Create an ILMProvider interface and have our current implementation use it

.github/actions/spelling/allow/allow.txt

@@ -28,11 +28,14 @@ godbolt
 gpt
 hyperlinking
 hyperlinks
+ILM
 kje
 libfuzzer
 liga
 lje
 Llast
+lm
+llm
 Lmid
 locl
 lol

src/cascadia/QueryExtension/AzureLLMProvider.cpp

@@ -0,0 +1,228 @@
+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT license.
+
+#include "pch.h"
+#include "AzureLLMProvider.h"
+#include "../../types/inc/utils.hpp"
+#include "LibraryResources.h"
+
+#include "AzureLLMProvider.g.cpp"
+#include "AzureResponse.g.cpp"
+
+using namespace winrt::Windows::Foundation;
+using namespace winrt::Windows::Foundation::Collections;
+using namespace winrt::Windows::UI::Core;
+using namespace winrt::Windows::UI::Xaml;
+using namespace winrt::Windows::UI::Xaml::Controls;
+using namespace winrt::Windows::System;
+namespace WWH = ::winrt::Windows::Web::Http;
+namespace WSS = ::winrt::Windows::Storage::Streams;
+namespace WDJ = ::winrt::Windows::Data::Json;
+
+static constexpr std::wstring_view acceptedModels[] = {
+    L"gpt-35-turbo",
+    L"gpt4",
+    L"gpt4-32k",
+    L"gpt4o",
+    L"gpt-35-turbo-16k"
+};
+static constexpr std::wstring_view acceptedSeverityLevel{ L"safe" };
+static constexpr std::wstring_view expectedDomain{ L"azure.com" };
+static constexpr std::wstring_view applicationJson{ L"application/json" };
+static constexpr std::wstring_view endpointString{ L"endpoint" };
+static constexpr std::wstring_view keyString{ L"key" };
+static constexpr std::wstring_view roleString{ L"role" };
+static constexpr std::wstring_view contentString{ L"content" };
+static constexpr std::wstring_view messageString{ L"message" };
+static constexpr std::wstring_view errorString{ L"error" };
+static constexpr std::wstring_view severityString{ L"severity" };
+
+const std::wregex azureOpenAIEndpointRegex{ LR"(^https.*openai\.azure\.com)" };
+
+namespace winrt::Microsoft::Terminal::Query::Extension::implementation
+{
+    void AzureLLMProvider::SetAuthentication(const Windows::Foundation::Collections::ValueSet& authValues)
+    {
+        _azureEndpoint = unbox_value_or<hstring>(authValues.TryLookup(endpointString).try_as<IPropertyValue>(), L"");
+        _azureKey = unbox_value_or<hstring>(authValues.TryLookup(keyString).try_as<IPropertyValue>(), L"");
+        _httpClient = winrt::Windows::Web::Http::HttpClient{};
+        _httpClient.DefaultRequestHeaders().Accept().TryParseAdd(applicationJson);
+        _httpClient.DefaultRequestHeaders().Append(L"api-key", _azureKey);
+    }
+
+    void AzureLLMProvider::ClearMessageHistory()
+    {
+        _jsonMessages.Clear();
+    }
+
+    void AzureLLMProvider::SetSystemPrompt(const winrt::hstring& systemPrompt)
+    {
+        WDJ::JsonObject systemMessageObject;
+        winrt::hstring systemMessageContent{ systemPrompt };
+        systemMessageObject.Insert(roleString, WDJ::JsonValue::CreateStringValue(L"system"));
+        systemMessageObject.Insert(contentString, WDJ::JsonValue::CreateStringValue(systemMessageContent));
+        _jsonMessages.Append(systemMessageObject);
+    }
+
+    void AzureLLMProvider::SetContext(const Extension::IContext context)
+    {
+        _context = context;
+    }
+
+    winrt::Windows::Foundation::IAsyncOperation<Extension::IResponse> AzureLLMProvider::GetResponseAsync(const winrt::hstring& userPrompt)
+    {
+        // Use a flag for whether the response the user receives is an error message
+        // we pass this flag back to the caller so they can handle it appropriately (specifically, ExtensionPalette will send the correct telemetry event)
+        // there is only one case downstream from here that sets this flag to false, so start with it being true
+        bool isError{ true };
+        hstring message{};
+
+        if (_azureEndpoint.empty())
+        {
+            message = RS_(L"CouldNotFindKeyErrorMessage");
+        }
+        else
+        {
+            // If the AI endpoint is not an azure open AI endpoint, return an error message
+            Windows::Foundation::Uri parsedUri{ _azureEndpoint };
+            if (!std::regex_search(_azureEndpoint.c_str(), azureOpenAIEndpointRegex) ||
+                parsedUri.Domain() != expectedDomain)
+            {
+                message = RS_(L"InvalidEndpointMessage");
+            }
+        }
+
+        // If we don't have a message string, that means the endpoint exists and matches the regex
+        // that we allow - now we can actually make the http request
+        if (message.empty())
+        {
+            // Make a copy of the prompt because we are switching threads
+            const auto promptCopy{ userPrompt };
+
+            // Make sure we are on the background thread for the http request
+            co_await winrt::resume_background();
+
+            WWH::HttpRequestMessage request{ WWH::HttpMethod::Post(), Uri{ _azureEndpoint } };
+            request.Headers().Accept().TryParseAdd(applicationJson);
+
+            WDJ::JsonObject jsonContent;
+            WDJ::JsonObject messageObject;
+
+            // _ActiveCommandline should be set already, we request for it the moment we become visible
+            winrt::hstring engineeredPrompt{ promptCopy };
+            if (_context && !_context.ActiveCommandline().empty())
+            {
+                engineeredPrompt = promptCopy + L". The shell I am running is " + _context.ActiveCommandline();
+            }
+            messageObject.Insert(roleString, WDJ::JsonValue::CreateStringValue(L"user"));
+            messageObject.Insert(contentString, WDJ::JsonValue::CreateStringValue(engineeredPrompt));
+            _jsonMessages.Append(messageObject);
+            jsonContent.SetNamedValue(L"messages", _jsonMessages);
+            jsonContent.SetNamedValue(L"max_tokens", WDJ::JsonValue::CreateNumberValue(800));
+            jsonContent.SetNamedValue(L"temperature", WDJ::JsonValue::CreateNumberValue(0.7));
+            jsonContent.SetNamedValue(L"frequency_penalty", WDJ::JsonValue::CreateNumberValue(0));
+            jsonContent.SetNamedValue(L"presence_penalty", WDJ::JsonValue::CreateNumberValue(0));
+            jsonContent.SetNamedValue(L"top_p", WDJ::JsonValue::CreateNumberValue(0.95));
+            jsonContent.SetNamedValue(L"stop", WDJ::JsonValue::CreateStringValue(L"None"));
+            const auto stringContent = jsonContent.ToString();
+            WWH::HttpStringContent requestContent{
+                stringContent,
+                WSS::UnicodeEncoding::Utf8,
+                L"application/json"
+            };
+
+            request.Content(requestContent);
+
+            // Send the request
+            try
+            {
+                const auto response = _httpClient.SendRequestAsync(request).get();
+                // Parse out the suggestion from the response
+                const auto string{ response.Content().ReadAsStringAsync().get() };
+                const auto jsonResult{ WDJ::JsonObject::Parse(string) };
+                if (jsonResult.HasKey(errorString))
+                {
+                    const auto errorObject = jsonResult.GetNamedObject(errorString);
+                    message = errorObject.GetNamedString(messageString);
+                }
+                else
+                {
+                    if (_verifyModelIsValidHelper(jsonResult))
+                    {
+                        const auto choices = jsonResult.GetNamedArray(L"choices");
+                        const auto firstChoice = choices.GetAt(0).GetObject();
+                        const auto messageObject = firstChoice.GetNamedObject(messageString);
+                        message = messageObject.GetNamedString(contentString);
+                        isError = false;
+                    }
+                    else
+                    {
+                        message = RS_(L"InvalidModelMessage");
+                    }
+                }
+            }
+            catch (...)
+            {
+                message = RS_(L"UnknownErrorMessage");
+            }
+        }
+
+        // Also make a new entry in our jsonMessages list, so the AI knows the full conversation so far
+        WDJ::JsonObject responseMessageObject;
+        responseMessageObject.Insert(roleString, WDJ::JsonValue::CreateStringValue(L"assistant"));
+        responseMessageObject.Insert(contentString, WDJ::JsonValue::CreateStringValue(message));
+        _jsonMessages.Append(responseMessageObject);
+
+        co_return winrt::make<AzureResponse>(message, isError);
+    }
+
+    bool AzureLLMProvider::_verifyModelIsValidHelper(const WDJ::JsonObject jsonResponse)
+    {
+        const auto model = jsonResponse.GetNamedString(L"model");
+        bool modelIsAccepted{ false };
+        for (const auto acceptedModel : acceptedModels)
+        {
+            if (model == acceptedModel)
+            {
+                modelIsAccepted = true;
+            }
+            break;
+        }
+        if (!modelIsAccepted)
+        {
+            return false;
+        }
+        WDJ::JsonObject contentFiltersObject;
+        // For some reason, sometimes the content filter results are in a key called "prompt_filter_results"
+        // and sometimes they are in a key called "prompt_annotations". Check for either.
+        if (jsonResponse.HasKey(L"prompt_filter_results"))
+        {
+            contentFiltersObject = jsonResponse.GetNamedArray(L"prompt_filter_results").GetObjectAt(0);
+        }
+        else if (jsonResponse.HasKey(L"prompt_annotations"))
+        {
+            contentFiltersObject = jsonResponse.GetNamedArray(L"prompt_annotations").GetObjectAt(0);
+        }
+        else
+        {
+            return false;
+        }
+        const auto contentFilters = contentFiltersObject.GetNamedObject(L"content_filter_results");
+        if (Feature_TerminalChatJailbreakFilter::IsEnabled() && !contentFilters.HasKey(L"jailbreak"))
+        {
+            return false;
+        }
+        for (const auto filterPair : contentFilters)
+        {
+            const auto filterLevel = filterPair.Value().GetObjectW();
+            if (filterLevel.HasKey(severityString))
+            {
+                if (filterLevel.GetNamedString(severityString) != acceptedSeverityLevel)
+                {
+                    return false;
+                }
+            }
+        }
+        return true;
+    }
+}

src/cascadia/QueryExtension/AzureLLMProvider.h

@@ -0,0 +1,54 @@
+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT license.
+
+#pragma once
+
+#include "AzureLLMProvider.g.h"
+#include "AzureResponse.g.h"
+
+namespace winrt::Microsoft::Terminal::Query::Extension::implementation
+{
+    struct AzureLLMProvider : AzureLLMProviderT<AzureLLMProvider>
+    {
+        AzureLLMProvider() = default;
+
+        void ClearMessageHistory();
+        void SetSystemPrompt(const winrt::hstring& systemPrompt);
+        void SetContext(const Extension::IContext context);
+
+        winrt::Windows::Foundation::IAsyncOperation<Extension::IResponse> GetResponseAsync(const winrt::hstring& userPrompt);
+
+        void SetAuthentication(const Windows::Foundation::Collections::ValueSet& authValues);
+        TYPED_EVENT(AuthChanged, winrt::Microsoft::Terminal::Query::Extension::ILMProvider, winrt::hstring);
+
+    private:
+        winrt::hstring _azureEndpoint;
+        winrt::hstring _azureKey;
+        winrt::Windows::Web::Http::HttpClient _httpClient{ nullptr };
+
+        Extension::IContext _context;
+
+        winrt::Windows::Data::Json::JsonArray _jsonMessages;
+
+        bool _verifyModelIsValidHelper(const Windows::Data::Json::JsonObject jsonResponse);
+    };
+
+    struct AzureResponse : AzureResponseT<AzureResponse>
+    {
+        AzureResponse(const winrt::hstring& message, const bool isError) :
+            _message{ message },
+            _isError{ isError } {}
+        winrt::hstring Message() { return _message; };
+        bool IsError() { return _isError; };
+
+    private:
+        winrt::hstring _message;
+        bool _isError;
+    };
+}
+
+namespace winrt::Microsoft::Terminal::Query::Extension::factory_implementation
+{
+    BASIC_FACTORY(AzureLLMProvider);
+    BASIC_FACTORY(AzureResponse);
+}

src/cascadia/QueryExtension/AzureLLMProvider.idl

@@ -0,0 +1,17 @@
+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT license.
+
+import "ILMProvider.idl";
+
+namespace Microsoft.Terminal.Query.Extension
+{
+    [default_interface] runtimeclass AzureLLMProvider : ILMProvider
+    {
+        AzureLLMProvider();
+    }
+
+    [default_interface] runtimeclass AzureResponse : IResponse
+    {
+        AzureResponse(String message, Boolean isError);
+    }
+}