Add comprehensive DocC documentation and remove dev artifacts

- Document CSVParser with usage examples, thread safety notes, and
  performance characteristics (SIMD acceleration, RFC 4180 compliance)
- Document CSVRowView zero-copy methods with complexity annotations
- Enhance CSVEncoder/CSVDecoder with usage examples, configuration
  guides, and thread safety guarantees
- Document CSVDecodingError intelligent suggestions system
- Document CSVEncodingError with nested type handling guidance
- Document LocaleUtilities for locale-aware parsing
- Remove MIGRATION_PLAN.md development artifact

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: g-cqd

MIGRATION_PLAN.md

@@ -7,10 +7,87 @@
 
 import Foundation
 
-/// A decoder that decodes CSV data into Decodable types.
+// MARK: - CSVDecoder
+
+/// A type-safe decoder that converts CSV data to `Decodable` types.
+///
+/// `CSVDecoder` provides a familiar API similar to `JSONDecoder`, with rich
+/// configuration options for dates, numbers, booleans, key transformations,
+/// and locale-aware parsing.
+///
+/// ## Basic Usage
+///
+/// ```swift
+/// struct Person: Codable {
+///     let name: String
+///     let age: Int
+/// }
+///
+/// let csv = """
+/// name,age
+/// Alice,30
+/// Bob,25
+/// """
+///
+/// let decoder = CSVDecoder()
+/// let people: [Person] = try decoder.decode(from: csv)
+/// ```
+///
+/// ## Configuration
+///
+/// Customize decoding via ``Configuration``:
+///
+/// ```swift
+/// var config = CSVDecoder.Configuration()
+/// config.delimiter = ";"
+/// config.dateDecodingStrategy = .flexible
+/// config.keyDecodingStrategy = .convertFromSnakeCase
+///
+/// let decoder = CSVDecoder(configuration: config)
+/// ```
+///
+/// ## Thread Safety
+///
+/// `CSVDecoder` is `Sendable` and safe to share across actor boundaries.
+/// The decoder is stateless; all configuration is immutable after initialization.
+/// Multiple concurrent decodes can safely share the same decoder instance.
+///
+/// ## Performance
+///
+/// - Zero-copy parsing using ``CSVParser`` for UTF-8 data
+/// - SIMD-accelerated field scanning for large files
+/// - Streaming and parallel decoding extensions available
+///
+/// ## RFC 4180 Compliance
+///
+/// Supports both lenient (default) and strict parsing modes. Lenient mode
+/// tolerates minor violations; strict mode enforces full compliance with
+/// detailed error locations.
+///
+/// ## See Also
+///
+/// - ``CSVEncoder`` for encoding types to CSV
+/// - ``Configuration`` for available options
+/// - ``CSVParser`` for low-level parsing
 public final class CSVDecoder: Sendable {
 
-    /// Configuration for CSV parsing.
+    /// Configuration options for CSV decoding.
+    ///
+    /// All properties have sensible defaults. Customize only what you need:
+    ///
+    /// ```swift
+    /// var config = CSVDecoder.Configuration()
+    /// config.delimiter = "\t"  // Tab-separated
+    /// config.trimWhitespace = false
+    /// ```
+    ///
+    /// ## Configuration Priority
+    ///
+    /// Header resolution follows this precedence:
+    /// 1. ``indexMapping`` - Explicit column index → property name mapping
+    /// 2. ``hasHeaders`` with ``keyDecodingStrategy`` - Transform header names
+    /// 3. `CSVIndexedDecodable` conformance - Use type's `CodingKeys` order
+    /// 4. Generated column names (`column0`, `column1`, ...)
     public struct Configuration: Sendable {
         /// The delimiter character used to separate fields. Default is comma (,).
         public var delimiter: Character

Sources/CSVCoder/Core/CSVDecoder.swift

@@ -7,7 +7,28 @@
 
 import Foundation
 
-/// Location information for errors in CSV data.
+// MARK: - CSVLocation
+
+/// Precise location information for errors in CSV data.
+///
+/// `CSVLocation` provides detailed context for debugging decoding failures,
+/// including row numbers, column identifiers, coding paths, and available
+/// keys for generating helpful suggestions.
+///
+/// ## Usage
+///
+/// Access location information from ``CSVDecodingError`` via the ``CSVDecodingError/location`` property:
+///
+/// ```swift
+/// do {
+///     let items = try decoder.decode([Item].self, from: data)
+/// } catch let error as CSVDecodingError {
+///     if let location = error.location {
+///         print("Error at \(location)")
+///         // e.g., "row 5, column 'price', path: items.price"
+///     }
+/// }
+/// ```
 public struct CSVLocation: Sendable, Equatable, CustomStringConvertible {
     /// The 1-based row number in the CSV file.
     public let row: Int?
@@ -52,7 +73,39 @@ public struct CSVLocation: Sendable, Equatable, CustomStringConvertible {
     }
 }
 
-/// Errors that can occur during CSV decoding.
+// MARK: - CSVDecodingError
+
+/// Errors that occur during CSV decoding with intelligent suggestions.
+///
+/// `CSVDecodingError` provides detailed error information including:
+/// - Precise location (row, column, coding path)
+/// - Typo detection using edit distance for key mismatches
+/// - Context-aware suggestions for fixing common issues
+///
+/// ## Error Handling
+///
+/// ```swift
+/// do {
+///     let items = try decoder.decode([Item].self, from: data)
+/// } catch let error as CSVDecodingError {
+///     print(error.errorDescription ?? "Unknown error")
+///     if let suggestion = error.suggestion {
+///         print("Suggestion: \(suggestion)")
+///     }
+/// }
+/// ```
+///
+/// ## Intelligent Suggestions
+///
+/// The error system provides context-aware suggestions:
+/// - **Key not found**: Suggests similar column names using Levenshtein distance
+/// - **Type mismatch**: Suggests appropriate decoding strategies (e.g., `.flexible` for European numbers)
+/// - **Parsing errors**: Suggests delimiter or quote fixes
+///
+/// ## See Also
+///
+/// - ``CSVLocation`` for error position details
+/// - ``CSVDecoder/Configuration`` for strategy options
 public enum CSVDecodingError: Error, LocalizedError, Sendable {
     /// The data could not be decoded with the specified encoding.
     case invalidEncoding