do normalization

Author: orsinium

dephell_setuptools/_base.py

@@ -1,13 +1,37 @@
 from pathlib import Path
 from typing import Union
 
+from ._constants import FIELDS
+
 
 class BaseReader:
     def __init__(self, path: Union[str, Path]):
+        self.path = self._normalize_path(path, default_name='setup.py')
+
+    @staticmethod
+    def _normalize_path(path: Union[str, Path], default_name: str) -> Path:
         if isinstance(path, str):
             path = Path(path)
         if not path.exists():
             raise FileNotFoundError(str(path))
         if path.is_dir():
-            path /= 'setup.py'
-        self.path = path
+            path /= default_name
+        return path
+
+    @staticmethod
+    def _clean(data: dict):
+        result = dict()
+        for k, v in data.items():
+            if k not in FIELDS:
+                continue
+            if not v or v == 'UNKNOWN':
+                continue
+            result[k] = v
+
+        # split keywords string by words
+        if 'keywords' in result:
+            if isinstance(result['keywords'], str):
+                result['keywords'] = [result['keywords']]
+            result['keywords'] = sum((kw.split() for kw in result['keywords']), [])
+
+        return result

dephell_setuptools/_cfg.py

@@ -1,5 +1,6 @@
 from copy import deepcopy
 from configparser import ConfigParser
+from pathlib import Path
 from typing import Dict, List, Optional, Union
 
 from setuptools.config import ConfigOptionsHandler, ConfigMetadataHandler
@@ -9,6 +10,9 @@
 
 
 class CfgReader(BaseReader):
+    def __init__(self, path: Union[str, Path]):
+        self.path = self._normalize_path(path, default_name='setup.cfg')
+
     @property
     def content(self) -> Optional[Dict[str, Union[List, Dict]]]:
         path = self.path
@@ -29,8 +33,4 @@ def content(self) -> Optional[Dict[str, Union[List, Dict]]]:
         ConfigOptionsHandler(container, options).parse()
         ConfigMetadataHandler(container, options).parse()
 
-        result = dict()
-        for k, v in vars(container).items():
-            if v is not None:
-                result[k] = v
-        return result
+        return self._clean(vars(container))

dephell_setuptools/_cli.py

@@ -1,12 +1,12 @@
 import json
 import sys
 
-from ._manager import ReadersManager
+from ._manager import read_setup
 
 
 def main(argv=None):
     if argv is None:
         argv = sys.argv[1:]
-    result = ReadersManager()(argv[0])
+    result = read_setup(path=argv[0])
     print(json.dumps(result, sort_keys=True, indent=2))
     return 0

dephell_setuptools/_cmd.py

@@ -24,28 +24,29 @@ def cd(path: Path):
 class CommandReader(BaseReader):
     @property
     def content(self):
+        # generate a temporary json file which contains the metadata
+        output_json = NamedTemporaryFile()
+        cmd = [
+            sys.executable,
+            self.path.name,
+            '-q',
+            '--command-packages', 'dephell_setuptools',
+            'distutils_cmd',
+            '-o', output_json.name,
+        ]
         with cd(self.path.parent):
-            # generate a temporary json file which contains the metadata
-            output_json = NamedTemporaryFile()
-            cmd = [
-                sys.executable,
-                self.path.name,
-                '-q',
-                '--command-packages', 'dephell_setuptools',
-                'distutils_cmd',
-                '-o', output_json.name,
-            ]
             result = subprocess.run(
                 cmd,
                 stderr=subprocess.PIPE,
                 stdout=subprocess.PIPE,
                 env={'PYTHONPATH': str(Path(__file__).parent.parent)},
             )
-            if result.returncode != 0:
-                return None
+        if result.returncode != 0:
+            return None
 
-            with open(output_json.name) as stream:
-                return json.load(stream)
+        with open(output_json.name) as stream:
+            result = json.load(stream)
+        return self._clean(result)
 
 
 class JSONCommand(Command):

dephell_setuptools/_manager.py

@@ -1,34 +1,32 @@
 from logging import getLogger
 from pathlib import Path
+from typing import Any, Callable, Iterable
 
 from ._cfg import CfgReader
 from ._cmd import CommandReader
 from ._pkginfo import PkgInfoReader
 from ._static import StaticReader
-from ._constants import FIELDS
 
 
 logger = getLogger('dephell_setuptools')
+ALL_READERS = (
+    StaticReader,
+    CfgReader,
+    CommandReader,
+    PkgInfoReader,
+)
 
 
-class ReadersManager:
-    error_handler = logger.exception
-    readers = (
-        StaticReader,
-        CfgReader,
-        CommandReader,
-        PkgInfoReader,
-    )
-
-    def __call__(self, path: Path):
-        result = dict()
-        for reader in self.readers:
-            try:
-                content = reader(path=path).content
-            except Exception as e:
-                self.error_handler(str(e))
-            else:
-                result.update(content)
-        exclude = (None, 0, [])
-        result = {k: v for k, v in result.items() if k in FIELDS and v not in exclude}
-        return result
+def read_setup(*,
+               path: Path,
+               error_handler: Callable[[Exception], Any] = logger.exception,
+               readers: Iterable = ALL_READERS):
+    result = dict()
+    for reader in readers:
+        try:
+            content = reader(path=path).content
+        except Exception as e:
+            error_handler(e)
+        else:
+            result.update(content)
+    return result

dephell_setuptools/_pkginfo.py

@@ -14,4 +14,5 @@ def content(self):
         result = subprocess.run(cmd, stderr=subprocess.PIPE, stdout=subprocess.PIPE)
         if result.returncode != 0:
             return None
-        return json.loads(result.stdout.decode())
+        content = json.loads(result.stdout.decode())
+        return self._clean(content)

dephell_setuptools/_static.py

@@ -10,7 +10,8 @@ class StaticReader(BaseReader):
     def content(self) -> Optional[Dict[str, Union[List, Dict]]]:
         if not self.call:
             return None
-        return self._get_call_kwargs(self.call)
+        result = self._get_call_kwargs(self.call)
+        return self._clean(result)
 
     @cached_property
     def tree(self) -> tuple: