mirror/QodeAssist
1
0
Fork 0
mirror of https://github.com/Palm1r/QodeAssist.git synced 2025-10-31 00:04:36 -04:00
Code Issues Packages Projects Releases Wiki Activity

Adapt new settings

Browse Source
This commit is contained in:
Petr Mironychev
2024-11-10 23:42:06 +01:00
parent 85d039cbd5
commit bc93bce03b
16 changed files with 264 additions and 159 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
providers/LMStudioProvider.cpp
View File

@ -26,7 +26,8 @@
#include <QNetworkReply>
#include "logger/Logger.hpp"
#include "settings/PresetPromptsSettings.hpp"
#include "settings/ChatAssistantSettings.hpp"
#include "settings/CodeCompletionSettings.hpp"
namespace QodeAssist::Providers {
@ -54,36 +55,43 @@ QString LMStudioProvider::chatEndpoint() const
void LMStudioProvider::prepareRequest(QJsonObject &request, LLMCore::RequestType type)
{
auto &promptSettings = Settings::presetPromptsSettings();
auto settings = promptSettings.getSettings(type);
auto prepareMessages = [](QJsonObject &req) -> QJsonArray {
QJsonArray messages;
if (req.contains("system")) {
messages.append(
QJsonObject{{"role", "system"}, {"content", req.take("system").toString()}});
}
if (req.contains("prompt")) {
messages.append(
QJsonObject{{"role", "user"}, {"content", req.take("prompt").toString()}});
}
return messages;
};
QJsonArray messages;
auto applyModelParams = [&request](const auto &settings) {
request["max_tokens"] = settings.maxTokens();
request["temperature"] = settings.temperature();
if (request.contains("system")) {
QJsonObject systemMessage{{"role", "system"},
{"content", request.take("system").toString()}};
messages.append(systemMessage);
}
if (request.contains("prompt")) {
QJsonObject userMessage{{"role", "user"}, {"content", request.take("prompt").toString()}};
messages.append(userMessage);
}
if (settings.useTopP())
request["top_p"] = settings.topP();
if (settings.useTopK())
request["top_k"] = settings.topK();
if (settings.useFrequencyPenalty())
request["frequency_penalty"] = settings.frequencyPenalty();
if (settings.usePresencePenalty())
request["presence_penalty"] = settings.presencePenalty();
};
QJsonArray messages = prepareMessages(request);
if (!messages.isEmpty()) {
request["messages"] = std::move(messages);
}
request["max_tokens"] = settings.maxTokens;
request["temperature"] = settings.temperature;
if (settings.useTopP)
request["top_p"] = settings.topP;
if (settings.useTopK)
request["top_k"] = settings.topK;
if (settings.useFrequencyPenalty)
request["frequency_penalty"] = settings.frequencyPenalty;
if (settings.usePresencePenalty)
request["presence_penalty"] = settings.presencePenalty;
if (type == LLMCore::RequestType::Fim) {
applyModelParams(Settings::codeCompletionSettings());
} else {
applyModelParams(Settings::chatAssistantSettings());
}
}
bool LMStudioProvider::handleResponse(QNetworkReply *reply, QString &accumulatedResponse)

40
providers/OllamaProvider.cpp
View File

@ -26,7 +26,8 @@
#include <QtCore/qeventloop.h>
#include "logger/Logger.hpp"
#include "settings/PresetPromptsSettings.hpp"
#include "settings/ChatAssistantSettings.hpp"
#include "settings/CodeCompletionSettings.hpp"
namespace QodeAssist::Providers {
@ -54,22 +55,29 @@ QString OllamaProvider::chatEndpoint() const
void OllamaProvider::prepareRequest(QJsonObject &request, LLMCore::RequestType type)
{
auto &promptSettings = Settings::presetPromptsSettings();
auto settings = promptSettings.getSettings(type);
auto applySettings = [&request](const auto &settings) {
QJsonObject options;
options["num_predict"] = settings.maxTokens();
options["temperature"] = settings.temperature();
QJsonObject options;
options["num_predict"] = settings.maxTokens;
options["temperature"] = settings.temperature;
if (settings.useTopP)
options["top_p"] = settings.topP;
if (settings.useTopK)
options["top_k"] = settings.topK;
if (settings.useFrequencyPenalty)
options["frequency_penalty"] = settings.frequencyPenalty;
if (settings.usePresencePenalty)
options["presence_penalty"] = settings.presencePenalty;
request["options"] = options;
request["keep_alive"] = settings.ollamaLivetime;
if (settings.useTopP())
options["top_p"] = settings.topP();
if (settings.useTopK())
options["top_k"] = settings.topK();
if (settings.useFrequencyPenalty())
options["frequency_penalty"] = settings.frequencyPenalty();
if (settings.usePresencePenalty())
options["presence_penalty"] = settings.presencePenalty();
request["options"] = options;
request["keep_alive"] = settings.ollamaLivetime();
};
if (type == LLMCore::RequestType::Fim) {
applySettings(Settings::codeCompletionSettings());
} else {
applySettings(Settings::chatAssistantSettings());
}
}
bool OllamaProvider::handleResponse(QNetworkReply *reply, QString &accumulatedResponse)

59
providers/OpenAICompatProvider.cpp
View File

@ -18,14 +18,14 @@
*/
#include "OpenAICompatProvider.hpp"
#include "settings/ChatAssistantSettings.hpp"
#include "settings/CodeCompletionSettings.hpp"
#include <QJsonArray>
#include <QJsonDocument>
#include <QJsonObject>
#include <QNetworkReply>
#include "settings/PresetPromptsSettings.hpp"
namespace QodeAssist::Providers {
OpenAICompatProvider::OpenAICompatProvider() {}
@ -52,39 +52,42 @@ QString OpenAICompatProvider::chatEndpoint() const
void OpenAICompatProvider::prepareRequest(QJsonObject &request, LLMCore::RequestType type)
{
auto &promptSettings = Settings::presetPromptsSettings();
auto settings = promptSettings.getSettings(type);
QJsonArray messages;
auto prepareMessages = [](QJsonObject &req) -> QJsonArray {
QJsonArray messages;
if (req.contains("system")) {
messages.append(
QJsonObject{{"role", "system"}, {"content", req.take("system").toString()}});
}
if (req.contains("prompt")) {
messages.append(
QJsonObject{{"role", "user"}, {"content", req.take("prompt").toString()}});
}
return messages;
};
if (request.contains("system")) {
QJsonObject systemMessage{{"role", "system"},
{"content", request.take("system").toString()}};
messages.append(systemMessage);
}
auto applyModelParams = [&request](const auto &settings) {
request["max_tokens"] = settings.maxTokens();
request["temperature"] = settings.temperature();
if (request.contains("prompt")) {
QJsonObject userMessage{{"role", "user"}, {"content", request.take("prompt").toString()}};
messages.append(userMessage);
}
if (settings.useTopP())
request["top_p"] = settings.topP();
if (settings.useTopK())
request["top_k"] = settings.topK();
if (settings.useFrequencyPenalty())
request["frequency_penalty"] = settings.frequencyPenalty();
if (settings.usePresencePenalty())
request["presence_penalty"] = settings.presencePenalty();
};
QJsonArray messages = prepareMessages(request);
if (!messages.isEmpty()) {
request["messages"] = std::move(messages);
}
request["max_tokens"] = settings.maxTokens;
request["temperature"] = settings.temperature;
if (settings.useTopP)
request["top_p"] = settings.topP;
if (settings.useTopK)
request["top_k"] = settings.topK;
if (settings.useFrequencyPenalty)
request["frequency_penalty"] = settings.frequencyPenalty;
if (settings.usePresencePenalty)
request["presence_penalty"] = settings.presencePenalty;
const QString &apiKey = settings.apiKey;
if (!apiKey.isEmpty()) {
request["api_key"] = apiKey;
if (type == LLMCore::RequestType::Fim) {
applyModelParams(Settings::codeCompletionSettings());
} else {
applyModelParams(Settings::chatAssistantSettings());
}
}