From 91b654c1bb43772607f12f36e9a41e99a2e44d35 Mon Sep 17 00:00:00 2001
From: jiatolentino <tolentino.joseireneo@gmail.com>
Date: Tue, 16 Jun 2026 12:28:36 +0800
Subject: [PATCH] feat: write split db-report downloads as a zip

---
 src/datamasque_cli/commands/discovery.py | 22 ++++++++++++++++--
 tests/commands/test_discovery.py         | 29 +++++++++++++++++++++++-
 2 files changed, 48 insertions(+), 3 deletions(-)

diff --git a/src/datamasque_cli/commands/discovery.py b/src/datamasque_cli/commands/discovery.py
index 9db83af..5420b1d 100644
--- a/src/datamasque_cli/commands/discovery.py
+++ b/src/datamasque_cli/commands/discovery.py
@@ -105,12 +105,30 @@ def sdd_report(
 @app.command("db-report")
 def db_discovery_report(
     run_id: int = typer.Argument(help="Run ID"),
-    output: Path | None = typer.Option(None, "--output", "-o", help="Write CSV to this path"),
+    output: Path | None = typer.Option(None, "--output", "-o", help="Write CSV (or zip) to this path"),
     profile: str | None = typer.Option(None, "--profile", "-p", help="Profile to use"),
 ) -> None:
-    """Download database discovery report (CSV) for a run."""
+    """Download database discovery report for a run.
+
+    Reports within Excel's row limit download as a single CSV. Larger reports are split
+    server-side into a zip of numbered CSV parts, written out as a binary `.zip`; that case
+    requires `-o`, since a zip can't be streamed to stdout.
+    """
     client = get_client(profile)
     report = client.get_db_discovery_result_report(RunId(run_id))
+
+    if isinstance(report, bytes):
+        if output is None:
+            abort(
+                "This report was split into a zip and can't be written to stdout.",
+                code=ErrorCode.INVALID_INPUT,
+                hint=f"Re-run with -o <path>.zip (e.g. -o discovery_report_{run_id}.zip).",
+            )
+        target = output if output.suffix.lower() == ".zip" else output.with_suffix(".zip")
+        target.write_bytes(report)
+        print_success(f"Database discovery report (split, zip) written to {target}")
+        return
+
     _write_or_echo(report, output, "Database discovery report")
 
 
diff --git a/tests/commands/test_discovery.py b/tests/commands/test_discovery.py
index c35218b..7a3d077 100644
--- a/tests/commands/test_discovery.py
+++ b/tests/commands/test_discovery.py
@@ -37,7 +37,7 @@ def test_sdd_report_echoes_to_stdout_without_output(mock_get_client: MagicMock,
 
 
 @patch(f"{MODULE}.get_client")
-def test_db_report_writes_to_output_file(mock_get_client: MagicMock, runner: CliRunner, tmp_path: Path) -> None:
+def test_db_report_writes_csv_to_output_file(mock_get_client: MagicMock, runner: CliRunner, tmp_path: Path) -> None:
     client = MagicMock()
     mock_get_client.return_value = client
     client.get_db_discovery_result_report.return_value = "header\nrow1\n"
@@ -49,6 +49,33 @@ def test_db_report_writes_to_output_file(mock_get_client: MagicMock, runner: Cli
     assert out.read_text() == "header\nrow1\n"
 
 
+@patch(f"{MODULE}.get_client")
+def test_db_report_writes_zip_when_split(mock_get_client: MagicMock, runner: CliRunner, tmp_path: Path) -> None:
+    client = MagicMock()
+    mock_get_client.return_value = client
+    zip_bytes = b"PK\x03\x04 fake zip bytes"
+    client.get_db_discovery_result_report.return_value = zip_bytes
+
+    out = tmp_path / "db.csv"
+    result = runner.invoke(app, ["discover", "db-report", "42", "--output", str(out)])
+
+    assert result.exit_code == 0
+    assert not out.exists()
+    assert (tmp_path / "db.zip").read_bytes() == zip_bytes
+
+
+@patch(f"{MODULE}.get_client")
+def test_db_report_split_without_output_aborts(mock_get_client: MagicMock, runner: CliRunner) -> None:
+    client = MagicMock()
+    mock_get_client.return_value = client
+    client.get_db_discovery_result_report.return_value = b"PK\x03\x04 fake zip bytes"
+
+    result = runner.invoke(app, ["discover", "db-report", "42"])
+
+    assert result.exit_code != 0
+    assert "-o" in result.stderr
+
+
 @patch(f"{MODULE}.get_client")
 def test_file_report_writes_json_to_output(mock_get_client: MagicMock, runner: CliRunner, tmp_path: Path) -> None:
     client = MagicMock()