[DGL-Go] Change name to dglgo (#3778)

* add

* remove

* fix

* rework the readme and some changes

* add png

* update png

* add recipe get
Co-authored-by: Minjie Wang <wmjlyjemaine@gmail.com>
Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

enter/recipes/linkpred_cora_sage.yml → dglgo/recipes/linkpred_cora_sage.yaml

@@ -31,4 +31,5 @@ general_pipeline:
     name: Adam
     lr: 0.005
   loss: BCELoss
+  save_path: "model.pth"
   num_runs: 1                 # Number of experiments to run

enter/recipes/nodepred-ns_arxiv_gcn.yml → dglgo/recipes/nodepred-ns_arxiv_gcn.yaml

@@ -31,4 +31,5 @@ general_pipeline:
     lr: 0.005
     weight_decay: 0.0
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 5

enter/recipes/nodepred-ns_product_sage.yml → dglgo/recipes/nodepred-ns_product_sage.yaml

@@ -35,4 +35,5 @@ general_pipeline:
     lr: 0.005
     weight_decay: 0.0
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 5                 # Number of experiments to run

enter/recipes/nodepred_citeseer_gat.yml → dglgo/recipes/nodepred_citeseer_gat.yaml

@@ -28,4 +28,5 @@ general_pipeline:
     lr: 0.005
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10               # Number of experiments to run

enter/recipes/nodepred_citeseer_gcn.yml → dglgo/recipes/nodepred_citeseer_gcn.yaml

@@ -24,4 +24,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_citeseer_sage.yml → dglgo/recipes/nodepred_citeseer_sage.yaml

@@ -23,4 +23,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_cora_gat.yml → dglgo/recipes/nodepred_cora_gat.yaml

@@ -28,4 +28,5 @@ general_pipeline:
     lr: 0.005
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_cora_gcn.yml → dglgo/recipes/nodepred_cora_gcn.yaml

@@ -24,4 +24,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_cora_sage.yml → dglgo/recipes/nodepred_cora_sage.yaml

@@ -23,4 +23,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_pubmed_gat.yml → dglgo/recipes/nodepred_pubmed_gat.yaml

@@ -28,4 +28,5 @@ general_pipeline:
     lr: 0.005
     weight_decay: 0.001
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_pubmed_gcn.yml → dglgo/recipes/nodepred_pubmed_gcn.yaml

@@ -24,4 +24,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/recipes/nodepred_pubmed_sage.yml → dglgo/recipes/nodepred_pubmed_sage.yaml

@@ -23,4 +23,5 @@ general_pipeline:
     lr: 0.01
     weight_decay: 0.0005
   loss: CrossEntropyLoss
+  save_path: "model.pth"
   num_runs: 10                # Number of experiments to run

enter/setup.py → dglgo/setup.py

@@ -3,7 +3,7 @@
 from setuptools import find_packages
 from distutils.core import setup
 
-setup(name='dglenter',
+setup(name='dglgo',
       version='0.0.1',
       description='DGL',
       author='DGL Team',
@@ -15,12 +15,15 @@ setup(name='dglenter',
           'autopep8>=1.6.0',
           'numpydoc>=1.1.0',
           "pydantic>=1.9.0",
-          "ruamel.yaml>=0.17.20"
+          "ruamel.yaml>=0.17.20",
+          "PyYAML>=5.1"
       ],
+      package_data={"": ["./*"]},
+      include_package_data=True,
       license='APACHE',
       entry_points={
           'console_scripts': [
-              "dgl-enter = dglenter.cli.cli:main"
+              "dgl = dglgo.cli.cli:main"
           ]
       },
       url='https://github.com/dmlc/dgl',

dglgo/tests/cfg.yml

+version: 0.0.1
+pipeline_name: nodepred
+device: cpu
+data:
+  name: cora
+  split_ratio:                # Ratio to generate split masks, for example set to [0.8, 0.1, 0.1] for 80% train/10% val/10% test. Leave blank to use builtin split in original dataset
+model:
+  name: sage
+  embed_size: -1              # The dimension of created embedding table. -1 means using original node embedding
+  hidden_size: 16             # Hidden size.
+  num_layers: 1               # Number of hidden layers.
+  activation: relu            # Activation function name under torch.nn.functional
+  dropout: 0.5                # Dropout rate.
+  aggregator_type: gcn        # Aggregator type to use (``mean``, ``gcn``, ``pool``, ``lstm``).
+general_pipeline:
+  early_stop:
+    patience: 20              # Steps before early stop
+    checkpoint_path: checkpoint.pth # Early stop checkpoint model file path
+  num_epochs: 200             # Number of training epochs
+  eval_period: 5              # Interval epochs between evaluations
+  optimizer:
+    name: Adam
+    lr: 0.01
+    weight_decay: 0.0005
+  loss: CrossEntropyLoss
+  num_runs: 1                 # Number of experiments to run

dglgo/tests/run_test.sh

+python -m pytest --pdb -vv --capture=tee-sys test_pipeline.py::test_recipe
\ No newline at end of file

dglgo/tests/test_pipeline.py

+import subprocess
+from typing import NamedTuple
+import pytest
+from pathlib import Path
+# class DatasetSpec:
+
+dataset_spec = {
+    "cora": {"timeout": 30}
+}
+
+
+
+class ExperimentSpec(NamedTuple):
+    pipeline: str
+    dataset: str
+    model: str
+    timeout: int
+    extra_cfg: dict = {}
+
+exps = [ExperimentSpec(pipeline="nodepred", dataset="cora", model="sage", timeout=0.5)]
+
+@pytest.mark.parametrize("spec", exps)
+def test_train(spec):
+    cfg_path = "/tmp/test.yaml"
+    run = subprocess.run(["dgl", "config", spec.pipeline, "--data", spec.dataset, "--model", spec.model, "--cfg", cfg_path], timeout=spec.timeout, capture_output=True)
+    assert run.stderr is None or len(run.stderr) == 0, "Found error message: {}".format(run.stderr)
+    output = run.stdout.decode("utf-8")
+    print(output)
+
+    run = subprocess.run(["dgl", "train", "--cfg", cfg_path], timeout=spec.timeout, capture_output=True)
+    assert run.stderr is None or len(run.stderr) == 0, "Found error message: {}".format(run.stderr)
+    output = run.stdout.decode("utf-8")
+    print(output)
+
+TEST_RECIPE_FOLDER = "my_recipes"
+
+@pytest.fixture
+def setup_recipe_folder():
+    run = subprocess.run(["dgl", "recipe", "copy", "--dir", TEST_RECIPE_FOLDER], timeout=15, capture_output=True)
+
+@pytest.mark.parametrize("file", [str(f) for f in Path(TEST_RECIPE_FOLDER).glob("*.yaml")])
+def test_recipe(file, setup_recipe_folder):
+    print("DGL enter train {}".format(file))
+    try:    
+        run = subprocess.run(["dgl", "train", "--cfg", file], timeout=5, capture_output=True)
+        sh_stdout, sh_stderr = run.stdout, run.stderr
+    except subprocess.TimeoutExpired as e:
+        sh_stdout = e.stdout
+        sh_stderr = e.stderr
+    if sh_stderr is not None and len(sh_stderr) != 0:
+        error_str = sh_stderr.decode("utf-8")
+        lines = error_str.split("\n")
+        for line in lines:
+            line = line.strip()
+            if line.startswith("WARNING") or line.startswith("Aborted") or line.startswith("0%"):
+                continue
+            else:
+                assert len(line) == 0, error_str
+    print("{} stdout: {}".format(file, sh_stdout))
+    print("{} stderr: {}".format(file, sh_stderr))
+
+# test_recipe( , None)
\ No newline at end of file