Added a preview of the document analysis assistant (#561)

Co-authored-by: Peer Schütt <20603780+peerschuett@users.noreply.github.com>

app/MindWork AI Studio/Settings/DataModel/Data.cs

@@ -84,6 +84,8 @@ public sealed class Data
     public DataCoding Coding { get; init; } = new();
     
     public DataERI ERI { get; init; } = new();
+    
+    public DataDocumentAnalysis DocumentAnalysis { get; init; } = new();
 
     public DataTextSummarizer TextSummarizer { get; init; } = new();
 

app/MindWork AI Studio/Settings/DataModel/DataDocumentAnalysis.cs

@@ -0,0 +1,9 @@
+namespace AIStudio.Settings.DataModel;
+
+public sealed class DataDocumentAnalysis
+{
+    /// <summary>
+    /// Configured document analysis policies.
+    /// </summary>
+    public List<DataDocumentAnalysisPolicy> Policies { get; set; } = [];
+}

app/MindWork AI Studio/Settings/DataModel/DataDocumentAnalysisPolicy.cs

@@ -0,0 +1,52 @@
+using AIStudio.Provider;
+
+namespace AIStudio.Settings.DataModel;
+
+public sealed class DataDocumentAnalysisPolicy
+{
+    /// <summary>
+    /// Preselect the policy name?
+    /// </summary>
+    public string PolicyName { get; set; } = string.Empty;
+    
+    /// <summary>
+    /// Preselect the policy description?
+    /// </summary>
+    public string PolicyDescription { get; set; } = string.Empty;
+
+    /// <summary>
+    /// Is this policy protected? If so, it cannot be deleted or modified by the user.
+    /// </summary>
+    public bool IsProtected { get; set; }
+
+    /// <summary>
+    /// Is this a managed policy? Managed policies are created and managed by the organization
+    /// and cannot be modified or deleted by the user.
+    /// </summary>
+    public bool IsManaged { get; set; }
+
+    /// <summary>
+    /// The rules for the document analysis policy.
+    /// </summary>
+    public string AnalysisRules { get; set; } = string.Empty;
+
+    /// <summary>
+    /// The rules for the output of the document analysis, e.g., the desired format, structure, etc.
+    /// </summary>
+    public string OutputRules { get; set; } = string.Empty;
+    
+    /// <summary>
+    /// The minimum confidence level required for a provider to be considered.
+    /// </summary>
+    public ConfidenceLevel MinimumProviderConfidence { get; set; } = ConfidenceLevel.NONE;
+    
+    /// <summary>
+    /// Which LLM provider should be preselected?
+    /// </summary>
+    public string PreselectedProvider { get; set; } = string.Empty;
+    
+    /// <summary>
+    /// Preselect a profile?
+    /// </summary>
+    public string PreselectedProfile { get; set; } = string.Empty;
+}

app/MindWork AI Studio/Settings/DataModel/PreviewFeatures.cs

@@ -11,4 +11,5 @@ public enum PreviewFeatures
     
     PRE_PLUGINS_2025,
     PRE_READ_PDF_2025,
+    PRE_DOCUMENT_ANALYSIS_2025,
 }

app/MindWork AI Studio/Settings/DataModel/PreviewFeaturesExtensions.cs

@@ -13,6 +13,7 @@ public static class PreviewFeaturesExtensions
         
         PreviewFeatures.PRE_PLUGINS_2025 => TB("Plugins: Preview of our plugin system where you can extend the functionality of the app"),
         PreviewFeatures.PRE_READ_PDF_2025 => TB("Read PDF: Preview of our PDF reading system where you can read and extract text from PDF files"),
+        PreviewFeatures.PRE_DOCUMENT_ANALYSIS_2025 => TB("Document Analysis: Preview of our document analysis system where you can analyze and extract information from documents"),
         
         _ => TB("Unknown preview feature")
     };

app/MindWork AI Studio/Settings/DataModel/PreviewVisibilityExtensions.cs

@@ -20,6 +20,7 @@ public static class PreviewVisibilityExtensions
         if (visibility >= PreviewVisibility.PROTOTYPE)
         {
             features.Add(PreviewFeatures.PRE_RAG_2024);
+            features.Add(PreviewFeatures.PRE_DOCUMENT_ANALYSIS_2025);
         }
         
         if (visibility >= PreviewVisibility.EXPERIMENTAL)

app/MindWork AI Studio/Settings/TolerantEnumConverter.cs

@@ -1,3 +1,4 @@
+using System.Text;
 using System.Text.Json;
 using System.Text.Json.Serialization;
 
@@ -9,6 +10,9 @@ namespace AIStudio.Settings;
 /// <remarks>
 /// When the target enum value does not exist, the value will be the default value.
 /// This converter handles enum values as property names and values.
+/// <br/><br/>
+/// We assume that enum names are in UPPER_SNAKE_CASE, and the JSON strings may be
+/// in any case style (e.g., camelCase, PascalCase, snake_case, UPPER_SNAKE_CASE, etc.)
 /// </remarks>
 public sealed class TolerantEnumConverter : JsonConverter<object>
 {
@@ -16,30 +20,46 @@ public sealed class TolerantEnumConverter : JsonConverter<object>
     
     public override bool CanConvert(Type typeToConvert) => typeToConvert.IsEnum;
 
-    public override object? Read(ref Utf8JsonReader reader, Type typeToConvert, JsonSerializerOptions options)
+    public override object? Read(ref Utf8JsonReader reader, Type enumType, JsonSerializerOptions options)
     {
         // Is this token a string?
         if (reader.TokenType == JsonTokenType.String)
+        {
             // Try to use that string as the name of the enum value:
-            if (Enum.TryParse(typeToConvert, reader.GetString(), out var result))
+            var text = reader.GetString();
+            
+            // Convert the text to UPPER_SNAKE_CASE:
+            text = ConvertToUpperSnakeCase(text);
+            
+            // Try to parse the enum value:
+            if (Enum.TryParse(enumType, text, out var result))
                 return result;
+        }
 
         // In any other case, we will return the default enum value:
-        LOG.LogWarning($"Cannot read '{reader.GetString()}' as '{typeToConvert.Name}' enum; token type: {reader.TokenType}");
-        return Activator.CreateInstance(typeToConvert);
+        LOG.LogWarning($"Cannot read '{reader.GetString()}' as '{enumType.Name}' enum; token type: {reader.TokenType}");
+        return Activator.CreateInstance(enumType);
     }
 
-    public override object ReadAsPropertyName(ref Utf8JsonReader reader, Type typeToConvert, JsonSerializerOptions options)
+    public override object ReadAsPropertyName(ref Utf8JsonReader reader, Type enumType, JsonSerializerOptions options)
     {
         // Is this token a property name?
         if (reader.TokenType == JsonTokenType.PropertyName)
+        {
             // Try to use that property name as the name of the enum value:
-            if (Enum.TryParse(typeToConvert, reader.GetString(), out var result))
+            var text = reader.GetString();
+            
+            // Convert the text to UPPER_SNAKE_CASE:
+            text = ConvertToUpperSnakeCase(text);
+            
+            // Try to parse the enum value:
+            if (Enum.TryParse(enumType, text, out var result))
                 return result;
+        }
 
         // In any other case, we will return the default enum value:
-        LOG.LogWarning($"Cannot read '{reader.GetString()}' as '{typeToConvert.Name}' enum; token type: {reader.TokenType}");
-        return Activator.CreateInstance(typeToConvert)!;
+        LOG.LogWarning($"Cannot read '{reader.GetString()}' as '{enumType.Name}' enum; token type: {reader.TokenType}");
+        return Activator.CreateInstance(enumType)!;
     }
 
     public override void Write(Utf8JsonWriter writer, object value, JsonSerializerOptions options)
@@ -51,4 +71,44 @@ public sealed class TolerantEnumConverter : JsonConverter<object>
     {
         writer.WritePropertyName(value.ToString()!);
     }
+    
+    /// <summary>
+    /// Converts a string to UPPER_SNAKE_CASE.
+    /// </summary>
+    /// <param name="text">The text to convert.</param>
+    /// <returns>The converted text as UPPER_SNAKE_CASE.</returns>
+    private static string ConvertToUpperSnakeCase(string? text)
+    {
+        // Handle null or empty strings:
+        if (string.IsNullOrWhiteSpace(text))
+            return string.Empty;
+        
+        // Create a string builder with the same length as the
+        // input text. We will add underscores as needed, which
+        // may increase the length -- we cannot predict how many
+        // underscores will be added, so we just start with the
+        // original length:
+        var sb = new StringBuilder(text.Length);
+        
+        // State to track if the last character was lowercase.
+        // This helps to determine when to add underscores:
+        var lastCharWasLowerCase = false;
+        
+        // Iterate through each character in the input text:
+        foreach(var c in text)
+        {
+            // If the current character is uppercase and the last
+            // character was lowercase, we need to add an underscore:
+            if (char.IsUpper(c) && lastCharWasLowerCase)
+                sb.Append('_');
+			
+            // Append the uppercase version of the current character:
+            sb.Append(char.ToUpperInvariant(c));
+            
+            // Keep track of whether the current character is lowercase:
+            lastCharWasLowerCase = char.IsLower(c);
+        }
+	
+        return sb.ToString();
+    }
 }