ad-hock word play (#20)

* words audio pre-processed

ConsoleApp/ConsoleApp.csproj

@@ -0,0 +1,22 @@
+<Project Sdk="Microsoft.NET.Sdk">
+
+  <PropertyGroup>
+    <OutputType>Exe</OutputType>
+    <TargetFramework>net8.0</TargetFramework>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <Nullable>enable</Nullable>
+    <UserSecretsId>716f7a77-339c-46aa-89a6-2dc47eab1d23</UserSecretsId>
+  </PropertyGroup>
+
+  <ItemGroup>
+    <ProjectReference Include="..\My1kWordsEe\My1kWordsEe.csproj" />
+  </ItemGroup>
+
+  <ItemGroup>
+    <PackageReference Include="Microsoft.Extensions.Configuration" Version="8.0.0" />
+    <PackageReference Include="Microsoft.Extensions.Configuration.UserSecrets" Version="8.0.0" />
+    <PackageReference Include="Microsoft.Extensions.Logging" Version="8.0.0" />
+    <PackageReference Include="Microsoft.Extensions.Logging.Console" Version="8.0.0" />
+  </ItemGroup>
+
+</Project>

ConsoleApp/Program.cs

@@ -0,0 +1,79 @@
+using Microsoft.Extensions.Configuration;
+using Microsoft.Extensions.Logging;
+
+using My1kWordsEe.Services;
+using My1kWordsEe.Services.Cqs;
+using My1kWordsEe.Services.Db;
+
+namespace ConsoleApp
+{
+    internal class Program
+    {
+
+        private static readonly string[] Words = new string[]
+        {
+        };
+
+        static async Task Main(string[] args)
+        {
+            using ILoggerFactory factory = LoggerFactory.Create(builder => builder.AddConsole());
+            IConfigurationRoot config = new ConfigurationBuilder()
+                .AddUserSecrets<Program>()
+                .Build();
+
+            var openAiKey = config["Secrets:OpenAiKey"];
+
+            if (string.IsNullOrWhiteSpace(openAiKey))
+            {
+                throw new ApplicationException("Secrets:OpenAiKey is missing");
+            }
+
+            var azureBlobConnectionString = config["Secrets:AzureBlobConnectionString"];
+
+            if (string.IsNullOrWhiteSpace(azureBlobConnectionString))
+            {
+                throw new ApplicationException("Secrets:AzureBlobConnectionString is missing");
+            }
+
+            var blob = new AzureBlobService(azureBlobConnectionString);
+            var openAi = new OpenAiService(factory.CreateLogger<OpenAiService>(), openAiKey);
+            var ensure = new EnsureWordCommand(blob, openAi);
+
+            var errors = new List<string>();
+
+            Console.WriteLine(Words.Length);
+            foreach (var word in Words)
+            {
+                Console.WriteLine($"Processing {word}");
+
+                var ensureCmd = await ensure.Invoke(word);
+                if (ensureCmd.IsFailure)
+                {
+                    Console.WriteLine($"Failed to ensure {word}: {ensureCmd.Error}");
+                    errors.Add($"Failed to ensure {word}: {ensureCmd.Error}");
+                }
+
+                var data = await blob.GetWordData(word);
+
+                if (data.IsSuccess)
+                {
+                    var wordData = data.Value;
+                    var update = wordData with
+                    {
+                        EeAudioUrl = new Uri(
+                            $"https://my1kee.blob.core.windows.net/audio/{wordData.EeWord}.wav")
+                    };
+
+                    await blob.SaveWordData(update);
+                }
+                else
+                {
+                    errors.Add($"Failed to get data for {word}");
+                    Console.WriteLine($"Failed to get data for {word}");
+                }
+            }
+
+            File.WriteAllLines("cmd-errors-5.txt", errors);
+        }
+    }
+}

My1kWordsEe/Components/Pages/Word.razor

@@ -33,11 +33,33 @@
         }
         this.isGenerationInProgress = false;
     }
+
+    private async Task SpeakWord(MouseEventArgs e)
+    {
+        await JS.InvokeVoidAsync("speakWord", EeWord);
+    }
+
+    private bool IsDataLoadedOk => WordMetadata.HasValue && WordMetadata.Value.IsSuccess;
 }
 
+<script>
+    window.speakWord = async (audioId) => {
+        const audio = document.getElementById(audioId);
+        await audio.play();
+    };
+</script>
+
 <div class="text-left">
     <div class="row">
-        <h1>@EeWord
+        <h1>
+            @EeWord
+
+            @if (IsDataLoadedOk)
+            {
+                <i class="bi bi-volume-down-fill small" @onclick="SpeakWord">
+                    <audio id="@EeWord" src="@WordMetadata.Value.Value.EeAudioUrl"></audio>
+                </i>
+            }
         </h1>
 
         @if (WordMetadata.HasValue)

My1kWordsEe/My1kWordsEe.sln

@@ -3,7 +3,9 @@ Microsoft Visual Studio Solution File, Format Version 12.00
 # Visual Studio Version 17
 VisualStudioVersion = 17.10.35122.118
 MinimumVisualStudioVersion = 10.0.40219.1
-Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "My1kWordsEe", "My1kWordsEe.csproj", "{497B021C-F03E-4CB6-BC42-3A9D8BC68F68}"
+Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "My1kWordsEe", "My1kWordsEe.csproj", "{497B021C-F03E-4CB6-BC42-3A9D8BC68F68}"
+EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "ConsoleApp", "..\ConsoleApp\ConsoleApp.csproj", "{ECE5CED5-98E9-4FAB-A3CB-0A5002C88D9C}"
 EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
@@ -15,6 +17,10 @@ Global
 		{497B021C-F03E-4CB6-BC42-3A9D8BC68F68}.Debug|Any CPU.Build.0 = Debug|Any CPU
 		{497B021C-F03E-4CB6-BC42-3A9D8BC68F68}.Release|Any CPU.ActiveCfg = Release|Any CPU
 		{497B021C-F03E-4CB6-BC42-3A9D8BC68F68}.Release|Any CPU.Build.0 = Release|Any CPU
+		{ECE5CED5-98E9-4FAB-A3CB-0A5002C88D9C}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{ECE5CED5-98E9-4FAB-A3CB-0A5002C88D9C}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{ECE5CED5-98E9-4FAB-A3CB-0A5002C88D9C}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{ECE5CED5-98E9-4FAB-A3CB-0A5002C88D9C}.Release|Any CPU.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE

My1kWordsEe/Program.cs

@@ -9,14 +9,15 @@ public class Program
     {
         public static void Main(string[] args)
         {
+            // default log: Console, Debug, EventSource
             var builder = WebApplication.CreateBuilder(args);
             builder.Configuration.AddEnvironmentVariables();
 
-            var openApiKey =
+            var openAiKey =
                 builder.Configuration["Secrets:OpenAiKey"] ??
                 Environment.GetEnvironmentVariable("Secrets_OpenAiKey");
 
-            if (string.IsNullOrWhiteSpace(openApiKey))
+            if (string.IsNullOrWhiteSpace(openAiKey))
             {
                 throw new ApplicationException("Secrets:OpenAiKey is missing");
             }
@@ -40,7 +41,7 @@ public static void Main(string[] args)
             }
 
             builder.Services.AddSingleton(new StabilityAiService(stabilityAiKey));
-            builder.Services.AddSingleton(new OpenAiService(openApiKey));
+            builder.Services.AddSingleton((p) => new OpenAiService(p.GetRequiredService<ILogger<OpenAiService>>(), openAiKey));
             builder.Services.AddSingleton(new AzureBlobService(azureBlobConnectionString));
             builder.Services.AddSingleton(new TartuNlpService());
             builder.Services.AddSingleton<EnsureWordCommand>();

My1kWordsEe/Services/Cqs/EnsureWordCommand.cs

@@ -31,6 +31,7 @@ public async Task<Result<SampleWord>> Invoke(string eeWord)
 
             if (sampleWord.IsSuccess)
             {
+                // generate audio
                 await azureBlobService.SaveWordData(sampleWord.Value);
             }
 

My1kWordsEe/Services/Db/AzureBlobService.cs

@@ -52,14 +52,16 @@ await blob.UploadAsync(
                 overwrite: true);
         }
 
-        public async Task<Uri> SaveAudio(Stream audioStream)
+        public async Task<Uri> SaveAudio(Stream audioStream, string blobName)
         {
             BlobContainerClient container = await GetAudioContainer();
-            BlobClient blob = container.GetBlobClient(WavBlobName());
-            await blob.UploadAsync(audioStream);
+            BlobClient blob = container.GetBlobClient(blobName);
+            await blob.UploadAsync(audioStream, overwrite: true);
             return blob.Uri;
         }
 
+        public Task<Uri> SaveAudio(Stream audioStream) => SaveAudio(audioStream, WavBlobName());
+
         public async Task<Uri> SaveImage(Stream imageStream)
         {
             BlobContainerClient container = await GetImageContainer();

My1kWordsEe/Services/OpenAiService.cs

@@ -12,9 +12,12 @@ namespace My1kWordsEe.Services
 {
     public class OpenAiService
     {
-        public OpenAiService(string apiKey)
+        private readonly ILogger logger;
+
+        public OpenAiService(ILogger<OpenAiService> logger, string apiKey)
         {
-            ApiKey = apiKey;
+            this.logger = logger;
+            this.ApiKey = apiKey;
         }
 
         private string ApiKey { get; }
@@ -93,21 +96,20 @@ public async Task<Result<Sentence>> GetSampleSentence(string eeWord)
             ChatCompletion chatCompletion = await client.CompleteChatAsync(
                 [
                     new SystemChatMessage(
-                        "Sa oled keeleõppe süsteemi abiline, mis aitab õppida enim levinud eesti keele sõnu. " +
-                        "Sinu sisend on üks sõna eesti keeles. " +
-                        "Sinu ülesanne on kirjutada selle kasutamise kohta lihtne lühike näitelause, kasutades seda sõna. " +
-                        "Lauses kasuta kõige levinuimaid ja lihtsamaid sõnu eesti keeles et toetada keeleõpet. " +
-                        "Sinu väljund on JSON objekt, milles on näitelaus eesti keeles ja selle vastav tõlge inglise keelde:\n" +
-                        "```\n" +
-                        "{\"ee_sentence\": \"<näide eesti keeles>\", \"en_sentence\": \"<näide inglise keeles>\" }" +
-                        "\n```" +
-                        "\n Tagastab ainult json-objekti!"),
+                        "Sa oled keeleõppe süsteemi abiline, mis aitab õppida enim levinud eesti keele sõnu.\n" +
+                        "Sinu sisend on üks sõna eesti keeles.\n" +
+                        "Sinu ülesanne on kirjutada selle kasutamise kohta lihtne lühike näitelause, kasutades seda sõna.\n" +
+                        "Lauses kasuta kõige levinuimaid ja lihtsamaid sõnu eesti keeles et toetada keeleõpet.\n" +
+                        "Teie väljundiks on JSON-objekt koos eestikeelse näidislausega ja sellele vastav tõlge inglise keelde vastavalt lepingule:\n" +
+                        "```\n{\n" +
+                        "\"ee_sentence\": \"<näide eesti keeles>\", \"en_sentence\": \"<näide inglise keeles>\"" +
+                        "\n}\n```"),
                     new UserChatMessage(eeWord),
                 ]);
 
             foreach (var c in chatCompletion.Content)
             {
-                var jsonStr = c.Text.Trim('`', ' ', '\'', '"');
+                var jsonStr = c.Text.Replace("json", "", StringComparison.OrdinalIgnoreCase).Trim('`', ' ', '\'', '"');
                 var sentence = JsonSerializer.Deserialize<Sentence>(jsonStr);
                 if (sentence == null)
                 {
@@ -124,26 +126,26 @@ public async Task<Result<Sentence>> GetSampleSentence(string eeWord)
 
         public async Task<Result<SampleWord>> GetWordMetadata(string word)
         {
-            ChatClient client = new(model: "gpt-4o-mini", ApiKey);
+            ChatClient client = new(model: "gpt-4o", ApiKey);
 
             ChatCompletion chatCompletion = await client.CompleteChatAsync(
                 [
                     new SystemChatMessage(
-                        "Your input is an Estonian word. " +
-                        "Your output is word metadata in JSON:\n" +
+                        "Sinu sisend on eestikeelne sõna.\n" +
+                        "Kui antud sõna ei ole eestikeelne, tagasta 404\n"+
+                        "Teie väljund on sõna metaandmed JSON-is vastavalt antud lepingule:\n" +
                         "```\n{\n" +
-                        "ee_word: \"<given word>\",\n" +
-                        "en_word: \"<english translation>\"\n" +
-                        "en_words: [<alternative translations if applicable>]\n" +
-                        "en_explanation: \"<explanation of the word in english>\"\n" +
-                        "}\n```\n" +
-                        "If the given word is not Estonian return 404"),
+                        "\"ee_word\": \"<antud sõna>\",\n" +
+                        "\"en_word\": \"<english translation>\"\n" +
+                        "\"en_words\": [<array of alternative english translations if applicable>]\n" +
+                        "\"en_explanation\": \"<explanation of the word meaning in english>\"\n" +
+                        "}\n```\n"),
                     new UserChatMessage(word),
                 ]);
 
             foreach (var c in chatCompletion.Content)
             {
-                var jsonStr = c.Text.Trim('`', ' ', '\'', '"');
+                var jsonStr = c.Text.Replace("json", "", StringComparison.OrdinalIgnoreCase).Trim('`', ' ', '\'', '"');
 
                 if (jsonStr.Contains("404"))
                 {
@@ -168,8 +170,9 @@ public async Task<Result<SampleWord>> GetWordMetadata(string word)
                         });
                     }
                 }
-                catch (JsonException)
+                catch (JsonException jsonException)
                 {
+                    this.logger.LogError(jsonException, "Failed to deserialize JSON: {jsonStr}", jsonStr);
                     return Result.Failure<SampleWord>("Unexpected data returned by AI");
                 }
             }