Fix parsing of author information

Instead of relying on regular expressions, this patch leverages Python’s
builtin `email.utils.parseaddr()` functionality to parse an RFC-822-compliant
email address string into its name and address parts.

This should also resolve issues with special characters in the name part; see
issues python-poetry#370 and python-poetry#798.

python-poetry#370
python-poetry#798

poetry/console/commands/init.py

@@ -6,6 +6,7 @@
 from typing import List
 from typing import Tuple
 
+from poetry.utils.helpers import parse_author
 from .command import Command
 from .env_command import EnvCommand
 
@@ -294,15 +295,13 @@ def _format_requirements(self, requirements):  # type: (List[str]) -> dict
         return requires
 
     def _validate_author(self, author, default):
-        from poetry.packages.package import AUTHOR_REGEX
-
         author = author or default
 
         if author in ["n", "no"]:
             return
 
-        m = AUTHOR_REGEX.match(author)
-        if not m:
+        name, email = parse_author(author)
+        if not name:
             raise ValueError(
                 "Invalid author string. Must be in the format: "
                 "John Smith <[email protected]>"

poetry/masonry/builders/builder.py

@@ -1,6 +1,5 @@
 # -*- coding: utf-8 -*-
 import os
-import re
 import shutil
 import tempfile
 
@@ -16,15 +15,14 @@
 from poetry.utils._compat import glob
 from poetry.utils._compat import lru_cache
 from poetry.utils._compat import to_str
+from poetry.utils.helpers import parse_author
 from poetry.vcs import get_vcs
 
 from ..metadata import Metadata
 from ..utils.module import Module
 from ..utils.package_include import PackageInclude
 
 
-AUTHOR_REGEX = re.compile(r"(?u)^(?P<name>[- .,\w\d'’\"()]+) <(?P<email>.+?)>$")
-
 METADATA_BASE = """\
 Metadata-Version: 2.1
 Name: {name}
@@ -228,10 +226,7 @@ def convert_entry_points(self):  # type: () -> dict
 
     @classmethod
     def convert_author(cls, author):  # type: () -> dict
-        m = AUTHOR_REGEX.match(author)
-
-        name = m.group("name")
-        email = m.group("email")
+        name, email = parse_author(author)
 
         return {"name": name, "email": email}
 

poetry/packages/package.py

@@ -1,6 +1,5 @@
 # -*- coding: utf-8 -*-
 import copy
-import re
 
 from contextlib import contextmanager
 from typing import Union
@@ -10,7 +9,7 @@
 from poetry.spdx import license_by_id
 from poetry.spdx import License
 from poetry.utils._compat import Path
-from poetry.utils.helpers import canonicalize_name
+from poetry.utils.helpers import canonicalize_name, parse_author
 from poetry.version.markers import AnyMarker
 from poetry.version.markers import parse_marker
 
@@ -22,8 +21,6 @@
 from .utils.utils import convert_markers
 from .utils.utils import create_nested_marker
 
-AUTHOR_REGEX = re.compile(r"(?u)^(?P<name>[- .,\w\d'’\"()]+)(?: <(?P<email>.+?)>)?$")
-
 
 class Package(object):
 
@@ -141,10 +138,7 @@ def _get_author(self):  # type: () -> dict
         if not self._authors:
             return {"name": None, "email": None}
 
-        m = AUTHOR_REGEX.match(self._authors[0])
-
-        name = m.group("name")
-        email = m.group("email")
+        name, email = parse_author(self._authors[0])
 
         return {"name": name, "email": email}
 

poetry/utils/helpers.py

@@ -3,6 +3,7 @@
 import shutil
 import stat
 import tempfile
+from email.utils import parseaddr
 
 from contextlib import contextmanager
 from typing import List
@@ -100,3 +101,21 @@ def _on_rm_error(func, path, exc_info):
 
 def safe_rmtree(path):
     shutil.rmtree(path, onerror=_on_rm_error)
+
+
+def parse_author(address):  # type: (str) -> tuple
+    """Parse name and address parts from an email address string.
+
+    .. note::
+
+       If the input string does not contain an `@` character, it is
+       assumed that it represents only a name without an email address.
+
+    :param address: the email address string to parse.
+    :return: a 2-tuple with the parsed name and email address.  If a
+             part is missing, ``None`` will be returned in its place.
+    """
+    if "@" not in address:
+        return (address, None)
+    name, email = parseaddr(address)
+    return (name or None, email or None)

tests/packages/test_package.py

@@ -14,3 +14,11 @@ def test_package_authors():
     package.authors.insert(0, "John Doe")
     assert package.author_name == "John Doe"
     assert package.author_email is None
+
+
+def test_package_authors_with_fancy_unicode():
+    package = Package("foo", "0.1.0")
+
+    package.authors.append("my·fancy·company <[email protected]>")
+    assert package.author_name == "my·fancy·company"
+    assert package.author_email == "[email protected]"

tests/utils/test_helpers.py

@@ -1,5 +1,9 @@
+# -*- coding: utf-8 -*-
+from __future__ import unicode_literals
+
 from poetry.utils.helpers import get_http_basic_auth
 from poetry.utils.helpers import parse_requires
+from poetry.utils.helpers import parse_author
 
 
 def test_parse_requires():
@@ -66,3 +70,51 @@ def test_get_http_basic_auth_without_password(config):
 
 def test_get_http_basic_auth_missing(config):
     assert get_http_basic_auth(config, "foo") is None
+
+
+def test_parse_author_simple_name_and_email():
+    name, email = parse_author("John Doe <[email protected]>")
+    assert name == "John Doe"
+    assert email == "[email protected]"
+
+
+def test_parse_author_simple_name_only():
+    name, email = parse_author("John Doe")
+    assert name == "John Doe"
+    assert email is None
+
+
+def test_parse_author_ascii_specialchars_name_and_email():
+    name, email = parse_author("R&D <[email protected]>")
+    assert name == "R&D"
+    assert email == "[email protected]"
+
+
+def test_parse_author_ascii_specialchars_name_only():
+    name, email = parse_author("R&D")
+    assert name == "R&D"
+    assert email is None
+
+
+def test_parse_author_unicode_name_and_email():
+    name, email = parse_author("my·fancy·corp <[email protected]>")
+    assert name == "my·fancy·corp"
+    assert email == "[email protected]"
+
+
+def test_parse_author_unicode_name_only():
+    name, email = parse_author("my·fancy·corp")
+    assert name == "my·fancy·corp"
+    assert email is None
+
+
+def test_parse_author_email_only_with_angular_brackets():
+    name, email = parse_author("<[email protected]>")
+    assert name is None
+    assert email == "[email protected]"
+
+
+def test_parse_author_email_only_without_angular_brackets():
+    name, email = parse_author("[email protected]")
+    assert name is None
+    assert email == "[email protected]"