fix: use relative path for model information (#1399)

* feat: relative path for data (part1) * chore: unit tests * fix: unit tests * fix: more * fix: windows * fix: comments * fix: use fs path * chore: move logic to file manager utils (#1405) --------- Co-authored-by: vansangpfiev <[email protected]> Co-authored-by: nguyenhoangthuan99 <[email protected]>
janhq · Oct 3, 2024 · 8657549 · 8657549
1 parent 65b1d20
commit 8657549
Show file tree

Hide file tree

Showing 13 changed files with 234 additions and 56 deletions.
diff --git a/engine/commands/chat_completion_cmd.cc b/engine/commands/chat_completion_cmd.cc
@@ -1,14 +1,14 @@
 #include "chat_completion_cmd.h"
 #include "httplib.h"
 
+#include "config/yaml_config.h"
 #include "cortex_upd_cmd.h"
 #include "database/models.h"
 #include "model_status_cmd.h"
 #include "run_cmd.h"
 #include "server_start_cmd.h"
 #include "trantor/utils/Logger.h"
 #include "utils/logging_utils.h"
-#include "config/yaml_config.h"
 
 namespace commands {
 namespace {
@@ -41,6 +41,8 @@ struct ChunkParser {
 
 void ChatCompletionCmd::Exec(const std::string& host, int port,
                              const std::string& model_handle, std::string msg) {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   cortex::db::Models modellist_handler;
   config::YamlHandler yaml_handler;
   try {
@@ -49,7 +51,10 @@ void ChatCompletionCmd::Exec(const std::string& host, int port,
       CLI_LOG("Error: " + model_entry.error());
       return;
     }
-    yaml_handler.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    yaml_handler.ModelConfigFromFile(
+        fmu::ToAbsoluteCortexDataPath(
+            fs::path(model_entry.value().path_to_model_yaml))
+            .string());
     auto mc = yaml_handler.GetModelConfig();
     Exec(host, port, model_handle, mc, std::move(msg));
   } catch (const std::exception& e) {

diff --git a/engine/commands/model_get_cmd.cc b/engine/commands/model_get_cmd.cc
@@ -11,6 +11,8 @@
 namespace commands {
 
 void ModelGetCmd::Exec(const std::string& model_handle) {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   cortex::db::Models modellist_handler;
   config::YamlHandler yaml_handler;
   try {
@@ -19,7 +21,10 @@ void ModelGetCmd::Exec(const std::string& model_handle) {
       CLI_LOG("Error: " + model_entry.error());
       return;
     }
-    yaml_handler.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    yaml_handler.ModelConfigFromFile(
+        fmu::ToAbsoluteCortexDataPath(
+            fs::path(model_entry.value().path_to_model_yaml))
+            .string());
     auto model_config = yaml_handler.GetModelConfig();
 
     std::cout << model_config.ToString() << std::endl;

diff --git a/engine/commands/model_import_cmd.cc b/engine/commands/model_import_cmd.cc
@@ -14,6 +14,8 @@ ModelImportCmd::ModelImportCmd(std::string model_handle, std::string model_path)
       model_path_(std::move(model_path)) {}
 
 void ModelImportCmd::Exec() {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   config::GGUFHandler gguf_handler;
   config::YamlHandler yaml_handler;
   cortex::db::Models modellist_utils_obj;
@@ -22,10 +24,13 @@ void ModelImportCmd::Exec() {
                                  std::filesystem::path("imported") /
                                  std::filesystem::path(model_handle_ + ".yml"))
                                     .string();
-  cortex::db::ModelEntry model_entry{
-      model_handle_,   "local",       "imported",
-      model_yaml_path, model_handle_};
   try {
+    // Use relative path for model_yaml_path. In case of import, we use absolute path for model
+    auto yaml_rel_path =
+        fmu::ToRelativeCortexDataPath(fs::path(model_yaml_path));
+    cortex::db::ModelEntry model_entry{model_handle_, "local", "imported",
+                                       yaml_rel_path.string(), model_handle_};
+
     std::filesystem::create_directories(
         std::filesystem::path(model_yaml_path).parent_path());
     gguf_handler.Parse(model_path_);

diff --git a/engine/commands/model_list_cmd.cc b/engine/commands/model_list_cmd.cc
@@ -10,6 +10,8 @@
 namespace commands {
 
 void ModelListCmd::Exec() {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   auto models_path = file_manager_utils::GetModelsContainerPath();
   cortex::db::Models modellist_handler;
   config::YamlHandler yaml_handler;
@@ -29,7 +31,10 @@ void ModelListCmd::Exec() {
     // auto model_entry = modellist_handler.GetModelInfo(model_handle);
     try {
       count += 1;
-      yaml_handler.ModelConfigFromFile(model_entry.path_to_model_yaml);
+      yaml_handler.ModelConfigFromFile(
+          fmu::ToAbsoluteCortexDataPath(
+              fs::path(model_entry.path_to_model_yaml))
+              .string());
       auto model_config = yaml_handler.GetModelConfig();
       table.add_row({std::to_string(count), model_entry.model,
                      model_entry.model_alias, model_config.engine,

diff --git a/engine/commands/model_upd_cmd.cc b/engine/commands/model_upd_cmd.cc
@@ -1,5 +1,5 @@
 #include "model_upd_cmd.h"
-
+#include "utils/file_manager_utils.h"
 #include "utils/logging_utils.h"
 
 namespace commands {
@@ -9,13 +9,17 @@ ModelUpdCmd::ModelUpdCmd(std::string model_handle)
 
 void ModelUpdCmd::Exec(
     const std::unordered_map<std::string, std::string>& options) {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   try {
     auto model_entry = model_list_utils_.GetModelInfo(model_handle_);
     if (model_entry.has_error()) {
       CLI_LOG("Error: " + model_entry.error());
       return;
     }
-    yaml_handler_.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    auto yaml_fp = fmu::ToAbsoluteCortexDataPath(
+        fs::path(model_entry.value().path_to_model_yaml));
+    yaml_handler_.ModelConfigFromFile(yaml_fp.string());
     model_config_ = yaml_handler_.GetModelConfig();
 
     for (const auto& [key, value] : options) {
@@ -25,7 +29,7 @@ void ModelUpdCmd::Exec(
     }
 
     yaml_handler_.UpdateModelConfig(model_config_);
-    yaml_handler_.WriteYamlFile(model_entry.value().path_to_model_yaml);
+    yaml_handler_.WriteYamlFile(yaml_fp.string());
     CLI_LOG("Successfully updated model ID '" + model_handle_ + "'!");
   } catch (const std::exception& e) {
     CLI_LOG("Failed to update model with model ID '" + model_handle_ +

diff --git a/engine/commands/run_cmd.cc b/engine/commands/run_cmd.cc
@@ -32,12 +32,17 @@ void RunCmd::Exec(bool chat_flag) {
   }
 
   try {
+    namespace fs = std::filesystem;
+    namespace fmu = file_manager_utils;
     auto model_entry = modellist_handler.GetModelInfo(*model_id);
     if (model_entry.has_error()) {
       CLI_LOG("Error: " + model_entry.error());
       return;
     }
-    yaml_handler.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    yaml_handler.ModelConfigFromFile(
+        fmu::ToAbsoluteCortexDataPath(
+            fs::path(model_entry.value().path_to_model_yaml))
+            .string());
     auto mc = yaml_handler.GetModelConfig();
 
     // Check if engine existed. If not, download it

diff --git a/engine/config/yaml_config.cc b/engine/config/yaml_config.cc
@@ -3,9 +3,9 @@
 #include <fstream>
 #include <iostream>
 #include <string>
-using namespace std;
 
 #include "utils/format_utils.h"
+#include "utils/file_manager_utils.h"
 #include "yaml_config.h"
 namespace config {
 // Method to read YAML file
@@ -14,6 +14,8 @@ void YamlHandler::Reset() {
   yaml_node_.reset();
 };
 void YamlHandler::ReadYamlFile(const std::string& file_path) {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   try {
     yaml_node_ = YAML::LoadFile(file_path);
     // incase of model.yml file, we don't have files yet, create them
@@ -24,8 +26,9 @@ void YamlHandler::ReadYamlFile(const std::string& file_path) {
       std::vector<std::string> v;
       if (yaml_node_["engine"] &&
           yaml_node_["engine"].as<std::string>() == "cortex.llamacpp") {
-        // TODO: change prefix to models:// with source from cortexso
-        v.emplace_back(s.substr(0, s.find_last_of('/')) + "/model.gguf");
+        auto abs_path = s.substr(0, s.find_last_of('/')) + "/model.gguf";
+        auto rel_path = fmu::ToRelativeCortexDataPath(fs::path(abs_path));
+        v.emplace_back(rel_path.string());
       } else {
         v.emplace_back(s.substr(0, s.find_last_of('/')));
       }
@@ -286,9 +289,7 @@ void YamlHandler::WriteYamlFile(const std::string& file_path) const {
       outFile << "version: " << yaml_node_["version"].as<std::string>() << "\n";
     }
     if (yaml_node_["files"] && yaml_node_["files"].size()) {
-      outFile << "files:             # can be universal protocol (models://) "
-                 "OR absolute local file path (file://) OR https remote URL "
-                 "(https://)\n";
+      outFile << "files:             # Can be relative OR absolute local file path\n";
       for (const auto& source : yaml_node_["files"]) {
         outFile << "  - " << source << "\n";
       }

diff --git a/engine/controllers/models.cc b/engine/controllers/models.cc
@@ -59,6 +59,8 @@ void Models::PullModel(const HttpRequestPtr& req,
 void Models::ListModel(
     const HttpRequestPtr& req,
     std::function<void(const HttpResponsePtr&)>&& callback) const {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   Json::Value ret;
   ret["object"] = "list";
   Json::Value data(Json::arrayValue);
@@ -73,8 +75,10 @@ void Models::ListModel(
     for (const auto& model_entry : list_entry.value()) {
       // auto model_entry = modellist_handler.GetModelInfo(model_handle);
       try {
-
-        yaml_handler.ModelConfigFromFile(model_entry.path_to_model_yaml);
+        yaml_handler.ModelConfigFromFile(
+            fmu::ToAbsoluteCortexDataPath(
+                fs::path(model_entry.path_to_model_yaml))
+                .string());
         auto model_config = yaml_handler.GetModelConfig();
         Json::Value obj = model_config.ToJson();
 
@@ -106,6 +110,8 @@ void Models::ListModel(
 void Models::GetModel(const HttpRequestPtr& req,
                       std::function<void(const HttpResponsePtr&)>&& callback,
                       const std::string& model_id) const {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   LOG_DEBUG << "GetModel, Model handle: " << model_id;
   Json::Value ret;
   ret["object"] = "list";
@@ -125,7 +131,10 @@ void Models::GetModel(const HttpRequestPtr& req,
       callback(resp);
       return;
     }
-    yaml_handler.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    yaml_handler.ModelConfigFromFile(
+        fmu::ToAbsoluteCortexDataPath(
+            fs::path(model_entry.value().path_to_model_yaml))
+            .string());
     auto model_config = yaml_handler.GetModelConfig();
 
     Json::Value obj = model_config.ToJson();
@@ -137,8 +146,8 @@ void Models::GetModel(const HttpRequestPtr& req,
     resp->setStatusCode(k200OK);
     callback(resp);
   } catch (const std::exception& e) {
-    std::string message = "Fail to get model information with ID '" +
-                          model_id + "': " + e.what();
+    std::string message =
+        "Fail to get model information with ID '" + model_id + "': " + e.what();
     LOG_ERROR << message;
     ret["data"] = data;
     ret["result"] = "Fail to get model information";
@@ -171,16 +180,20 @@ void Models::DeleteModel(const HttpRequestPtr& req,
 void Models::UpdateModel(const HttpRequestPtr& req,
                          std::function<void(const HttpResponsePtr&)>&& callback,
                          const std::string& model_id) const {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   auto json_body = *(req->getJsonObject());
   try {
     cortex::db::Models model_list_utils;
     auto model_entry = model_list_utils.GetModelInfo(model_id);
     config::YamlHandler yaml_handler;
-    yaml_handler.ModelConfigFromFile(model_entry.value().path_to_model_yaml);
+    auto yaml_fp = fmu::ToAbsoluteCortexDataPath(
+        fs::path(model_entry.value().path_to_model_yaml));
+    yaml_handler.ModelConfigFromFile(yaml_fp.string());
     config::ModelConfig model_config = yaml_handler.GetModelConfig();
     model_config.FromJson(json_body);
     yaml_handler.UpdateModelConfig(model_config);
-    yaml_handler.WriteYamlFile(model_entry.value().path_to_model_yaml);
+    yaml_handler.WriteYamlFile(yaml_fp.string());
     std::string message = "Successfully update model ID '" + model_id +
                           "': " + json_body.toStyledString();
     LOG_INFO << message;
@@ -210,6 +223,8 @@ void Models::UpdateModel(const HttpRequestPtr& req,
 void Models::ImportModel(
     const HttpRequestPtr& req,
     std::function<void(const HttpResponsePtr&)>&& callback) const {
+  namespace fs = std::filesystem;
+  namespace fmu = file_manager_utils;
   if (!http_util::HasFieldInReq(req, callback, "model") ||
       !http_util::HasFieldInReq(req, callback, "modelPath")) {
     return;
@@ -219,14 +234,18 @@ void Models::ImportModel(
   config::GGUFHandler gguf_handler;
   config::YamlHandler yaml_handler;
   cortex::db::Models modellist_utils_obj;
-
   std::string model_yaml_path = (file_manager_utils::GetModelsContainerPath() /
                                  std::filesystem::path("imported") /
                                  std::filesystem::path(modelHandle + ".yml"))
                                     .string();
-  cortex::db::ModelEntry model_entry{modelHandle, "local", "imported",
-                                     model_yaml_path, modelHandle};
+
   try {
+    // Use relative path for model_yaml_path. In case of import, we use absolute path for model
+    auto yaml_rel_path =
+        fmu::ToRelativeCortexDataPath(fs::path(model_yaml_path));
+    cortex::db::ModelEntry model_entry{modelHandle, "local", "imported",
+                                       yaml_rel_path.string(), modelHandle};
+
     std::filesystem::create_directories(
         std::filesystem::path(model_yaml_path).parent_path());
     gguf_handler.Parse(modelPath);
@@ -295,13 +314,13 @@ void Models::SetModelAlias(
     if (result.has_error()) {
       std::string message = result.error();
       LOG_ERROR << message;
-        Json::Value ret;
-        ret["result"] = "Set alias failed!";
-        ret["modelHandle"] = model_handle;
-        ret["message"] = message;
-        auto resp = cortex_utils::CreateCortexHttpJsonResponse(ret);
-        resp->setStatusCode(k400BadRequest);
-        callback(resp);
+      Json::Value ret;
+      ret["result"] = "Set alias failed!";
+      ret["modelHandle"] = model_handle;
+      ret["message"] = message;
+      auto resp = cortex_utils::CreateCortexHttpJsonResponse(ret);
+      resp->setStatusCode(k400BadRequest);
+      callback(resp);
     } else {
       if (result.value()) {
         std::string message = "Successfully set model alias '" + model_alias +

diff --git a/engine/main.cc b/engine/main.cc
@@ -46,8 +46,8 @@ void RunServer() {
       std::filesystem::path(config.logFolderPath) /
       std::filesystem::path(cortex_utils::logs_folder));
   trantor::FileLogger asyncFileLogger;
-  asyncFileLogger.setFileName(config.logFolderPath + "/" +
-                              cortex_utils::logs_base_name);
+  asyncFileLogger.setFileName((std::filesystem::path(config.logFolderPath) /
+      std::filesystem::path(cortex_utils::logs_base_name)).string());
   asyncFileLogger.setMaxLines(config.maxLogLines);  // Keep last 100000 lines
   asyncFileLogger.startLogging();
   trantor::Logger::setOutputFunction(

diff --git a/engine/services/download_service.cc b/engine/services/download_service.cc
@@ -173,6 +173,7 @@ cpp::result<bool, std::string> DownloadService::Download(
       CTL_INF("Existing file size: " << download_item.downloadUrl << " - "
                                      << download_item.localPath.string()
                                      << " - " << existing_file_size);
+      CTL_INF("Download item size: " << download_item.bytes.value());                               
       auto missing_bytes = download_item.bytes.value() - existing_file_size;
       if (missing_bytes > 0) {
         CLI_LOG("Found unfinished download! Additional "