jpt-archive/saucebrush
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

drop Python 2 stuff, including Unicode/StringFilter

Browse Source
This commit is contained in:
James Turk 2022-11-10 21:29:31 -06:00
parent a7e3fc63b3
commit 230039fa90
2 changed files with 6 additions and 59 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
saucebrush/filters.py
View File

@ -221,7 +221,7 @@ class FieldModifier(FieldFilter):
def process_field(self, item): def process_field(self, item):
return self._filter_func(item) return self._filter_func(item)
def __unicode__(self): def __str__(self):
return "%s( %s, %s )" % ( return "%s( %s, %s )" % (
self.__class__.__name__, self.__class__.__name__,
str(self._target_keys), str(self._target_keys),
@ -263,7 +263,7 @@ class FieldRemover(Filter):
record.pop(key, None) record.pop(key, None)
return record return record
def __unicode__(self): def __str__(self):
return "%s( %s )" % (self.__class__.__name__, str(self._target_keys)) return "%s( %s )" % (self.__class__.__name__, str(self._target_keys))
@ -291,7 +291,7 @@ class FieldMerger(Filter):
record[to_col] = self._merge_func(*values) record[to_col] = self._merge_func(*values)
return record return record
def __unicode__(self): def __str__(self):
return "%s( %s, %s )" % ( return "%s( %s, %s )" % (
self.__class__.__name__, self.__class__.__name__,
str(self._field_mapping), str(self._field_mapping),
@ -517,36 +517,6 @@ class UniqueIDValidator(UniqueID):
validator = True validator = True
class UnicodeFilter(Filter):
"""Convert all str elements in the record to Unicode."""
def __init__(self, encoding="utf-8", errors="ignore"):
super(UnicodeFilter, self).__init__()
self._encoding = encoding
self._errors = errors
def process_record(self, record):
for key, value in record.items():
if isinstance(value, str):
record[key] = unicode(value, self._encoding, self._errors)
elif isinstance(value, unicode):
record[key] = value.decode(self._encoding, self._errors)
return record
class StringFilter(Filter):
def __init__(self, encoding="utf-8", errors="ignore"):
super(StringFilter, self).__init__()
self._encoding = encoding
self._errors = errors
def process_record(self, record):
for key, value in record.items():
if isinstance(value, unicode):
record[key] = value.encode(self._encoding, self._errors)
return record
########################### ###########################
## Commonly Used Filters ## ## Commonly Used Filters ##
########################### ###########################
@ -601,7 +571,7 @@ class NameCleaner(Filter):
# first middle? last suffix? # first middle? last suffix?
FIRST_LAST = re.compile( FIRST_LAST = re.compile(
"""^\s*(?:(?P<firstname>\w+)(?:\.?) r"""^\s*(?:(?P<firstname>\w+)(?:\.?)
\s+(?:(?P<middlename>\w+)\.?\s+)? \s+(?:(?P<middlename>\w+)\.?\s+)?
(?P<lastname>[A-Za-z'-]+)) (?P<lastname>[A-Za-z'-]+))
(?:\s+(?P<suffix>JR\.?|II|III|IV))? (?:\s+(?P<suffix>JR\.?|II|III|IV))?
@ -611,7 +581,7 @@ class NameCleaner(Filter):
# last, first middle? suffix? # last, first middle? suffix?
LAST_FIRST = re.compile( LAST_FIRST = re.compile(
"""^\s*(?:(?P<lastname>[A-Za-z'-]+), r"""^\s*(?:(?P<lastname>[A-Za-z'-]+),
\s+(?P<firstname>\w+)(?:\.?) \s+(?P<firstname>\w+)(?:\.?)
(?:\s+(?P<middlename>\w+)\.?)?) (?:\s+(?P<middlename>\w+)\.?)?)
(?:\s+(?P<suffix>JR\.?|II|III|IV))? (?:\s+(?P<suffix>JR\.?|II|III|IV))?

25
saucebrush/utils.py
View File

@ -1,10 +1,5 @@
import collections
import os import os
from urllib.request import urlopen
try:
from urllib.request import urlopen # attemp py3 first
except ImportError:
from urllib2 import urlopen # fallback to py2
""" """
General utilities used within saucebrush that may be useful elsewhere. General utilities used within saucebrush that may be useful elsewhere.
@ -72,24 +67,6 @@ def str_or_list(obj):
# #
class FallbackCounter(collections.defaultdict):
"""Python 2.6 does not have collections.Counter.
This is class that does the basics of what we need from Counter.
"""
def __init__(self, *args, **kwargs):
super(FallbackCounter, self).__init__(int)
def most_common(n=None):
l = sorted(self.items(), cmp=lambda x, y: cmp(x[1], y[1]))
if n is not None:
l = l[:n]
return l
class Files(object): class Files(object):
"""Iterate over multiple files as a single file. Pass the paths of the """Iterate over multiple files as a single file. Pass the paths of the
files as arguments to the class constructor: files as arguments to the class constructor: