[Test Framework] Add test scripts for serve.py (#92)

* add test for serve.py

* nit

* cut down config combos

* fix

* fix compile

* change to self-hosted

* test

* add log

* update

* add exit code check

* nit

* nit

* nit

* Update tests/inference/test_serve.py

Signed-off-by: Xiaochang Wu <[email protected]>

* Update tests/inference/test_serve.py

Signed-off-by: Xiaochang Wu <[email protected]>

---------

Signed-off-by: Yizhong Zhang <[email protected]>
Signed-off-by: Xiaochang Wu <[email protected]>
Co-authored-by: Xiaochang Wu <[email protected]>

tests/inference/test_serve.py

@@ -0,0 +1,124 @@
+import subprocess
+import pytest
+
+# Config matrix
+# config_file_array = ["inference/models/gpt2.yaml", None]
+# model_id_or_path_array = ["gpt2", None]
+# models_array = ["gpt2", "gpt2 gpt-j-6b", "gpt2 bloom-560m falcon-7b"]
+# port_array = [8000, None]
+# route_prefix_array = [None]
+# cpus_per_worker_array = [24, None]
+# gpus_per_worker_array = [0, None]
+# hpus_per_worker_array = [0, None]
+# deepspeed_array = [True, False]
+# workers_per_group_array = [2, None]
+# ipex_array = [True, False]
+# device_array = ["cpu", None]
+# serve_local_only_array = [True, False]
+# simple_array = [True, False]
+# keep_serve_termimal_array = [True, False]
+
+
+# Parametrize the test function with different combinations of parameters
+@pytest.mark.parametrize(
+    "config_file, model_id_or_path, models, port, route_prefix, cpus_per_worker, gpus_per_worker, hpus_per_worker, deepspeed, workers_per_group, ipex, device, serve_local_only, simple, keep_serve_termimal",
+    [
+        (
+            config_file,
+            model_id_or_path,
+            models,
+            port,
+            route_prefix,
+            cpus_per_worker,
+            gpus_per_worker,
+            hpus_per_worker,
+            deepspeed,
+            workers_per_group,
+            ipex,
+            device,
+            serve_local_only,
+            simple,
+            keep_serve_termimal,
+        )
+        for config_file in ["../.github/workflows/config/gpt2-ci.yaml"]
+        for model_id_or_path in ["gpt2"]
+        for models in ["gpt2"]
+        for port in [8000]
+        for route_prefix in [None]
+        for cpus_per_worker in [24]
+        for gpus_per_worker in [0]
+        for hpus_per_worker in [0]
+        for deepspeed in [False]
+        for workers_per_group in [None]
+        for ipex in [False]
+        for device in ["cpu"]
+        for serve_local_only in [False]
+        for simple in [False]
+        for keep_serve_termimal in [False]
+    ],
+)
+def test_script(
+    config_file,
+    model_id_or_path,
+    models,
+    port,
+    route_prefix,
+    cpus_per_worker,
+    gpus_per_worker,
+    hpus_per_worker,
+    deepspeed,
+    workers_per_group,
+    ipex,
+    device,
+    serve_local_only,
+    simple,
+    keep_serve_termimal,
+):
+    cmd_serve = ["python", "../llm_on_ray/inference/serve.py"]
+    if config_file is not None:
+        cmd_serve.append("--config_file")
+        cmd_serve.append(str(config_file))
+    if model_id_or_path is not None:
+        cmd_serve.append("--model_id_or_path")
+        cmd_serve.append(str(model_id_or_path))
+    if models is not None:
+        cmd_serve.append("--models")
+        cmd_serve.append(str(models))
+    if port is not None:
+        cmd_serve.append("--port")
+        cmd_serve.append(str(port))
+    if route_prefix is not None:
+        cmd_serve.append("--route_prefix")
+        cmd_serve.append(str(route_prefix))
+    if cpus_per_worker is not None:
+        cmd_serve.append("--cpus_per_worker")
+        cmd_serve.append(str(cpus_per_worker))
+    if gpus_per_worker is not None:
+        cmd_serve.append("--gpus_per_worker")
+        cmd_serve.append(str(gpus_per_worker))
+    if hpus_per_worker is not None:
+        cmd_serve.append("--hpus_per_worker")
+        cmd_serve.append(str(hpus_per_worker))
+    if deepspeed:
+        cmd_serve.append("--deepspeed")
+    if workers_per_group is not None:
+        cmd_serve.append("--workers_per_group")
+        cmd_serve.append(str(workers_per_group))
+    if ipex:
+        cmd_serve.append("--ipex")
+    if device is not None:
+        cmd_serve.append("--device")
+        cmd_serve.append(str(device))
+    if serve_local_only:
+        cmd_serve.append("--serve_local_only")
+    if simple:
+        cmd_serve.append("--simple")
+    if keep_serve_termimal:
+        cmd_serve.append("--keep_serve_termimal")
+
+    result_serve = subprocess.run(cmd_serve, capture_output=True, text=True)
+
+    assert "Error" not in result_serve.stderr
+    assert result_serve.returncode == 0
+    print("Output of stderr:")
+    print(result_serve.stderr)