Recipe scraper implementation (unfinished)

2023-11-07 21:28:24 +01:00
parent 1343d8c658
commit 3217c7b4aa
6 changed files with 234 additions and 3 deletions
--- a/Client.xcodeproj/project.pbxproj
+++ b/Client.xcodeproj/project.pbxproj
@@ -32,8 +32,10 @@
 		A703226F2ABB1DD700D7C4ED /* ColorExtension.swift in Sources */ = {isa = PBXBuildFile; fileRef = A703226E2ABB1DD700D7C4ED /* ColorExtension.swift */; };
 		A70D7CA12AC73CA800D53DBF /* RecipeEditView.swift in Sources */ = {isa = PBXBuildFile; fileRef = A70D7CA02AC73CA700D53DBF /* RecipeEditView.swift */; };
 		A70D7CA32AC74B3B00D53DBF /* DateExtension.swift in Sources */ = {isa = PBXBuildFile; fileRef = A70D7CA22AC74B3B00D53DBF /* DateExtension.swift */; };
+		A74D33BE2AF82AAE00D06555 /* SwiftSoup in Frameworks */ = {isa = PBXBuildFile; productRef = A74D33BD2AF82AAE00D06555 /* SwiftSoup */; };
 		A76B8A6F2ADFFA8800096CEC /* SupportedLanguage.swift in Sources */ = {isa = PBXBuildFile; fileRef = A76B8A6E2ADFFA8800096CEC /* SupportedLanguage.swift */; };
 		A76B8A712AE002AE00096CEC /* AlertHandler.swift in Sources */ = {isa = PBXBuildFile; fileRef = A76B8A702AE002AE00096CEC /* AlertHandler.swift */; };
+		A781E7612AF822D000452F6F /* RecipeScraper.swift in Sources */ = {isa = PBXBuildFile; fileRef = A781E7602AF822CF00452F6F /* RecipeScraper.swift */; };
 		A7AEAE642AD5521400135378 /* Localizable.xcstrings in Resources */ = {isa = PBXBuildFile; fileRef = A7AEAE632AD5521400135378 /* Localizable.xcstrings */; };
 		A7F3F8E82ACBFC760076C227 /* KeywordPickerView.swift in Sources */ = {isa = PBXBuildFile; fileRef = A7F3F8E72ACBFC760076C227 /* KeywordPickerView.swift */; };
 		A7F3F8EA2ACC221C0076C227 /* CategoryPickerView.swift in Sources */ = {isa = PBXBuildFile; fileRef = A7F3F8E92ACC221C0076C227 /* CategoryPickerView.swift */; };
@@ -86,8 +88,10 @@
 		A703226E2ABB1DD700D7C4ED /* ColorExtension.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = ColorExtension.swift; sourceTree = "<group>"; };
 		A70D7CA02AC73CA700D53DBF /* RecipeEditView.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = RecipeEditView.swift; sourceTree = "<group>"; };
 		A70D7CA22AC74B3B00D53DBF /* DateExtension.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = DateExtension.swift; sourceTree = "<group>"; };
+		A74D33BF2AF82CB500D06555 /* Scraper.playground */ = {isa = PBXFileReference; lastKnownFileType = file.playground; path = Scraper.playground; sourceTree = "<group>"; xcLanguageSpecificationIdentifier = xcode.lang.swift; };
 		A76B8A6E2ADFFA8800096CEC /* SupportedLanguage.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = SupportedLanguage.swift; sourceTree = "<group>"; };
 		A76B8A702AE002AE00096CEC /* AlertHandler.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = AlertHandler.swift; sourceTree = "<group>"; };
+		A781E7602AF822CF00452F6F /* RecipeScraper.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = RecipeScraper.swift; sourceTree = "<group>"; };
 		A7AEAE632AD5521400135378 /* Localizable.xcstrings */ = {isa = PBXFileReference; lastKnownFileType = text.json.xcstrings; path = Localizable.xcstrings; sourceTree = "<group>"; };
 		A7F3F8E72ACBFC760076C227 /* KeywordPickerView.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = KeywordPickerView.swift; sourceTree = "<group>"; };
 		A7F3F8E92ACC221C0076C227 /* CategoryPickerView.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = CategoryPickerView.swift; sourceTree = "<group>"; };
@@ -98,6 +102,7 @@
 			isa = PBXFrameworksBuildPhase;
 			buildActionMask = 2147483647;
 			files = (
+				A74D33BE2AF82AAE00D06555 /* SwiftSoup in Frameworks */,
 			);
 			runOnlyForDeploymentPostprocessing = 0;
 		};
@@ -149,6 +154,7 @@
 				A70171BA2AB4980100064C43 /* Views */,
 				A70171B72AB2445700064C43 /* ViewModels */,
 				A70171B22AB211F000064C43 /* Network */,
+				A781E75F2AF8228100452F6F /* RecipeImport */,
 				A703226B2ABAF60D00D7C4ED /* Extensions */,
 				A7AEAE632AD5521400135378 /* Localizable.xcstrings */,
 				A70171852AA8E71F00064C43 /* Assets.xcassets */,
@@ -246,6 +252,15 @@
 			path = Screenshots;
 			sourceTree = "<group>";
 		};
+		A781E75F2AF8228100452F6F /* RecipeImport */ = {
+			isa = PBXGroup;
+			children = (
+				A781E7602AF822CF00452F6F /* RecipeScraper.swift */,
+				A74D33BF2AF82CB500D06555 /* Scraper.playground */,
+			);
+			path = RecipeImport;
+			sourceTree = "<group>";
+		};
 /* End PBXGroup section */

 /* Begin PBXNativeTarget section */
@@ -262,6 +277,9 @@
 			dependencies = (
 			);
 			name = "Nextcloud Cookbook iOS Client";
+			packageProductDependencies = (
+				A74D33BD2AF82AAE00D06555 /* SwiftSoup */,
+			);
 			productName = "Nextcloud Cookbook iOS Client";
 			productReference = A701717E2AA8E71900064C43 /* Nextcloud Cookbook iOS Client.app */;
 			productType = "com.apple.product-type.application";
@@ -337,6 +355,9 @@
 				fr,
 			);
 			mainGroup = A70171752AA8E71900064C43;
+			packageReferences = (
+				A74D33BC2AF82AAE00D06555 /* XCRemoteSwiftPackageReference "SwiftSoup" */,
+			);
 			productRefGroup = A701717F2AA8E71900064C43 /* Products */;
 			projectDirPath = "";
 			projectRoot = "";
@@ -400,6 +421,7 @@
 				A703226A2ABAF49800D7C4ED /* JSONCoderExtension.swift in Sources */,
 				A703226D2ABAF90D00D7C4ED /* APIController.swift in Sources */,
 				A70171822AA8E71900064C43 /* Nextcloud_Cookbook_iOS_ClientApp.swift in Sources */,
+				A781E7612AF822D000452F6F /* RecipeScraper.swift in Sources */,
 				A70171AD2AA8EF4700064C43 /* MainViewModel.swift in Sources */,
 				A76B8A6F2ADFFA8800096CEC /* SupportedLanguage.swift in Sources */,
 				A70171C92AB4CBB400064C43 /* OnboardingView.swift in Sources */,
@@ -579,7 +601,7 @@
 				LD_RUNPATH_SEARCH_PATHS = "@executable_path/Frameworks";
 				"LD_RUNPATH_SEARCH_PATHS[sdk=macosx*]" = "@executable_path/../Frameworks";
 				MACOSX_DEPLOYMENT_TARGET = 13.0;
-				MARKETING_VERSION = 1.3;
+				MARKETING_VERSION = 1.4;
 				PRODUCT_BUNDLE_IDENTIFIER = "VincentMeilinger.Nextcloud-Cookbook-iOS-Client";
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				SDKROOT = auto;
@@ -620,7 +642,7 @@
 				LD_RUNPATH_SEARCH_PATHS = "@executable_path/Frameworks";
 				"LD_RUNPATH_SEARCH_PATHS[sdk=macosx*]" = "@executable_path/../Frameworks";
 				MACOSX_DEPLOYMENT_TARGET = 13.0;
-				MARKETING_VERSION = 1.3;
+				MARKETING_VERSION = 1.4;
 				PRODUCT_BUNDLE_IDENTIFIER = "VincentMeilinger.Nextcloud-Cookbook-iOS-Client";
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				SDKROOT = auto;
@@ -761,6 +783,25 @@
 			defaultConfigurationName = Release;
 		};
 /* End XCConfigurationList section */
+
+/* Begin XCRemoteSwiftPackageReference section */
+		A74D33BC2AF82AAE00D06555 /* XCRemoteSwiftPackageReference "SwiftSoup" */ = {
+			isa = XCRemoteSwiftPackageReference;
+			repositoryURL = "https://github.com/scinfu/SwiftSoup.git";
+			requirement = {
+				kind = upToNextMajorVersion;
+				minimumVersion = 2.6.1;
+			};
+		};
+/* End XCRemoteSwiftPackageReference section */
+
+/* Begin XCSwiftPackageProductDependency section */
+		A74D33BD2AF82AAE00D06555 /* SwiftSoup */ = {
+			isa = XCSwiftPackageProductDependency;
+			package = A74D33BC2AF82AAE00D06555 /* XCRemoteSwiftPackageReference "SwiftSoup" */;
+			productName = SwiftSoup;
+		};
+/* End XCSwiftPackageProductDependency section */
 	};
 	rootObject = A70171762AA8E71900064C43 /* Project object */;
 }
--- a/Client.xcodeproj/project.xcworkspace/xcshareddata/swiftpm/Package.resolved
+++ b/Client.xcodeproj/project.xcworkspace/xcshareddata/swiftpm/Package.resolved
@@ -0,0 +1,14 @@
+{
+  "pins" : [
+    {
+      "identity" : "swiftsoup",
+      "kind" : "remoteSourceControl",
+      "location" : "https://github.com/scinfu/SwiftSoup.git",
+      "state" : {
+        "revision" : "8b6cf29eead8841a1fa7822481cb3af4ddaadba6",
+        "version" : "2.6.1"
+      }
+    }
+  ],
+  "version" : 2
+}
--- a/Client.xcodeproj/xcuserdata/vincentmeilinger.xcuserdatad/xcschemes/xcschememanagement.plist
+++ b/Client.xcodeproj/xcuserdata/vincentmeilinger.xcuserdatad/xcschemes/xcschememanagement.plist
@@ -4,10 +4,52 @@
 <dict>
 	<key>SchemeUserState</key>
 	<dict>
+		<key>MyPlayground (Playground) 1.xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>2</integer>
+		</dict>
+		<key>MyPlayground (Playground) 2.xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>3</integer>
+		</dict>
+		<key>MyPlayground (Playground).xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>0</integer>
+		</dict>
 		<key>Nextcloud Cookbook iOS Client.xcscheme_^#shared#^_</key>
 		<dict>
 			<key>orderHint</key>
-			<integer>0</integer>
+			<integer>1</integer>
+		</dict>
+		<key>Scraper (Playground) 1.xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>5</integer>
+		</dict>
+		<key>Scraper (Playground) 2.xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>6</integer>
+		</dict>
+		<key>Scraper (Playground).xcscheme</key>
+		<dict>
+			<key>isShown</key>
+			<false/>
+			<key>orderHint</key>
+			<integer>4</integer>
 		</dict>
 	</dict>
 </dict>
--- a/Client/RecipeImport/RecipeScraper.swift
+++ b/Client/RecipeImport/RecipeScraper.swift
@@ -0,0 +1,67 @@
+//
+//  RecipeScraper.swift
+//  Nextcloud Cookbook iOS Client
+//
+//  Created by Vincent Meilinger on 05.11.23.
+//
+
+import Foundation
+import SwiftSoup
+
+class RecipeScraper {
+    func scrape(url: String) -> RecipeDetail? {
+        var contents: String? = nil
+        if let url = URL(string: url) {
+            do {
+                contents = try String(contentsOf: url)
+            } catch {
+                print("ERROR: Could not load url content.")
+            }
+            
+        } else {
+            print("ERROR: Bad url.")
+        }
+
+        guard let html = contents else {
+            print("ERROR: no contents")
+            exit(1)
+        }
+        
+        let doc: Document = try SwiftSoup.parse(html)
+        let elements: Elements = try doc.select("script")
+        for elem in elements.array() {
+            for attr in elem.getAttributes()!.asList() {
+                if attr.getValue() == "application/ld+json" {
+                    toDict(elem)
+                }
+            }
+        }
+    }
+    
+    
+    private func toDict(_ elem: Element) -> [String: Any] {
+        do {
+            let jsonString = try elem.html()
+            //print(json)
+            let json = try JSONSerialization.jsonObject(with: jsonString.data(using: .utf8)!, options: .fragmentsAllowed)
+            if let recipe = json as? [String : Any] {
+                return recipe
+            } else if let recipe = (json as! [Any])[0] as? [String : Any] {
+                return recipe
+            }            
+        } catch {
+            print("COULD NOT DECODE")
+        }
+    }
+    
+    private func getRecipe(fromDict recipe: Dictionary<String, Any>) {
+        if recipe["@type"] as? String ?? "" == "Recipe" {
+            print(recipe["name"] ?? "No name")
+            print(recipe["recipeIngredient"] ?? "No ingredients")
+            print(recipe["recipeInstruction"] ?? "No instruction")
+        } else if (recipe["@type"] as? [String] ?? []).contains("Recipe") {
+            print(recipe["name"] ?? "No name")
+        }
+    }
+    
+}
--- a/Client/RecipeImport/Scraper.playground/Contents.swift
+++ b/Client/RecipeImport/Scraper.playground/Contents.swift
@@ -0,0 +1,63 @@
+import SwiftSoup
+import Foundation
+
+
+
+//let url = "https://www.chefkoch.de/rezepte/1385981243676608/Knusprige-Entenbrust.html"
+let url = "https://www.allrecipes.com/recipe/234620/mascarpone-mashed-potatoes/"
+var contents: String? = nil
+if let url = URL(string: url) {
+    do {
+        contents = try String(contentsOf: url)
+        //print(contents)
+    } catch {
+        print("ERROR: Could not load url content.")
+    }
+    
+} else {
+    print("ERROR: Bad url.")
+}
+
+guard let html = contents else {
+    print("ERROR: no contents")
+    exit(1)
+}
+
+let doc: Document = try SwiftSoup.parse(html)
+let elements: Elements = try doc.select("script")
+for elem in elements.array() {
+    for attr in elem.getAttributes()!.asList() {
+        //print(attr.getValue())
+        if attr.getValue() == "application/ld+json" {
+            
+            do {
+                let jsonString = try elem.html()
+                //print(json)
+                let json = try JSONSerialization.jsonObject(with: jsonString.data(using: .utf8)!, options: .fragmentsAllowed)
+                if let recipe = json as? [String : Any] {
+                    print("1")
+                    getRecipe(fromDict: recipe)
+                } else if let recipe = (json as! [Any])[0] as? [String : Any] {
+                    print("2")
+                    getRecipe(fromDict: recipe)
+                }
+                
+                
+            } catch {
+                print("COULD NOT DECODE")
+            }
+        }
+    }
+}
+
+
+func getRecipe(fromDict recipe: Dictionary<String, Any>) {
+    
+    if recipe["@type"] as? String ?? "" == "Recipe" {
+        print(recipe["name"] ?? "No name")
+        print(recipe["recipeIngredient"] ?? "No ingredients")
+        print(recipe["recipeInstruction"] ?? "No instruction")
+    } else if (recipe["@type"] as? [String] ?? []).contains("Recipe") {
+        print(recipe["name"] ?? "No name")
+    }
+}
--- a/Client/RecipeImport/Scraper.playground/contents.xcplayground
+++ b/Client/RecipeImport/Scraper.playground/contents.xcplayground
@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<playground version='5.0' target-platform='ios' buildActiveScheme='true' importAppTypes='true'>
+    <timeline fileName='timeline.xctimeline'/>
+</playground>