bug: fix ephemeral container state when DATA_DIR is not set

[nullable-eth]

Fix: Restore Ephemeral Mode When DATA_DIR is Not Set

🐛 Issue Summary

PR #9 introduced persistent storage as a major enhancement, but inadvertently broke the application's previous ephemeral behavior when DATA_DIR is not explicitly set. This caused the application to always attempt to create a /data directory, leading to permission errors in Docker containers and breaking backward compatibility.

Related Issues:

• Issue #12 - Permission denied when creating /data directory

📋 Root Cause Analysis

What PR #9 Changed

PR #9 added several excellent features:

• ✅ Persistent Storage - Tracks processed items across restarts
• ✅ Radarr/Sonarr Integration - Automatic TMDb ID detection
• ✅ Keyword Normalization - Intelligent formatting
• ✅ Verbose Logging - Detailed debugging
• ✅ Force Update Mode - Reprocess all items

The Problem

However, the persistent storage implementation made storage mandatory rather than optional:

// Before this PR - always tries to create storage
DataDir: getEnvWithDefault("DATA_DIR", "/data"),

// In NewProcessor - always initializes storage
stor, err := storage.NewStorage(cfg.DataDir)
if err != nil {
    return nil, fmt.Errorf("failed to initialize storage: %w", err)
}

This meant that even when users didn't set DATA_DIR, the application would:

1. Default to /data directory
2. Try to create it with os.MkdirAll(dataDir, 0755)
3. Fail with permission errors in Docker containers running as non-root

The Error

❌ Failed to initialize processor: failed to initialize storage: failed to create data directory: mkdir /data: permission denied

🔧 Solution

This PR restores the ephemeral behavior when DATA_DIR is not set while preserving all the new functionality when it is set.

Changes Made

1. Optional Storage Configuration

// OLD - Always defaults to /data
DataDir: getEnvWithDefault("DATA_DIR", "/data"),

// NEW - Only set when explicitly provided
DataDir: os.Getenv("DATA_DIR"), // No default - ephemeral if not set

2. Conditional Storage Initialization

// NEW - Only create storage if DataDir is set
var stor *storage.Storage
if cfg.DataDir != "" {
    var err error
    stor, err = storage.NewStorage(cfg.DataDir)
    if err != nil {
        return nil, fmt.Errorf("failed to initialize storage: %w", err)
    }
}

3. Conditional Storage Operations

// NEW - Check if storage exists before using it
if p.storage != nil {
    if processed, storageExists := p.storage.Get(item.GetRatingKey()); storageExists && processed.KeywordsSynced && processed.UpdateField == p.config.UpdateField && !p.config.ForceUpdate {
        // Skip if already processed
        continue
    }
}

// NEW - Only save if storage is enabled
if p.storage != nil {
    if err := p.storage.Set(processedItem); err != nil {
        fmt.Printf("⚠️ Warning: Failed to save processed item to storage: %v\n", err)
    }
}

4. Clear User Feedback

// NEW - Clear messaging about storage mode
if stor != nil {
    count := stor.Count()
    if count > 0 {
        fmt.Printf("📁 Loaded %d previously processed items from storage\n", count)
    }
} else {
    fmt.Printf("🔄 Running in ephemeral mode - no persistent storage (set DATA_DIR to enable)\n")
}

📊 Behavior Comparison

Before This Fix

DATA_DIR Set	Behavior	Result
❌ Not set	Always tries to create /data	❌ Permission error
✅ Set to path	Creates storage at path	✅ Persistent storage

After This Fix

DATA_DIR Set	Behavior	Result
❌ Not set	Runs ephemerally	✅ Processes all items every run
✅ Set to path	Creates storage at path	✅ Persistent storage

🧪 Testing

Build Success

go build -o labelarr ./cmd/labelarr
# ✅ Builds successfully

Test Results

go test ./...
# ✅ All tests pass

Runtime Behavior

Without DATA_DIR:

🔄 Running in ephemeral mode - no persistent storage (set DATA_DIR to enable)

With DATA_DIR:

📁 Loaded 1250 previously processed items from storage

🔄 Migration Guide

For Users Running Ephemerally (No Change Required)

# This will now work without permission errors
services:
  labelarr:
    image: ghcr.io/nullable-eth/labelarr:latest
    environment:
      - PLEX_TOKEN=your_token
      - TMDB_READ_ACCESS_TOKEN=your_token
      # No DATA_DIR needed - runs ephemerally

For Users Wanting Persistent Storage

# Add DATA_DIR and volume mount
services:
  labelarr:
    image: ghcr.io/nullable-eth/labelarr:latest
    volumes:
      - ./labelarr-data:/data
    environment:
      - PLEX_TOKEN=your_token
      - TMDB_READ_ACCESS_TOKEN=your_token
      - DATA_DIR=/data  # Enable persistent storage

🎯 Benefits

1. ✅ Backward Compatibility - Restores ephemeral behavior when DATA_DIR not set
2. ✅ No Permission Errors - No more Docker permission issues
3. ✅ Clear User Control - Explicit opt-in for persistent storage
4. ✅ Preserves All Features - Radarr/Sonarr, normalization, verbose logging all work
5. ✅ Graceful Degradation - Application works with or without storage

🚀 Impact

This fix resolves the breaking change introduced in PR #9 while preserving all the valuable new functionality. Users can now:

• Run the application without any storage concerns (ephemeral mode)
• Explicitly enable persistent storage when desired
• Migrate from ephemeral to persistent operation seamlessly

The application now behaves correctly in both modes, restoring the flexibility that existed before PR #9 while maintaining all the new capabilities.