'Diana'],\n age: [25, 30, 35, 28],\n city: ['New York', 'London', 'Tokyo', 'Paris'],\n salary: [75000, 85000, 95000, 70000]\n};\n\nconst df = new dfd.DataFrame(data);\nconsole.log('Created DataFrame:');\nconsole.log('Shape:', df.shape);\nconsole.log('Columns:', df.columns);\n\n// Explicit serialization - user controls the format\nconst serializedData = {\n columns: df.columns,\n values: df.values,\n shape: df.shape,\n dtypes: df.dtypes\n};\n\n// Store in the notebook's persistent storage\nstorage.set('employee_data', serializedData);\nstorage.set('last_modified', new Date().toISOString());\n\nconsole.log('Data stored in notebook storage');\nconsole.log('Storage keys:', storage.keys());\n\n// Export for reactive use in other cells\nexports.df = df;\nexports.originalDataShape = df.shape;"

},

{

"type": "markdown",

"id": "md_02",

"content": "## Data Stored Successfully\n\nThe DataFrame has been stored using:\n- **storage.set('employee_data', serializedData)** - Custom serialization format\n- **storage.set('last_modified', timestamp)** - Simple timestamp storage\n\nThe storage contains **{{storage.keys().length}}** entries: {{storage.keys().join(', ')}}\n\nStored at: **{{storage.get('last_modified')}}**"

},

{

"type": "code",

"id": "code_02",

"code": "// Restore data from storage (this would work even after notebook reload)\nif (storage.has('employee_data')) {\n console.log('Found stored data! Restoring...');\n \n const storedData = storage.get('employee_data');\n console.log('Stored data shape:', storedData.shape);\n \n // Reconstruct DataFrame using danfojs\n const restoredDf = new dfd.DataFrame(storedData.values, { columns: storedData.columns });\n \n console.log('Restored DataFrame:');\n console.log('Shape:', restoredDf.shape);\n console.log('Columns:', restoredDf.columns);\n \n // Show first few rows\n restoredDf.head().print();\n \n // Export the restored DataFrame\n exports.restoredDf = restoredDf;\n \n} else {\n console.log('No stored data found');\n}\n\n// Show storage info\nconsole.log('\\n=== Storage Info ===');\nconsole.log('Total keys:', storage.keys().length);\nconsole.log('Keys:', storage.keys());\nconsole.log('Last modified:', storage.get('last_modified'));"

},

{

"type": "code",

"id": "code_03",

"code": "// Demonstrate data editing and re-storage\nif (restoredDf) {\n console.log('Modifying the restored DataFrame...');\n \n // Add a new column\n const bonuses = [5000, 7000, 10000, 4000];\n restoredDf.addColumn('bonus', bonuses, { inplace: true });\n \n console.log('Added bonus column:');\n restoredDf.head().print();\n \n // Store the modified version\n const modifiedData = {\n columns: restoredDf.columns,\n values: restoredDf.values,\n shape: restoredDf.shape,\n dtypes: restoredDf.dtypes,\n version: 2 // Track versions\n };\n \n storage.set('employee_data', modifiedData);\n storage.set('last_modified', new Date().toISOString());\n \n console.log('\\nUpdated data stored!');\n console.log('New shape:', restoredDf.shape);\n \n // Export the modified DataFrame\n exports.modifiedDf = restoredDf;\n \n} else {\n console.log('No DataFrame to modify (run previous cell first)');\n}"

},

{

"type": "markdown",

"id": "md_04",

"content": "## Key Benefits of Explicit Storage\n\n### ✅ **User Control**\n- **You decide** what to store and how to serialize it\n- **Custom formats**: Use DataFrame.toJSON(), CSV, or any format\n- **Versioning**: Add version numbers, timestamps, metadata\n\n### ✅ **No Override Issues**\n- **Explicit restore**: Only restore when you call `storage.get()`\n- **Conditional logic**: Check if stored data exists before using\n- **Cell independence**: Each cell controls its own data flow\n\n### ✅ **Flexible Serialization**\n```javascript\n// DataFrame with custom serialization\nstorage.set('df_csv', df.toCSV());\nstorage.set('df_json', df.toJSON());\nstorage.set('df_custom', {\n data: df.values,\n meta: { created: Date.now(), user: 'me' }\n});\n\n// Restore with appropriate method\nconst df1 = dfd.readCSV(storage.get('df_csv'));\nconst df2 = dfd.readJSON(storage.get('df_json'));\nconst custom = storage.get('df_custom');\n```\n\n### ✅ **Storage Management**\n```javascript\n// Check what's stored\nconsole.log('Keys:', storage.keys());\nconsole.log('Has data:', storage.has('my_data'));\n\n// Clean up\nstorage.delete('old_data');\nstorage.clear(); // Remove everything\n```"

},

{

"type": "code",

"id": "code_04",

"code": "// Advanced storage patterns\n\n// 1. Versioned storage with metadata\nfunction storeDataWithMetadata(key, data, metadata = {}) {\n const entry = {\n data: data,\n metadata: {\n version: 1,\n created: new Date().toISOString(),\n ...metadata\n }\n };\n \n // Check if exists and increment version\n if (storage.has(key)) {\n const existing = storage.get(key);\n entry.metadata.version = (existing.metadata?.version || 0) + 1;\n entry.metadata.previousVersion = existing.metadata;\n }\n \n storage.set(key, entry);\n return entry.metadata.version;\n}\n\n// 2. Compressed storage for large data\nfunction storeCompressed(key, data) {\n // In a real scenario, you might use a compression library\n const serialized = JSON.stringify(data);\n const compressed = serialized; // Placeholder for compression\n \n storage.set(key, {\n type: 'compressed',\n data: compressed,\n originalSize: serialized.length,\n compressedSize: compressed.length\n });\n}\n\n// 3. Storage with expiration\nfunction storeWithExpiration(key, data, expirationHours = 24) {\n const entry = {\n data: data,\n expires: new Date(Date.now() + expirationHours * 60 * 60 * 1000).toISOString()\n };\n storage.set(key, entry);\n}\n\nfunction getWithExpiration(key) {\n if (!storage.has(key)) return null;\n \n const entry = storage.get(key);\n if (entry.expires && new Date() > new Date(entry.expires)) {\n storage.delete(key);\n return null;\n }\n \n return entry.data;\n}\n\n// Demo the advanced patterns\nif (modifiedDf) {\n // Store with metadata\n const version = storeDataWithMetadata('employee_versioned', {\n columns: modifiedDf.columns,\n values: modifiedDf.values,\n shape: modifiedDf.shape\n }, {\n user: 'demo',\n description: 'Employee data with bonuses'\n });\n \n console.log(`Stored employee data as version ${version}`);\n \n // Store with expiration (expires in 1 hour)\n storeWithExpiration('temp_calc', { result: 42, computation: 'demo' }, 1);\n \n console.log('\\n=== Storage Summary ===');\n storage.keys().forEach(key => {\n const value = storage.get(key);\n console.log(`${key}:`, typeof value === 'object' ? JSON.stringify(value).substring(0, 100) + '...' : value);\n });\n \n} else {\n console.log('Run previous cells first to see advanced storage patterns');\n}\n\nexports.storeDataWithMetadata = storeDataWithMetadata;\nexports.getWithExpiration = getWithExpiration;"

},

{

"type": "markdown",

"id": "md_05",

"content": "## Summary\n\nThe explicit storage system provides:\n\n1. **Full Control**: Users decide what, when, and how to store data\n2. **No Conflicts**: Explicit restore prevents automatic override issues\n3. **Flexible Serialization**: Support for any data format (CSV, JSON, binary, etc.)\n4. **Metadata Support**: Add versions, timestamps, descriptions\n5. **Storage Management**: List, check, delete, and clear stored data\n\nThis approach is much more robust than automatic serialization because:\n- **DataFrames can use optimized serialization** (toCSV, toJSON, etc.)\n- **Users control the data lifecycle** (when to save/restore)\n- **No surprise overrides** from cell re-execution\n- **Version control and metadata** can be added as needed\n\nThe storage persists with the notebook file, so when you save and reload the notebook, your data is still there!"

}

]

}