feat: write json to file

.gitignore

@@ -1,2 +1,4 @@
 .venv
 __pycache__
+
+data/

oc4ids_datastore_pipeline/pipeline.py

@@ -1,4 +1,6 @@
+import json
 import logging
+import os
 from typing import Any
 
 import requests
@@ -41,11 +43,23 @@ def validate_json(dataset_name: str, json_data: Any) -> None:
         raise Exception("Validation failed", e)
 
 
+def write_json_to_file(file_name: str, json_data: Any) -> None:
+    logger.info(f"Writing dataset to file {file_name}")
+    try:
+        os.makedirs(os.path.dirname(file_name), exist_ok=True)
+        with open(file_name, "w") as file:
+            json.dump(json_data, file, indent=4)
+        logger.info(f"Finished writing to {file_name}")
+    except Exception as e:
+        raise Exception("Error while writing to JSON file", e)
+
+
 def process_dataset(dataset_name: str, dataset_url: str) -> None:
     logger.info(f"Processing dataset {dataset_name}")
     try:
         json_data = download_json(dataset_url)
         validate_json(dataset_name, json_data)
+        write_json_to_file(f"data/{dataset_name}.json", json_data)
         logger.info(f"Processed dataset {dataset_name}")
     except Exception as e:
         logger.warning(f"Failed to process dataset {dataset_name} with error {e}")

tests/test_pipeline.py

@@ -1,10 +1,15 @@
+import os
+import tempfile
+from textwrap import dedent
+
 import pytest
 from pytest_mock import MockerFixture
 
 from oc4ids_datastore_pipeline.pipeline import (
     download_json,
     process_dataset,
     validate_json,
+    write_json_to_file,
 )
 
 
@@ -47,6 +52,34 @@ def test_validate_json_raises_validation_errors_exception(
     assert "Dataset has 2 validation errors" in str(exc_info.value)
 
 
+def test_write_json_to_file_writes_in_correct_format() -> None:
+    with tempfile.TemporaryDirectory() as dir:
+        file_name = os.path.join(dir, "test_dataset.json")
+        write_json_to_file(file_name=file_name, json_data={"key": "value"})
+
+        expected = dedent(
+            """\
+            {
+                "key": "value"
+            }"""
+        )
+        with open(file_name) as file:
+            assert file.read() == expected
+
+
+def test_write_json_to_file_raises_failure_exception(mocker: MockerFixture) -> None:
+    patch_json_dump = mocker.patch("oc4ids_datastore_pipeline.pipeline.json.dump")
+    patch_json_dump.side_effect = Exception("Mocked exception")
+
+    with pytest.raises(Exception) as exc_info:
+        with tempfile.TemporaryDirectory() as dir:
+            file_name = os.path.join(dir, "test_dataset.json")
+            write_json_to_file(file_name=file_name, json_data={"key": "value"})
+
+            assert "Error while writing to JSON file" in str(exc_info.value)
+            assert "Mocked exception" in str(exc_info.value)
+
+
 def test_process_dataset_catches_exception(mocker: MockerFixture) -> None:
     patch_download_json = mocker.patch(
         "oc4ids_datastore_pipeline.pipeline.download_json"