chore: import upstream snapshot with attribution

tests/test_cli.py

@@ -0,0 +1,80 @@
+from contextlib import redirect_stdout
+from io import StringIO
+from unittest import mock
+
+import pytest
+
+from litgpt.__main__ import main
+
+
+def test_cli():
+    out = StringIO()
+    with pytest.raises(SystemExit), redirect_stdout(out), mock.patch("sys.argv", ["litgpt", "-h"]):
+        main()
+    out = out.getvalue()
+    assert "usage: litgpt" in out
+    assert (
+        "{download,chat,finetune,finetune_lora,finetune_full,finetune_adapter,finetune_adapter_v2,"
+        "pretrain,generate,generate_full,generate_adapter,generate_adapter_v2,generate_sequentially,"
+        "generate_speculatively,generate_tp,convert_to_litgpt,convert_from_litgpt,convert_pretrained_checkpoint,"
+        "merge_lora,evaluate,serve,validate}" in out
+    )
+    assert (
+        """Available subcommands:
+    download            Download weights or tokenizer data from the Hugging
+                        Face Hub.
+    chat                Chat with a model."""
+        in out
+    )
+    assert """evaluate            Evaluate a model with the LM Evaluation Harness.""" in out
+    assert """serve               Serve a LitGPT model using LitServe.""" in out
+    out = StringIO()
+    with pytest.raises(SystemExit), redirect_stdout(out), mock.patch("sys.argv", ["litgpt", "finetune_lora", "-h"]):
+        main()
+    out = out.getvalue()
+    assert (
+        """--lora_alpha LORA_ALPHA
+                        The LoRA alpha. (type: int, default: 16)"""
+        in out
+    )
+
+    out = StringIO()
+    with pytest.raises(SystemExit), redirect_stdout(out), mock.patch("sys.argv", ["litgpt", "pretrain", "-h"]):
+        main()
+    out = out.getvalue()
+    print(out)
+    assert (
+        """--train.max_tokens MAX_TOKENS
+                        Total number of tokens to train on (type: int | None,
+                        default: 3000000000000)"""
+        in out
+    )
+
+
+def test_pretrain_allows_max_steps():
+    # Ensure --train.max_steps is accepted by the CLI for pretrain
+    # and only emits a warning instead of raising a validation error.
+    args = [
+        "litgpt",
+        "pretrain",
+        "pythia-14m",
+        "--train.max_steps=1",
+        "--out_dir=out/test-cli",
+    ]
+
+    with pytest.warns(UserWarning, match="max_steps"):
+        try:
+            with mock.patch("sys.argv", args):
+                main()
+        except Exception:
+            pass
+
+
+def test_rewrite_finetune_command():
+    out1 = StringIO()
+    with pytest.raises(SystemExit), redirect_stdout(out1), mock.patch("sys.argv", ["litgpt", "fineune", "-h"]):
+        main()
+    out2 = StringIO()
+    with pytest.raises(SystemExit), redirect_stdout(out2), mock.patch("sys.argv", ["litgpt", "fineune_lora", "-h"]):
+        main()
+    assert out1.getvalue() == out2.getvalue()