NPO-Accounting/conservancy_beancount Changeset - 499f18ff623c

Changeset - 499f18ff623c

Parent rev.

Child rev.

[Not reviewed]

0 3 0

Brett Smith - 4 years ago 2020-04-01 17:38:37
brettcsmith@brettcsmith.org

meta_entity: Adjust what entities are allowed based on today's books.

See the comments throughout for more discussion about what cases are
or aren't allowed, and why.

3 files changed with 52 insertions and 3 deletions:

conservancy_beancount/plugin/meta_entity.py

setup.py

tests/test_meta_entity.py

0 comments (0 inline, 0 general)

conservancy_beancount/plugin/meta_entity.py

➞

Show inline comments

@@ ... / @@ -14,7 +14,12 @@ @@
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
 import re
 # Type stubs aren't available for regex.
 # Fortunately, we're using it in a way that's API-compatible with the re
 # module. We mitigate the lack of type stubs by providing type declarations
 # for returned objects. This way, the only thing that isn't type checked are
 # the calls to regex functions.
 import regex  # type:ignore[import]
 from . import core
 from .. import data
@@ ... / @@ -23,10 +28,25 @@ from ..beancount_types import ( @@
     Transaction,
+)
 from typing import (
     Pattern,
+)
 class MetaEntity(core.TransactionHook):
     METADATA_KEY = 'entity'
     HOOK_GROUPS = frozenset(['posting', 'metadata', METADATA_KEY])
     ENTITY_RE = re.compile(r'^[A-Za-z0-9]+(?:-[A-Za-z0-9]+)*$')
     # alnum is the set of characters we always accept in entity metadata:
     # letters and digits, minus the Latin 1 supplement (i.e., Roman letters
     # with diacritics: áÁàÀâÂåÅäÄãÃ çÇ ðÐ ñÑ øØ ß etc.)
     # See the tests for specific cases.
     alnum = r'\p{Letter}\p{Digit}--\p{Block=Latin_1_Supplement}'
     # A regexp that would be reasonably stricter would be:
     #   f'^[{alnum}][.{alnum}]*(?:-[.{alnum}])*$'
     # However, current producers fail that regexp in a few different ways.
     # See the tests for specific cases.
     ENTITY_RE: Pattern[str] = regex.compile(f'^[{alnum}][-.{alnum}]*$', regex.VERSION1)
     del alnum
     def run(self, txn: Transaction) -> errormod.Iter:
         txn_entity = txn.meta.get(self.METADATA_KEY)

setup.py

➞

Show inline comments

@@ ... / @@ -12,6 +12,7 @@ setup( @@
     install_requires=[
         'beancount>=2.2',
         'regex',
         'rt>=2.0',
     ],
     setup_requires=[

tests/test_meta_entity.py

➞

Show inline comments

@@ ... / @@ -21,19 +21,47 @@ from . import testutil @@
 from conservancy_beancount.plugin import meta_entity
 VALID_VALUES = {
     # Classic entity: LastName-FirstName
     'Smith-Alex',
     # Various people and companies have one-word names
     # Digits are allowed, as part of a name or standalone
     'Company19',
     'Company-19',
     # No case requirements
     'boyd-danah',
     # No limit on the number of parts of the name
     'B-van-der-A',
     # Names that have no ASCII are allowed, with or without dash separators
     '田中流星',
     '田中-流星',
     'スミスダコタ',
     'スミス-ダコタ',
     'Яшин-Данила',
     # The PayPal importer produces . in entity metadata
     'Du-Bois-W.-E.-B.',
     # import2ledger produces entities that end with -
     # That's probably a bug, but allow it for now.
     'foo-',
+}
 INVALID_VALUES = {
     # Starting with a - is not allowed
     '-foo',
     'foo-',
     '-',
     # Names that can be reduced to ASCII should be
     # Producers should change this to Uberentity or Ueberentity
     # I am not wild about this rule and would like to relax it—it's mostly
     # based on an expectation that entities are typed in by a human. That's true
     # less and less and it seems like we should reduce the amount of mangling
     # producers are expected to do. But it's the rule for today.
     'Überentity',
     # Whitespace is never allowed
     'Alex Smith',
     '田中 流星',
     'スミス ダコタ',
     'Яшин Данила',
     ' ',
     # An empty string is not valid
     '',
+}

0 comments (0 inline, 0 general)