fetch_dataset function uploads dataset from docker image, ruff format

Author: Nigma-Ks

backend/launch_service/app_setup.py

@@ -1,6 +1,5 @@
 import os
 import subprocess
-import requests
 from backend.config.paths import (
     RESOURCES_DIR,
     DATASET_FILE,
@@ -10,19 +9,33 @@
 from backend.utils.methods_handler import Methods
 
 IMAGE_NAME = "pysymgym-test"
-URL = "https://raw.githubusercontent.com/PySymGym/PySymGym/main/maps/DotNet/Maps/dataset.json"
 
 
-def fetch_dataset(url, data_upload_file):
-    print(f"Downloading dataset from {url} ...")
-    resp = requests.get(url)
-    resp.raise_for_status()
+def fetch_dataset(data_upload_file):
+    container_name = "temp-fetch-dataset"
 
-    with open(data_upload_file, "wb") as f:
-        f.write(resp.content)
+    try:
+        subprocess.run(
+            ["docker", "create", "--name", container_name, IMAGE_NAME],
+            check=True,
+            capture_output=True,
+        )
 
-    print(f"Dataset saved to {data_upload_file}")
-    return data_upload_file
+        subprocess.run(
+            [
+                "docker",
+                "cp",
+                f"{container_name}:/workspace/PySymGym/maps/DotNet/Maps/dataset.json",
+                data_upload_file,
+            ],
+            check=True,
+        )
+
+        print(f"Dataset copied to {data_upload_file}")
+        return data_upload_file
+
+    finally:
+        subprocess.run(["docker", "rm", container_name], capture_output=True)
 
 
 def build_container():
@@ -43,5 +56,5 @@ def update_frontend_selection_options(dataset_file, selection_options_file):
 
 if __name__ == "__main__":
     build_container()
-    fetch_dataset(URL, DATASET_FILE)
+    fetch_dataset(DATASET_FILE)
     update_frontend_selection_options(DATASET_FILE, METHODS_TS_FILE)

backend/launch_service/tests/test_setup_functions.py

@@ -1,29 +1,39 @@
-import os
+import subprocess
 from unittest.mock import Mock, patch
 
 from backend.config.paths import DOCKER_DIR, RESOURCES_DIR
 from backend.launch_service.app_setup import fetch_dataset, IMAGE_NAME, build_container
 
 
-def test_fetch_dataset(tmp_path):
-    url = "https://example.com/dataset.csv"
-    test_content = b"dll,method\nManuallyCollected.dll,BinSearchMain"
-    data_file = tmp_path / "dataset.json"
+def test_fetch_dataset_success(tmp_path, monkeypatch):
+    test_file = tmp_path / "dataset.json"
+    commands_executed = []
 
-    with patch("requests.get") as mock_get:
-        mock_response = Mock()
-        mock_response.content = test_content
-        mock_response.raise_for_status = Mock()
-        mock_get.return_value = mock_response
+    class MockResult:
+        def __init__(self, returncode=0, stdout=b"", stderr=b""):
+            self.returncode = returncode
+            self.stdout = stdout
+            self.stderr = stderr
 
-        result = fetch_dataset(url, data_file)
+    def mock_subprocess_run(*args, **kwargs):
+        commands_executed.append(args[0])
 
-    assert result == data_file
-    assert os.path.exists(data_file)
-    with open(data_file, "rb") as f:
-        assert f.read() == test_content
-    mock_get.assert_called_once_with(url)
-    mock_response.raise_for_status.assert_called_once()
+        if args[0][0] == "docker" and args[0][1] == "cp":
+            test_file.write_text('{"test": "data"}')
+
+        return MockResult()
+
+    monkeypatch.setattr(subprocess, "run", mock_subprocess_run)
+
+    result = fetch_dataset(str(test_file))
+
+    assert result == str(test_file)
+    assert test_file.exists()
+    assert test_file.read_text() == '{"test": "data"}'
+    assert len(commands_executed) == 3
+    assert commands_executed[0][:3] == ["docker", "create", "--name"]
+    assert commands_executed[1][:2] == ["docker", "cp"]
+    assert commands_executed[2][:2] == ["docker", "rm"]
 
 
 def test_successful_build():

backend/utils/tests/test_methods_handler.py

@@ -18,7 +18,7 @@ def test_write_selection_dataset_to_front_file():
         content = temp_file.read()
 
         assert content.startswith("export const METHODS = ")
-        json_str = content[len("export const METHODS = "):]
+        json_str = content[len("export const METHODS = ") :]
         written_data = json.loads(json_str)
         assert written_data == test_data
 
@@ -29,7 +29,7 @@ def test_write_selection_dataset_to_front_file_empty():
 
         temp_file.seek(0)
         content = temp_file.read()
-        json_str = content[len("export const METHODS = "):]
+        json_str = content[len("export const METHODS = ") :]
         assert json.loads(json_str) == []