Fix package data in setup.py for single files. (#416)

* Fix package data in setup.py for single files.

* Add specification of GPT jit build in README.

* Remove tests from package.

* Fix package data leaving out files with same names.

Author: guoshengCS

paddlenlp/__init__.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = '2.0.0rc19'  # Maybe dev is better
+__version__ = '2.0.0'  # Maybe dev is better
 
 from . import data
 from . import datasets

paddlenlp/ops/README.md

@@ -164,7 +164,7 @@ python sample/gpt_sample.py --model_name_or_path gpt2-medium-en --decoding_lib .
 
 其中，各个选项的意义如下：
 * `--model_name_or_path`: 预训练模型的名称或是路径。
-* `--decoding_lib`: 指向 `libdecoding_op.so` 的路径。需要包含 `libdecoding_op.so`。
+* `--decoding_lib`: 指向 `libdecoding_op.so` 的路径。需要包含 `libdecoding_op.so`。若不存在则将自动进行 jit 编译产出该 lib。
 * `--batch_size`: 一个 batch 内，样本数目的大小。
 * `--candidate_num`: 执行 topk-sampling 的时候的 `k` 的大小，默认是 4。
 * `--probability_threshold`: 执行 topp-sampling 的时候的阈值的大小，默认是 0.0 表示不执行 topp-sampling。

paddlenlp/ops/ext_utils.py

@@ -27,7 +27,7 @@ def __init__(self, name, source_dir=None):
         # A CMakeExtension needs a source_dir instead of a file list.
         Extension.__init__(self, name, sources=[])
         if source_dir is None:
-            self.source_dir = Path(__file__).parent.resolve()
+            self.source_dir = str(Path(__file__).parent.resolve())
         else:
             self.source_dir = os.path.abspath(os.path.expanduser(source_dir))
         self.sources = [

setup.py

@@ -31,11 +31,15 @@ def get_package_data_files(package, data, package_dir=None):
     all_files = []
     for f in data:
         path = os.path.join(package_dir, f)
+        if os.path.isfile(path):
+            all_files.append(f)
+            continue
         for root, _dirs, files in os.walk(path, followlinks=True):
             root = os.path.relpath(root, package_dir)
             for file in files:
+                file = os.path.join(root, file)
                 if file not in all_files:
-                    all_files.append(os.path.join(root, file))
+                    all_files.append(file)
     return all_files
 
 
@@ -49,7 +53,7 @@ def get_package_data_files(package, data, package_dir=None):
     long_description_content_type="text/plain",
     url="https://github.com/PaddlePaddle/PaddleNLP",
     packages=setuptools.find_packages(
-        where='.', exclude=('./examples')),
+        where='.', exclude=('examples*', 'tests*')),
     package_data={
         'paddlenlp.ops': get_package_data_files('paddlenlp.ops', [
             'CMakeLists.txt', 'README.md', 'cmake', 'src', 'sample', 'patches',