def parse_file(file,
report_filename=None,
report_firstline=1,
encoding=None,
**kw):
"""Parse a beancount input file and return Ledger with the list of
transactions and tree of accounts.
Args:
file: file object or path to the file to be parsed.
kw: a dict of keywords to be applied to the C parser.
Returns:
A tuple of (
list of entries parsed in the file,
list of errors that were encountered during parsing, and
a dict of the option values that were parsed from the file.)
"""
if encoding is not None and codecs.lookup(encoding).name != 'utf-8':
raise ValueError('Only UTF-8 encoded files are supported.')
with contextlib.ExitStack() as ctx:
if file == '-':
file = sys.stdin.buffer
# It would be more appropriate here to check for io.RawIOBase but
# that does not work for io.BytesIO despite it implementing the
# readinto() method.
elif not isinstance(file, io.IOBase):
file = ctx.enter_context(open(file, 'rb'))
builder = grammar.Builder()
parser = _parser.Parser(builder)
parser.parse(file,
filename=report_filename,
lineno=report_firstline,
**kw)
return builder.finalize()
def test_parser_parse(self):
# Do not use a string to avoid issues due to string interning.
name = object()
self.assertEqual(sys.getrefcount(name), 2)
f = io.BytesIO(b"")
f.name = name
self.assertEqual(sys.getrefcount(f.name), 3)
builder = grammar.Builder()
parser = _parser.Parser(builder)
parser.parse(f)
# The Parser object keeps a reference to the input file.
self.assertEqual(sys.getrefcount(f), 3)
# There are references to the file name from the Parser object
# and from the the parsing results. In the case of an empty
# input file from the options dictionary stored in the builder.
self.assertEqual(sys.getrefcount(name), 5)
builder.options = {}
self.assertEqual(sys.getrefcount(name), 4)
del parser
# Once the Parser object is gone we should have just the local
# reference to the file object and two references to name.
self.assertEqual(sys.getrefcount(name), 3)
self.assertEqual(sys.getrefcount(f), 2)
def parse_file(file, report_filename=None, report_firstline=1, **kw):
"""Parse a beancount input file and return Ledger with the list of
transactions and tree of accounts.
Args:
file: file object or path to the file to be parsed.
kw: a dict of keywords to be applied to the C parser.
Returns:
A tuple of (
list of entries parsed in the file,
list of errors that were encountered during parsing, and
a dict of the option values that were parsed from the file.)
"""
close_file = None
if file == '-':
close_file = file = sys.stdin.buffer
# It would be more appropriate here to check for io.RawIOBase but
# that does not work for io.BytesIO despite it implementing the
# readinto() method.
elif not isinstance(file, io.IOBase):
close_file = file = open(file, 'rb')
builder = grammar.Builder()
parser = _parser.Parser(builder)
parser.parse(file, filename=report_filename, lineno=report_firstline, **kw)
if close_file:
close_file.close()
return builder.finalize()
def test_parse_string(self):
# TODO(blais): Remove, this is temporary, for testing locally.
filename = os.getenv("L")
assert filename
builder = grammar.Builder()
out = extmodule.parse(builder, filename)
print(out)
def parse_string(string, **kw):
"""Parse a beancount input file and return Ledger with the list of
transactions and tree of accounts.
Args:
string: a str, the contents to be parsed instead of a file's.
**kw: See parse.c. This function parses out 'dedent' which removes
whitespace from the front of the text (default is False).
Return:
Same as the output of parse_file().
"""
if kw.pop('dedent', None):
string = textwrap.dedent(string)
builder = grammar.Builder(None)
_parser.parse_string(string, builder, **kw)
builder.options['filename'] = '<string>'
return builder.finalize()
def parse_file(filename, **kw):
"""Parse a beancount input file and return Ledger with the list of
transactions and tree of accounts.
Args:
filename: the name of the file to be parsed.
kw: a dict of keywords to be applied to the C parser.
Returns:
A tuple of (
list of entries parsed in the file,
list of errors that were encountered during parsing, and
a dict of the option values that were parsed from the file.)
"""
abs_filename = path.abspath(filename) if filename else None
builder = grammar.Builder(abs_filename)
_parser.parse_file(filename, builder, **kw)
return builder.finalize()
def parse_string(string, report_filename=None, **kw):
"""Parse a beancount input file and return Ledger with the list of
transactions and tree of accounts.
Args:
string: A string, the contents to be parsed instead of a file's.
report_filename: A string, the source filename from which this string
has been extracted, if any. This is stored in the metadata of the
parsed entries.
**kw: See parse.c. This function parses out 'dedent' which removes
whitespace from the front of the text (default is False).
Return:
Same as the output of parse_file().
"""
if kw.pop('dedent', None):
string = textwrap.dedent(string)
builder = grammar.Builder(report_filename or '<string>')
_parser.parse_string(string, builder, report_filename=report_filename, **kw)
return builder.finalize()
def test_parse_lineno(self):
f = io.BytesIO(b"2020-07-30 open Assets:Test")
builder = grammar.Builder()
parser = _parser.Parser(builder)
parser.parse(f, lineno=42)
self.assertEqual(builder.entries[0].meta['lineno'], 42)
