.keepAll buffering strategy added

Author: dehesa

Sources/Utils/Buffer.swift Sources/Active/ActiveBuffer.swiftSources/Utils/Buffer.swift renamed to Sources/Active/ActiveBuffer.swift

@@ -1,7 +1,7 @@
 import Foundation
 
 /// Buffer used to stored previously read unicode scalars.
-internal final class Buffer: IteratorProtocol {
+internal final class ScalarBuffer: IteratorProtocol {
     /// Unicode scalars read inferring configuration variables that were unknown.
     ///
     /// This buffer is reversed to make it efficient to remove elements.

Sources/Utils/BOM.swift Sources/Active/BOM.swiftSources/Utils/BOM.swift renamed to Sources/Active/BOM.swift

@@ -13,7 +13,7 @@ public final class CSVReader: IteratorProtocol, Sequence {
     /// The unicode scalar iterator providing all inputs.
     private let iterator: AnyIterator<Unicode.Scalar>
     /// Unicode scalar buffer to keep scalars that hasn't yet been analysed.
-    private let buffer: Buffer
+    private let buffer: ScalarBuffer
     /// Check whether the given unicode scalar is part of the field delimiter sequence.
     private let isFieldDelimiter: DelimiterChecker
     /// Check whether the given unicode scalar is par of the row delimiter sequence.
@@ -28,7 +28,7 @@ public final class CSVReader: IteratorProtocol, Sequence {
         self.headers = nil
         self.status = .reading
         self.iterator = AnyIterator(iterator)
-        self.buffer = Buffer()
+        self.buffer = ScalarBuffer()
         self.settings = try Settings(configuration: configuration, iterator: self.iterator, buffer: self.buffer)
 
         self.isFieldDelimiter = CSVReader.makeMatcher(delimiter: self.settings.delimiters.field, buffer: self.buffer, iterator: self.iterator)
@@ -244,7 +244,7 @@ extension CSVReader {
     /// - parameter buffer: A unicode character buffer containing further characters to parse.
     /// - parameter iterator: A unicode character buffer containing further characters to parse.
     /// - returns: A closure which given the targeted unicode character and the buffer and iterrator, returns a Boolean indicating whether there is a delimiter.
-    private static func makeMatcher(delimiter view: String.UnicodeScalarView, buffer: Buffer, iterator: AnyIterator<Unicode.Scalar>) -> DelimiterChecker {
+    private static func makeMatcher(delimiter view: String.UnicodeScalarView, buffer: ScalarBuffer, iterator: AnyIterator<Unicode.Scalar>) -> DelimiterChecker {
         // This should never be triggered.
         precondition(!view.isEmpty, "Delimiters must include at least one unicode scalar.")
 

Sources/Utils/Data.swift Sources/Active/Data.swiftSources/Utils/Data.swift renamed to Sources/Active/Data.swift

@@ -48,7 +48,7 @@ extension CSVReader {
         /// - parameter iterator: Source of the unicode scalar data. Note, that you can only iterate once through it.
         /// - parameter buffer: Buffer containing all read scalars used to infer not specified information.
         /// - throws: `CSVReader.Error` exclusively.
-        init(configuration: Configuration, iterator: AnyIterator<Unicode.Scalar>, buffer: Buffer) throws {
+        init(configuration: Configuration, iterator: AnyIterator<Unicode.Scalar>, buffer: ScalarBuffer) throws {
             switch configuration.trimStrategry {
             case .none:         self.trimCharacters = nil
             case .whitespaces:  self.trimCharacters = CharacterSet.whitespaces
@@ -95,21 +95,21 @@ extension CSVReader {
 extension CSVReader {
     /// Tries to infer the field delimiter given the row delimiter.
     /// - throws: `CSVReader.Error` exclusively.
-    fileprivate static func inferFieldDelimiter(iterator: AnyIterator<Unicode.Scalar>, rowDelimiter: String.UnicodeScalarView, buffer: Buffer) throws -> Delimiter.RawPair {
+    fileprivate static func inferFieldDelimiter(iterator: AnyIterator<Unicode.Scalar>, rowDelimiter: String.UnicodeScalarView, buffer: ScalarBuffer) throws -> Delimiter.RawPair {
         //#warning("TODO:")
         fatalError()
     }
 
     /// Tries to infer the row delimiter given the field delimiter.
     /// - throws: `CSVReader.Error` exclusively.
-    fileprivate static func inferRowDelimiter(iterator: AnyIterator<Unicode.Scalar>, fieldDelimiter: String.UnicodeScalarView, buffer: Buffer) throws -> Delimiter.RawPair {
+    fileprivate static func inferRowDelimiter(iterator: AnyIterator<Unicode.Scalar>, fieldDelimiter: String.UnicodeScalarView, buffer: ScalarBuffer) throws -> Delimiter.RawPair {
         //#warning("TODO:")
         fatalError()
     }
 
     /// Tries to infer both the field and row delimiter from the raw data.
     /// - throws: `CSVReader.Error` exclusively.
-    fileprivate static func inferDelimiters(iterator: AnyIterator<Unicode.Scalar>, buffer: Buffer) throws -> Delimiter.RawPair {
+    fileprivate static func inferDelimiters(iterator: AnyIterator<Unicode.Scalar>, buffer: ScalarBuffer) throws -> Delimiter.RawPair {
         //#warning("TODO:")
         fatalError()
     }
@@ -118,7 +118,7 @@ extension CSVReader {
 extension CSVReader {
     /// Tries to infer whether the CSV data has a header row or not.
     /// - throws: `CSVReader.Error` exclusively.
-    fileprivate static func inferHeaderStatus(iterator: AnyIterator<Unicode.Scalar>, buffer: Buffer) throws -> Bool {
+    fileprivate static func inferHeaderStatus(iterator: AnyIterator<Unicode.Scalar>, buffer: ScalarBuffer) throws -> Bool {
         //#warning("TODO:")
         fatalError()
     }

Sources/Active/Reader/Reader.swift

@@ -7,7 +7,7 @@ public final class CSVWriter {
     /// Encoder used to transform unicode scalars into a bunch of bytes.
     private let encoder: Unicode.Scalar.Encoder
     /// Unicode scalar buffer to keep scalars that hasn't yet been analysed.
-    private let buffer: Buffer
+    private let buffer: ScalarBuffer
     /// Check whether the given unicode scalar is part of the field delimiter sequence.
     private let isFieldDelimiter: DelimiterChecker
     /// Check whether the given unicode scalar is par of the row delimiter sequence.
@@ -40,7 +40,7 @@ public final class CSVWriter {
     internal init(output: (stream: OutputStream, closeAtEnd: Bool), configuration: Configuration, encoder: @escaping Unicode.Scalar.Encoder) throws {
         self.settings = try Settings(configuration: configuration)
 
-        self.buffer = Buffer(reservingCapacity: max(self.settings.delimiters.field.count, self.settings.delimiters.row.count) + 1)
+        self.buffer = ScalarBuffer(reservingCapacity: max(self.settings.delimiters.field.count, self.settings.delimiters.row.count) + 1)
         self.encoder = encoder
         self.isFieldDelimiter = CSVWriter.matchCreator(delimiter: self.settings.delimiters.field, buffer: self.buffer)
         self.isRowDelimiter = CSVWriter.matchCreator(delimiter: self.settings.delimiters.row, buffer: self.buffer)
@@ -362,7 +362,7 @@ extension CSVWriter {
     private typealias DelimiterChecker = (_ scalar: Unicode.Scalar, _ iterator: inout String.UnicodeScalarView.Iterator) -> Bool
 
     /// Creates a delimiter identifier closure.
-    private static func matchCreator(delimiter view: String.UnicodeScalarView, buffer: Buffer) -> DelimiterChecker  {
+    private static func matchCreator(delimiter view: String.UnicodeScalarView, buffer: ScalarBuffer) -> DelimiterChecker  {
         // This should never be triggered.
         precondition(!view.isEmpty, "Delimiters must include at least one unicode scalar.")
 

Sources/Active/Reader/ReaderConfiguration.swift

@@ -17,6 +17,8 @@ extension CSVDecoder {
         public var dateStrategy: Strategy.DateDecoding = .deferredToDate
         /// The strategy to use when decoding binary data.
         public var dataStrategy: Strategy.DataDecoding = .base64
+        /// The amount of CSV rows kept in memory after decoding to allow the random-order jumping exposed by keyed containers.
+        public var bufferingStrategy: Strategy.Buffering = .keepAll
 
         /// General configuration for CSV codecs and parsers.
         /// - parameter fieldDelimiter: The delimiter between CSV fields.

Sources/Utils/String+Encoding.swift Sources/Active/String+Encoding.swiftSources/Utils/String+Encoding.swift renamed to Sources/Active/String+Encoding.swift

@@ -0,0 +1,36 @@
+import Foundation
+
+extension ShadowDecoder.Source {
+    /// The decoder buffer caching the decoded CSV rows.
+    internal final class Buffer {
+        /// The buffering strategy.
+        let strategy: Strategy.Buffering
+        /// The underlying storage.
+        private var storage: [Int:[String]]
+        /// The first index being stored.
+        private var firstIndex: Int
+        /// The last index being stored.
+        private var lastIndex: Int
+        
+        /// Designated initializer.
+        init(strategy: Strategy.Buffering) {
+            self.strategy = strategy
+            self.storage = .init(minimumCapacity: 8)
+            self.firstIndex = -1
+            self.lastIndex = -1
+            #warning("TODO: Buffering strategy")
+        }
+        
+        /// Stores the given row at the given position.
+        func store(_ row: [String], at index: Int) {
+            precondition(index >= 0)
+            self.storage[index] = row
+        }
+        
+        /// Retrieves the row at the given position.
+        func retrieve(at index: Int) -> [String]? {
+            precondition(index >= 0)
+            return self.storage[index]
+        }
+    }
+}