Improve reader delimiter handling

Author: dehesa

.github/workflows/tests.yml

@@ -1,19 +1,19 @@
-name: Tests
+name: Unit Tests
 on: [ push, pull_request ]
 
 jobs:
-  tests_on_macOS:
-    name: Tests on macOS
+  unittests_on_macOS:
+    name: Unit tests on macOS
     runs-on: macos-latest
     steps:
       - uses: actions/checkout@v2
       - name: Build
         run: swift build -v
       - name: Run tests
-        run: swift test -v
+        run: swift test --filter UnitTests -v
 
-  tests_on_Ubuntu:
-    name: Tests on Ubuntu
+  unittests_on_Ubuntu:
+    name: Unit tests on Ubuntu
     runs-on: ubuntu-latest
     container:
       image: swift:latest
@@ -22,4 +22,4 @@ jobs:
       - name: Build
         run: swift build -v
       - name: Run tests
-        run: swift test -v --enable-test-discovery
+        run: swift test --filter UnitTests --enable-test-discovery -v

Package.swift

@@ -8,12 +8,11 @@ let package = Package(
     ],
     products: [
         .library(name: "CodableCSV", targets: ["CodableCSV"]),
-//        .executable(name: "CodableCSV-Benchmarks", targets: ["CodableCSV-Benchmarks"])
     ],
     dependencies: [],
     targets: [
         .target(name: "CodableCSV", dependencies: [], path: "sources"),
-        .testTarget(name: "CodableCSVTests", dependencies: ["CodableCSV"], path: "tests"),
-//        .target(name: "CodableCSV-Benchmarks", dependencies: ["CodableCSV"], path: "benchmarks"),
+        .testTarget(name: "UnitTests", dependencies: ["CodableCSV"], path: "tests"),
+        .testTarget(name: "Benchmarks", dependencies: ["CodableCSV"], path: "benchmarks")
     ]
 )

README.md

@@ -11,13 +11,14 @@
 
 [CodableCSV](https://github.com/dehesa/CodableCSV) provides:
 
--   Row-by-row CSV reader & writer.
--   Codable interface.
--   Support for multiple inputs/outputs: `String`s, `Data` blobs, and CSV files (represented by `URL` addresses).
--   CSV encoding & configuration inference (e.g. what field/row delimiters are being used).
+-   Imperative CSV reader/writer (row-by-row and/or field-by-field).
+-   Declarative `Codable` encoder/decoder.
+-   Support for multiple inputs/outputs: `String`s, `Data` blobs, and `URL`s.
+-   Support for multiple string encodings and Byte Order Markers (BOM).
+-   Extremely configurable: delimiters, escaping scalar, trim strategy, presampling, and numerous codable strategies.
 -   Multiplatform support with no dependencies.
 
-> The Swift standard library and Foundation are considered implicit requirements.
+> The Swift Standard Library and Foundation are considered implicit requirements.
 
 # Usage
 
@@ -64,7 +65,7 @@ import CodableCSV
 
 There are two ways to use this library:
 
-1. imperatively, as a row-by-row and field-by-field reader or writer.
+1. imperatively, as a row-by-row and field-by-field reader/writer.
 2. declaratively, through Swift's `Codable` interface.
 
 ## Imperative Reader/Writer

benchmarks/PerformanceTests.swift

@@ -0,0 +1,19 @@
+import XCTest
+import CodableCSV
+
+/// Tests checking the regular encoding usage.
+final class PerformanceTests: XCTestCase {
+    override func setUp() {
+        self.continueAfterFailure = false
+    }
+}
+
+extension PerformanceTests {
+    /// Tests the encoding of an empty.
+    func testEmptyFile() throws {
+//        XCTSkipUnless(<#T##expression: Bool##Bool#>, <#T##message: String?##String?#>)
+        #if !DEBUG
+        print("Hello RELEASE")
+        #endif
+    }
+}

benchmarks/main.swift

@@ -54,28 +54,29 @@ extension CSVReader {
         /// - throws: `CSVError<CSVReader>` exclusively.
         init(configuration: Configuration, decoder: ScalarDecoder, buffer: ScalarBuffer) throws {
             // 1. Figure out the field and row delimiters.
-            switch (configuration.delimiters.field.rawValue, configuration.delimiters.row.rawValue) {
-            case (nil, nil):
+            let (field, row) = (configuration.delimiters.field.rawValue, configuration.delimiters.row.rawValue)
+            switch (field.isEmpty, row.isEmpty) {
+            case (true,  true):
                 self.delimiters = try CSVReader.inferDelimiters(decoder: decoder, buffer: buffer)
-            case (nil, let row):
+            case (true,  false):
                 self.delimiters = try CSVReader.inferFieldDelimiter(rowDelimiter: row, decoder: decoder, buffer: buffer)
-            case (let field, nil):
+            case (false, true):
                 self.delimiters = try CSVReader.inferRowDelimiter(fieldDelimiter: field, decoder: decoder, buffer: buffer)
-            case (let field, let row) where !field.elementsEqual(row):
+            case (false, false) where field.elementsEqual(row):
+                throw Error.invalidDelimiters(field)
+            default:
                 self.delimiters = (.init(field), .init(row))
-            case (let delimiter, _):
-                throw Error.invalidDelimiters(delimiter)
             }
             // 2. Set the escaping scalar.
             self.escapingScalar = configuration.escapingStrategy.scalar
             // 3. Set the trim characters set.
             self.trimCharacters = configuration.trimStrategry
             // 4. Ensure trim character set doesn't contain the field delimiter.
-            guard delimiters.field.allSatisfy({ !self.trimCharacters.contains($0) }) else {
+            guard self.delimiters.field.allSatisfy({ !self.trimCharacters.contains($0) }) else {
                 throw Error.invalidTrimCharacters(self.trimCharacters, delimiter: configuration.delimiters.field.rawValue)
             }
             // 5. Ensure trim character set doesn't contain the row delimiter.
-            guard delimiters.row.allSatisfy({ !self.trimCharacters.contains($0) }) else {
+            guard self.delimiters.row.allSatisfy({ !self.trimCharacters.contains($0) }) else {
                 throw Error.invalidTrimCharacters(self.trimCharacters, delimiter: configuration.delimiters.row.rawValue)
             }
             // 6. Ensure trim character set does not include escaping scalar

sources/imperative/reader/internal/ReaderInternals.swift

@@ -175,8 +175,14 @@ extension CSVWriter {
 }
 
 // MARK: -
-
 extension CSVWriter {
+    /// Writes the given delimiter using the instance's `encoder`.
+    /// - parameter delimiter: The array of `Unicode.Scalar` representing a delimiter.
+    /// - throws: `CSVError<CSVWriter>` exclusively.
+    @inline(__always) private func lowlevelWrite(delimiter: [Unicode.Scalar]) throws {
+        try delimiter.forEach { try self.encoder($0) }
+    }
+    
     /// Writes the given `String` into the receiving writer's stream.
     /// - parameter field: The field to be checked for characters to escape and subsequently written.
     /// - throws: `CSVError<CSVWriter>` exclusively.
@@ -239,13 +245,6 @@ extension CSVWriter {
 
         try result.forEach { try self.encoder($0) }
     }
-    
-    /// Writes the given delimiter using the instance's `encoder`.
-    /// - parameter delimiter: The array of `Unicode.Scalar` representing a delimiter.
-    /// - throws: `CSVError<CSVWriter>` exclusively.
-    @inline(__always) private func lowlevelWrite(delimiter: [Unicode.Scalar]) throws {
-        try delimiter.forEach { try self.encoder($0) }
-    }
 }
 
 fileprivate extension CSVWriter.Error {

sources/imperative/writer/Writer.swift

@@ -44,7 +44,7 @@ extension CSVWriter {
             if field.isEmpty || row.isEmpty {
                 throw Error.invalidEmptyDelimiter()
             } else if field.elementsEqual(row) {
-                throw Error.invalidDelimiters(field)
+                throw Error.invalidSameDelimiters(field)
             } else {
                 self.delimiters = (.init(field), .init(row))
             }
@@ -67,7 +67,7 @@ fileprivate extension CSVWriter.Error {
 
     /// Error raised when the field and row delimiters are the same.
     /// - parameter delimiter: The indicated field and row delimiters.
-    static func invalidDelimiters(_ delimiter: String.UnicodeScalarView) -> CSVError<CSVWriter> {
+    static func invalidSameDelimiters(_ delimiter: String.UnicodeScalarView) -> CSVError<CSVWriter> {
         .init(.invalidConfiguration,
               reason: "The field and row delimiters cannot be the same.",
               help: "Set different delimiters for field and rows.",

sources/imperative/writer/internal/WriterInternals.swift

@@ -1,5 +1,5 @@
 import XCTest
-@testable import CodableCSV
+import CodableCSV
 
 /// Tests for empty or single value CSVs.
 final class DecodingSinglesTests: XCTestCase {
@@ -8,6 +8,29 @@ final class DecodingSinglesTests: XCTestCase {
     }
 }
 
+extension DecodingSinglesTests {
+    /// The coding key used to identify encoding/decoding containers.
+    private struct IndexKey: CodingKey {
+        /// The integer value of the coding key.
+        let index: Int
+        /// Designated initializer.
+        init(_ index: Int) { self.index = index }
+        
+        init?(intValue: Int) {
+            guard intValue >= 0 else { return nil }
+            self.init(intValue)
+        }
+        
+        init?(stringValue: String) {
+            guard let intValue = Int(stringValue) else { return nil }
+            self.init(intValue)
+        }
+        
+        var stringValue: String { String(self.index) }
+        var intValue: Int? { self.index }
+    }
+}
+
 extension DecodingSinglesTests {
     /// Tests the decoding of a completely empty file.
     func testEmptyFile() throws {

tests/CodableTests/DecodingSinglesTests.swift

@@ -1,5 +1,5 @@
 import XCTest
-@testable import CodableCSV
+import CodableCSV
 
 /// Tests for the decodable car dealer data.
 final class DecodingWrappersTests: XCTestCase {