CI: self-hosted runner (#75)

* CI: self-hosted runner

* Fix hosted runner label

* Tests: custom download location, delete when done.

Author: pcuenca

.github/workflows/unit-tests.yml

@@ -9,7 +9,7 @@ on:
 
 jobs:
   build-and-test:
-    runs-on: macos-13
+    runs-on: [ self-hosted, macOS ]
     steps:
     - uses: actions/checkout@v4
     - name: Build

Tests/HubTests/HubTests.swift

@@ -9,17 +9,26 @@ import XCTest
 
 
 class HubTests: XCTestCase {
-    override func setUp() {
-        // Put setup code here. This method is called before the invocation of each test method in the class.
-    }
+    let downloadDestination: URL = {
+        let base = FileManager.default.urls(for: .cachesDirectory, in: .userDomainMask).first!
+        return base.appending(component: "huggingface-tests")
+    }()
+
+    override func setUp() {}
 
     override func tearDown() {
-        // Put teardown code here. This method is called after the invocation of each test method in the class.
+        do {
+            try FileManager.default.removeItem(at: downloadDestination)
+        } catch {
+            print("Can't remove test download destination \(downloadDestination), error: \(error)")
+        }
     }
 
+    var hubApi: HubApi { HubApi(downloadBase: downloadDestination) }
+
     func testConfigDownload() async {
         do {
-            let configLoader = LanguageModelConfigurationFromHub(modelName: "t5-base")
+            let configLoader = LanguageModelConfigurationFromHub(modelName: "t5-base", hubApi: hubApi)
             let config = try await configLoader.modelConfig
 
             // Test leaf value (Int)
@@ -62,7 +71,7 @@ class HubTests: XCTestCase {
 
     func testConfigCamelCase() async {
         do {
-            let configLoader = LanguageModelConfigurationFromHub(modelName: "t5-base")
+            let configLoader = LanguageModelConfigurationFromHub(modelName: "t5-base", hubApi: hubApi)
             let config = try await configLoader.modelConfig
 
             // Test leaf value (Int)

Tests/TokenizersTests/FactoryTests.swift

@@ -7,16 +7,38 @@
 
 import XCTest
 import Tokenizers
+import Hub
 
-class FactoryTests: XCTestCase {
+class TestWithCustomHubDownloadLocation: XCTestCase {
+    let downloadDestination: URL = {
+        let base = FileManager.default.urls(for: .cachesDirectory, in: .userDomainMask).first!
+        return base.appending(component: "huggingface-tests")
+    }()
+
+    override func setUp() {}
+
+    override func tearDown() {
+        do {
+            try FileManager.default.removeItem(at: downloadDestination)
+        } catch {
+            print("Can't remove test download destination \(downloadDestination), error: \(error)")
+        }
+    }
+
+    var hubApi: HubApi {
+        return HubApi(downloadBase: downloadDestination)
+    }
+}
+
+class FactoryTests: TestWithCustomHubDownloadLocation {
     func testFromPretrained() async throws {
-        let tokenizer = try await AutoTokenizer.from(pretrained: "coreml-projects/Llama-2-7b-chat-coreml")
+        let tokenizer = try await AutoTokenizer.from(pretrained: "coreml-projects/Llama-2-7b-chat-coreml", hubApi: hubApi)
         let inputIds = tokenizer("Today she took a train to the West")
         XCTAssertEqual(inputIds, [1, 20628, 1183, 3614, 263, 7945, 304, 278, 3122])
     }
 
     func testWhisper() async throws {
-        let tokenizer = try await AutoTokenizer.from(pretrained: "openai/whisper-large-v2")
+        let tokenizer = try await AutoTokenizer.from(pretrained: "openai/whisper-large-v2", hubApi: hubApi)
         let inputIds = tokenizer("Today she took a train to the West")
         XCTAssertEqual(inputIds, [50258, 50363, 27676, 750, 1890, 257, 3847, 281, 264, 4055, 50257])
     }

Tests/TokenizersTests/TokenizerTests.swift

@@ -83,8 +83,8 @@ class TokenizerTester {
     private var edgeCases: [EdgeCase]? = nil
     private var _tokenizer: Tokenizer? = nil
 
-    init(hubModelName: String, encodedSamplesFilename: String, unknownTokenId: Int?) {
-        configuration = LanguageModelConfigurationFromHub(modelName: hubModelName)
+    init(hubModelName: String, encodedSamplesFilename: String, unknownTokenId: Int?, hubApi: HubApi) {
+        configuration = LanguageModelConfigurationFromHub(modelName: hubModelName, hubApi: hubApi)
         self.encodedSamplesFilename = encodedSamplesFilename
         self.unknownTokenId = unknownTokenId
 
@@ -198,17 +198,33 @@ class TokenizerTests: XCTestCase {
 
     // Known id retrieved from Python, to verify it was parsed correctly
     class var unknownTokenId: Int? { nil }
-    
+
+    static var downloadDestination: URL = {
+        let base = FileManager.default.urls(for: .cachesDirectory, in: .userDomainMask).first!
+        return base.appending(component: "huggingface-tests")
+    }()
+
+    class var hubApi: HubApi { HubApi(downloadBase: downloadDestination) }
+
     override class func setUp() {
         if let hubModelName = hubModelName, let encodedSamplesFilename = encodedSamplesFilename {
             _tester = TokenizerTester(
                 hubModelName: hubModelName,
                 encodedSamplesFilename: encodedSamplesFilename,
-                unknownTokenId: unknownTokenId
+                unknownTokenId: unknownTokenId,
+                hubApi: hubApi
             )
         }
     }
-        
+
+    override class func tearDown() {
+        do {
+            try FileManager.default.removeItem(at: downloadDestination)
+        } catch {
+            print("Can't remove test download destination \(downloadDestination), error: \(error)")
+        }
+    }
+
     func testTokenize() async {
         if let tester = Self._tester {
             await tester.testTokenize()