Add output cleanup policy

tests/test_app.py

@@ -1,4 +1,6 @@
 import io
+import json
+from pathlib import Path
 
 from app import create_app
 
@@ -132,6 +134,95 @@ def test_download_route_returns_generated_file(client):
     download_response.close()
 
 
+def test_download_route_can_cleanup_files_after_download(tmp_path):
+    class CleanupAfterDownloadConfig:
+        TESTING = True
+        SECRET_KEY = "test-secret"
+        MAX_CONTENT_LENGTH = 100 * 1024 * 1024
+        PREVIEW_RECORD_LIMIT = 5
+        OUTPUT_DIRECTORY = tmp_path / "download-cleanup-outputs"
+        OUTPUT_RETENTION_HOURS = 24
+        CLEANUP_ON_STARTUP = False
+        CLEANUP_AFTER_DOWNLOAD = True
+
+    app = create_app(CleanupAfterDownloadConfig)
+    client = app.test_client()
+    log_file = io.BytesIO(SAMPLE_LOG.encode("utf-8"))
+
+    convert_response = client.post(
+        "/convert",
+        data={
+            "mode": "vendor",
+            "output_format": "csv",
+            "sort_by": "datetime",
+            "order": "asc",
+            "policy_cs": "",
+            "policy_ci": "",
+            "severity_cs": "",
+            "severity_ci": "",
+            "log_file": (log_file, "sample.log"),
+        },
+        content_type="multipart/form-data",
+    )
+    log_file.close()
+
+    html = convert_response.data.decode("utf-8")
+    marker = "/download/"
+    start = html.index(marker) + len(marker)
+    end = html.index('"', start)
+    result_id = html[start:end]
+    metadata_path = Path(app.config["OUTPUT_DIRECTORY"]) / f"{result_id}.json"
+
+    download_response = client.get(f"/download/{result_id}")
+    download_response.close()
+    convert_response.close()
+
+    assert not metadata_path.exists()
+
+
+def test_cleanup_on_startup_removes_expired_outputs(tmp_path):
+    output_dir = tmp_path / "startup-cleanup-outputs"
+    output_dir.mkdir(parents=True)
+    result_id = "expired-result"
+    file_path = output_dir / f"{result_id}.csv"
+    metadata_path = output_dir / f"{result_id}.json"
+    file_path.write_text("header\nvalue\n", encoding="utf-8")
+    metadata_path.write_text(
+        json.dumps(
+            {
+                "result_id": result_id,
+                "file_path": str(file_path),
+                "download_name": "waf-report.csv",
+                "mimetype": "text/csv; charset=utf-8",
+            }
+        ),
+        encoding="utf-8",
+    )
+    old_timestamp = 946684800
+    file_path.touch()
+    metadata_path.touch()
+    Path(file_path).touch()
+    import os
+
+    os.utime(file_path, (old_timestamp, old_timestamp))
+    os.utime(metadata_path, (old_timestamp, old_timestamp))
+
+    class StartupCleanupConfig:
+        TESTING = True
+        SECRET_KEY = "test-secret"
+        MAX_CONTENT_LENGTH = 100 * 1024 * 1024
+        PREVIEW_RECORD_LIMIT = 5
+        OUTPUT_DIRECTORY = output_dir
+        OUTPUT_RETENTION_HOURS = 1
+        CLEANUP_ON_STARTUP = True
+        CLEANUP_AFTER_DOWNLOAD = False
+
+    create_app(StartupCleanupConfig)
+
+    assert not file_path.exists()
+    assert not metadata_path.exists()
+
+
 def test_default_upload_limit_is_100_mib(app):
     assert app.config["MAX_CONTENT_LENGTH"] == 100 * 1024 * 1024