initial tokenizer

This commit is contained in:
ari committed 2020-10-13 10:13:20 -04:00
1 parent fe9fd304ab
commit 460d213700
10 files changed
+868 -1

No files matched your search

+1 -1
View File
@@ -6,4 +6,4 @@ indent_size = 4
end_of_line = lf
charset = utf-8
trim_trailing_whitespace = false
insert_final_newline = false
insert_final_newline = true
+46
View File
@@ -0,0 +1,46 @@
import io
from typing import List
from pdfmajor.tokenizer.token import Token
from pdfmajor.tokenizer.exceptions import TokenizerEOF
from pdfmajor.tokenizer.parse_funcs import ParseFunc
class PSTokenizer:
def __init__(self, fp: io.BufferedIOBase) -> None:
self.fp = fp
self.seek(0)
def seek(self, pos: int):
"""Seeks tokenizer to the given position."""
self.fp.seek(pos)
# reset the status for nextline()
self.bufpos: int = pos
self.buf: bytes = b""
self.charpos: int = 0
# reset the status for nexttoken()
self._current_parse_func: ParseFunc = self._parse_main
self._curtoken: bytes = b""
self._curtokenpos: int = 0
self._tokens: List[Token] = []
return
def __repr__(self):
return "<%s: %r, bufpos=%d>" % (self.__class__.__name__, self.fp, self.bufpos)
def nexttoken(self):
while not self._tokens:
self.fillbuf()
self.charpos = self._current_parse_func(self.buf, self.charpos)
token = self._tokens.pop(0)
return token
def fillbuf(self):
if self.charpos < len(self.buf):
return
# fetch next chunk.
self.bufpos = self.fp.tell()
self.buf = self.fp.read(self.BUFSIZ)
if not self.buf:
raise TokenizerEOF("Unexpected EOF")
self.charpos = 0
return
+37
View File
@@ -0,0 +1,37 @@
import re
from .types import PSLiteral, PSSymbolTable, PSKeyword
PSLiteralTable = PSSymbolTable(PSLiteral)
PSKeywordTable = PSSymbolTable(PSKeyword)
LIT = PSLiteralTable.intern
KWD = PSKeywordTable.intern
KEYWORD_PROC_BEGIN = KWD(b"{")
KEYWORD_PROC_END = KWD(b"}")
KEYWORD_ARRAY_BEGIN = KWD(b"[")
KEYWORD_ARRAY_END = KWD(b"]")
KEYWORD_DICT_BEGIN = KWD(b"<<")
KEYWORD_DICT_END = KWD(b">>")
## PSBaseParser
##
EOL = re.compile(br"[\r\n]")
SPC = re.compile(br"\s")
NONSPC = re.compile(br"\S")
HEX = re.compile(br"[0-9a-fA-F]")
END_LITERAL = re.compile(br"[#/%\[\]()<>{}\s]")
END_HEX_STRING = re.compile(br"[^\s0-9a-fA-F]")
HEX_PAIR = re.compile(br"[0-9a-fA-F]{2}|.")
END_NUMBER = re.compile(br"[^0-9]")
END_KEYWORD = re.compile(br"[#/%\[\]()<>{}\s]")
END_STRING = re.compile(br"[()\134]")
OCT_STRING = re.compile(br"[0-7]")
ESC_STRING = {
b"b": 8,
b"t": 9,
b"n": 10,
b"f": 12,
b"r": 13,
b"(": 40,
b")": 41,
b"\\": 92,
}
+16
View File
@@ -0,0 +1,16 @@
from pdfmajor.execptions import PDFMajorException
class TokenizerError(PDFMajorException):
pass
class TokenizerEOF(TokenizerError):
pass
class InvalidToken(TokenizerError):
def __init__(self, pos: int, token: bytes) -> None:
super().__init__("Invalid Token Found at {pos}, of value={v}".format(
pos=pos,
v=token
))
+341
View File
@@ -0,0 +1,341 @@
from typing import List, Tuple, Union
from io import TextIOWrapper
from pdfmajor.utils import int2byte
from pdfmajor.execptions import PSEOF
from .types import PSLiteral
from .constants import OCT_STRING, ESC_STRING, KEYWORD_DICT_BEGIN, KEYWORD_DICT_END
from .constants import END_NUMBER, END_STRING, END_KEYWORD, END_HEX_STRING
from .constants import SPC, HEX_PAIR
from .constants import EOL, NONSPC, KWD, END_LITERAL, LIT, HEX
class PSTokenizer(object):
"""Tokenizatizes bytes stream to PS-Tokens"""
BUFSIZ = 4096
def __init__(self, fp: TextIOWrapper):
self.fp = fp
self.seek(0)
return
def __repr__(self):
return "<%s: %r, bufpos=%d>" % (self.__class__.__name__, self.fp, self.bufpos)
def flush(self):
return
def close(self):
self.flush()
return
def tell(self):
return self.bufpos + self.charpos
def poll(self, pos=None, n=80):
pos0 = self.fp.tell()
if not pos:
pos = self.bufpos + self.charpos
self.fp.seek(pos)
self.fp.read(n)
self.fp.seek(pos0)
return
def seek(self, pos: int):
"""Seeks the parser to the given position."""
self.fp.seek(pos)
# reset the status for nextline()
self.bufpos: int = pos
self.buf: bytes = b""
self.charpos: int = 0
# reset the status for nexttoken()
self._current_parse_func = self._parse_main
self._curtoken: bytes = b""
self._curtokenpos: int = 0
self._tokens: List[
Tuple[int, Union[int, float, bool, str, bytes, PSLiteral]]
] = []
return
def fillbuf(self):
if self.charpos < len(self.buf):
return
# fetch next chunk.
self.bufpos = self.fp.tell()
self.buf = self.fp.read(self.BUFSIZ)
if not self.buf:
raise PSEOF("Unexpected EOF")
self.charpos = 0
return
def nextline(self):
"""Fetches a next line that ends either with \\r or \\n."""
linebuf = b""
linepos = self.bufpos + self.charpos
eol = False
while 1:
self.fillbuf()
if eol:
c = self.buf[self.charpos : self.charpos + 1]
# handle b'\r\n'
if c == b"\n":
linebuf += c
self.charpos += 1
break
m = EOL.search(self.buf, self.charpos)
if m:
linebuf += self.buf[self.charpos : m.end(0)]
self.charpos = m.end(0)
if linebuf[-1:] == b"\r":
eol = True
else:
break
else:
linebuf += self.buf[self.charpos :]
self.charpos = len(self.buf)
return (linepos, linebuf)
def revreadlines(self):
"""Fetches a next line backword.
This is used to locate the trailers at the end of a file.
"""
self.fp.seek(0, 2)
pos = self.fp.tell()
buf = b""
while 0 < pos:
prevpos = pos
pos = max(0, pos - self.BUFSIZ)
self.fp.seek(pos)
s = self.fp.read(prevpos - pos)
if not s:
break
while 1:
n = max(s.rfind(b"\r"), s.rfind(b"\n"))
if n == -1:
buf = s + buf
break
yield s[n:] + buf
s = s[:n]
buf = b""
return
def _parse_main(self, s, i):
m = NONSPC.search(s, i)
if not m:
return len(s)
j = m.start(0)
c = s[j : j + 1]
self._curtokenpos = self.bufpos + j
if c == b"%":
self._curtoken = b"%"
self._current_parse_func = self._parse_comment
return j + 1
elif c == b"/":
self._curtoken = b""
self._current_parse_func = self._parse_literal
return j + 1
elif c in b"-+" or c.isdigit():
self._curtoken = c
self._current_parse_func = self._parse_number
return j + 1
elif c == b".":
self._curtoken = c
self._current_parse_func = self._parse_float
return j + 1
elif c.isalpha():
self._curtoken = c
self._current_parse_func = self._parse_keyword
return j + 1
elif c == b"(":
self._curtoken = b""
self.paren = 1
self._current_parse_func = self._parse_string
return j + 1
elif c == b"<":
self._curtoken = b""
self._current_parse_func = self._parse_wopen
return j + 1
elif c == b">":
self._curtoken = b""
self._current_parse_func = self._parse_wclose
return j + 1
else:
self._add_token(KWD(c))
return j + 1
def _add_token(self, obj):
self._tokens.append((self._curtokenpos, obj))
return
def _parse_comment(self, s, i):
m = EOL.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
self._current_parse_func = self._parse_main
# We ignore comments.
# self._tokens.append(self._curtoken)
return j
def _parse_literal(self, s, i):
m = END_LITERAL.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
c = s[j : j + 1]
if c == b"#":
self.hex = b""
self._current_parse_func = self._parse_literal_hex
return j + 1
try:
self._curtoken = str(self._curtoken, "utf-8")
except:
pass
self._add_token(LIT(self._curtoken))
self._current_parse_func = self._parse_main
return j
def _parse_literal_hex(self, s, i):
c = s[i : i + 1]
if HEX.match(c) and len(self.hex) < 2:
self.hex += c
return i + 1
if self.hex:
self._curtoken += int2byte(int(self.hex, 16))
self._current_parse_func = self._parse_literal
return i
def _parse_number(self, s, i):
m = END_NUMBER.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
c = s[j : j + 1]
if c == b".":
self._curtoken += c
self._current_parse_func = self._parse_float
return j + 1
try:
self._add_token(int(self._curtoken))
except ValueError:
pass
self._current_parse_func = self._parse_main
return j
def _parse_float(self, s, i):
m = END_NUMBER.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
try:
self._add_token(float(self._curtoken))
except ValueError:
pass
self._current_parse_func = self._parse_main
return j
def _parse_keyword(self, s, i):
m = END_KEYWORD.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
if self._curtoken == b"true":
token = True
elif self._curtoken == b"false":
token = False
else:
token = KWD(self._curtoken)
self._add_token(token)
self._current_parse_func = self._parse_main
return j
def _parse_string(self, s, i):
m = END_STRING.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
c = s[j : j + 1]
if c == b"\\":
self.oct = b""
self._current_parse_func = self._parse_string_1
return j + 1
if c == b"(":
self.paren += 1
self._curtoken += c
return j + 1
if c == b")":
self.paren -= 1
if self.paren: # WTF, they said balanced parens need no special treatment.
self._curtoken += c
return j + 1
self._add_token(self._curtoken)
self._current_parse_func = self._parse_main
return j + 1
def _parse_string_1(self, s, i):
c = s[i : i + 1]
if OCT_STRING.match(c) and len(self.oct) < 3:
self.oct += c
return i + 1
if self.oct:
self._curtoken += int2byte(int(self.oct, 8))
self._current_parse_func = self._parse_string
return i
if c in ESC_STRING:
self._curtoken += int2byte(ESC_STRING[c])
self._current_parse_func = self._parse_string
return i + 1
def _parse_wopen(self, s, i):
c = s[i : i + 1]
if c == b"<":
self._add_token(KEYWORD_DICT_BEGIN)
self._current_parse_func = self._parse_main
i += 1
else:
self._current_parse_func = self._parse_hexstring
return i
def _parse_wclose(self, s, i):
c = s[i : i + 1]
if c == b">":
self._add_token(KEYWORD_DICT_END)
i += 1
self._current_parse_func = self._parse_main
return i
def _parse_hexstring(self, s, i):
m = END_HEX_STRING.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
token = HEX_PAIR.sub(
lambda m: int2byte(int(m.group(0), 16)), SPC.sub(b"", self._curtoken)
)
self._add_token(token)
self._current_parse_func = self._parse_main
return j
def nexttoken(self):
while not self._tokens:
self.fillbuf()
self.charpos = self._current_parse_func(self.buf, self.charpos)
token = self._tokens.pop(0)
return token
+100
View File
@@ -0,0 +1,100 @@
from pdfmajor.tokenizer.token import Token, TokenKeyword
from typing import Callable, NamedTuple, Optional, Union
from .constants import EOL, NONSPC
ParseFunc = Callable[["ParseInput"], "NextParseAction"]
class NextParseAction(NamedTuple):
next_parse_action: ParseFunc
charpos: int
curtokenpos: int
curtoken: Union[Token, bytes] = b""
paren: Optional[int] = None
class ParseInput(NamedTuple):
bufvalue: bytes
charpos: int
bufpos: int
def parse_main(inp: ParseInput) -> NextParseAction:
m = NONSPC.search(inp.bufvalue, inp.charpos)
if not m:
return len(inp.bufvalue)
j = m.start(0)
c = inp.bufvalue[j : j + 1]
if c == b"%":
return NextParseAction(
curtoken=b"%",
next_parse_action=__parse_comment,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c == b"/":
return NextParseAction(
next_parse_action=__parse_literal,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c in b"-+" or c.isdigit():
return NextParseAction(
curtoken=c,
next_parse_action=__parse_number,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c == b".":
return NextParseAction(
curtoken=c,
next_parse_action=__parse_float,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c.isalpha():
return NextParseAction(
curtoken=c,
next_parse_action=__parse_keyword,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c == b"(":
return NextParseAction(
next_parse_action=__parse_string,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
paren=1,
)
elif c == b"<":
return NextParseAction(
next_parse_action=__parse_wopen,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
elif c == b">":
return NextParseAction(
next_parse_action=__parse_wclose,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
)
else:
return NextParseAction(
next_parse_action=parse_main,
curtokenpos=inp.bufpos + j,
charpos=j + 1,
curtoken=TokenKeyword(inp.bufpos + j, c),
)
def __parse_comment(s, i):
m = EOL.search(s, i)
if not m:
self._curtoken += s[i:]
return len(s)
j = m.start(0)
self._curtoken += s[i:j]
self._current_parse_func = self._parse_main
# We ignore comments.
# self._tokens.append(self._curtoken)
return j
+44
View File
@@ -0,0 +1,44 @@
from typing import Generic, TypeVar, Union
from decimal import Decimal
T = TypeVar("T")
class Token(Generic[T]):
__slots__ = ["pos", "size", "value"]
def __init__(self, pos: int, size: int, value: T) -> None:
self.size = size
self.pos = pos
self.value = value
def __repr__(self) -> str:
return "<{cls} pos={pos}>{value}</{cls}>".format(
cls=self.__class__.__name__, pos=self.pos, value=self.value
)
class TokenComment(Token[bytes]):
pass
class TokenHex(Token[bytes]):
pass
class TokenKeyword(Token[bytes]):
pass
class TokenString(Token[str]):
pass
class TokenInteger(Token[int]):
pass
class TokenDecimal(Token[Decimal]):
pass
TokenNumber = Union[TokenInteger, TokenDecimal]
+139
View File
@@ -0,0 +1,139 @@
from decimal import Decimal
from pdfmajor.tokenizer.exceptions import InvalidToken
from pdfmajor.tokenizer.token import Token, TokenComment, TokenDecimal, TokenNumber, TokenInteger
from pdfmajor.tokenizer.constants import END_NUMBER, EOL
from typing import Callable, Iterator, NamedTuple
class PInput(NamedTuple):
pos: int
buf: bytes
ParserFunc = Callable[[int, Iterator[PInput]], Token]
def cmp_tsize(curpos: int, initialpos: int, tokenend_idx: int) -> int:
"""Computes the size of the token based on its positional data
Args:
curpos (int): current position of the buffer
initialpos (int): initial position where the parser started working
tokenend_idx (int): last relevant character index within the last buffered value
Returns:
int: length of index
"""
return curpos-initialpos+tokenend_idx
def parse_comment(initialpos: int, inp: Iterator[PInput]) -> TokenComment:
"""Parses input stream into a comment
Args:
initialpos (int): initial position where we started
inp (Iterator[PInput])
Returns:
TokenComment
"""
curtoken = b""
for curpos, s in inp:
m = EOL.search(s, 0)
if not m:
curtoken += s
else:
j = m.start(0)
curtoken += s[:j]
return TokenComment(
initialpos,
cmp_tsize(curpos, initialpos, j),
curtoken
)
def parse_number(initialpos: int, inp: Iterator[PInput]) -> TokenNumber:
"""Parses input stream into a comment
Args:
initialpos (int): initial position where we started
inp (Iterator[PInput])
Returns:
TokenComment
"""
curtoken = b""
is_dec = False
for curpos, s in inp:
m = END_NUMBER.search(s, 0)
if not m:
curtoken += s
else:
j = m.start(0)
curtoken += s[:j]
c = s[j : j + 1]
if (c == b".") and not is_dec:
is_dec = True
curtoken += c
s = s[j+1:]
m = END_NUMBER.search(s, 0)
if not m:
curtoken += s
continue
else:
nj = m.start(0)
curtoken += s[:nj]
j += nj
return TokenDecimal(
initialpos,
cmp_tsize(curpos, initialpos, j),
Decimal(curtoken.decode())
)
try:
if is_dec:
return TokenDecimal(
initialpos,
cmp_tsize(curpos, initialpos, j),
Decimal(curtoken.decode())
)
else:
return TokenInteger(
initialpos,
cmp_tsize(curpos, initialpos, j),
int(curtoken)
)
except ValueError:
raise InvalidToken(initialpos, curtoken)
# def parse_float(
# initialpos: int,
# inp: Iterator[PInput],
# initialtokenval = b""
# ) -> TokenFloat:
# curtoken = b"" + initialtokenval
# for curpos, s in inp:
# print('->',curpos, s)
# m = END_NUMBER.search(s, 0)
# if not m:
# curtoken += s
# else:
# j = m.start(0)
# curtoken += s[:j]
# next_t = s[j:j+1]
# if next_t == b'.':
# curtoken += b'.'
# s = s[j+1:]
# m = END_NUMBER.search(s, 0)
# if not m:
# curtoken += s
# continue
# else:
# j = m.start(0)
# curtoken += s[:j]
# next_t = s[j+1]
# try:
# print(curpos,j, initialpos)
# return TokenFloat(
# initialpos,
# cmp_tsize(curpos, initialpos, j),
# Decimal(curtoken.decode())
# )
# except ValueError:
# raise InvalidToken(initialpos, curtoken)
+80
View File
@@ -0,0 +1,80 @@
## Basic PostScript Types
##
from typing import Union
## PSObject
##
class PSObject(object):
"""Base class for all PS or PDF-related data types."""
pass
## PSLiteral
##
class PSLiteral(PSObject):
"""A class that represents a PostScript literal.
Postscript literals are used as identifiers, such as
variable names, property names and dictionary keys.
Literals are case sensitive and denoted by a preceding
slash sign (e.g. "/Name")
Note: Do not create an instance of PSLiteral directly.
Always use PSLiteralTable.intern().
"""
def __init__(self, name):
self.name = name
def __repr__(self):
name = self.name
return "/%r" % name
## PSKeyword
##
class PSKeyword(PSObject):
"""A class that represents a PostScript keyword.
PostScript keywords are a dozen of predefined words.
Commands and directives in PostScript are expressed by keywords.
They are also used to denote the content boundaries.
Note: Do not create an instance of PSKeyword directly.
Always use PSKeywordTable.intern().
"""
def __init__(self, name):
self.name = name
return
def __repr__(self):
name = self.name
return "/%r" % name
## PSSymbolTable
##
class PSSymbolTable(object):
"""A utility class for storing PSLiteral/PSKeyword objects.
Interned objects can be checked its identity with "is" operator.
"""
def __init__(self, klass):
self.dict = {}
self.klass = klass
return
def intern(self, name):
if name in self.dict:
lit = self.dict[name]
else:
lit = self.klass(name)
self.dict[name] = lit
return lit
+64
View File
@@ -0,0 +1,64 @@
from decimal import Decimal
import io
from pdfmajor.tokenizer.token import TokenComment, TokenDecimal, TokenInteger
from pdfmajor.tokenizer.token_parser import PInput, parse_comment, parse_number
from unittest import TestCase
def make_stream_iter(data: bytes, initpos: int = 0):
stream = io.BytesIO(data)
def iter_stream(buf_size: int):
while True:
bufpos = stream.tell()
buf = stream.read(buf_size)
if buf:
yield PInput(initpos+bufpos, buf)
else:
raise "EOF"
return iter_stream
class Basics(TestCase):
def test_comment_parser_ln(self):
iter_stream = make_stream_iter(
b"this is a lengthy comment that ends here\nso this is not reachable"
)
token = parse_comment(0, iter_stream(3))
self.assertIsInstance(token, TokenComment)
self.assertEqual(token.pos, 0)
self.assertEqual(token.value, b"this is a lengthy comment that ends here")
self.assertEqual(token.size, 40)
def test_comment_parser_lr(self):
iter_stream = make_stream_iter(
b"this is a lengthy comment that ends here\rso this is not reachable"
)
token = parse_comment(0, iter_stream(7))
self.assertIsInstance(token, TokenComment)
self.assertEqual(token.pos, 0)
self.assertEqual(token.value, b"this is a lengthy comment that ends here")
self.assertEqual(token.size, 40)
def test_parse_float(self):
iter_stream = make_stream_iter(
b"0.120 + some invalid text",
5
)
token = parse_number(5, iter_stream(3))
self.assertIsInstance(token, TokenDecimal)
self.assertEqual(token.pos, 5)
self.assertEqual(token.value, Decimal('0.12'))
self.assertEqual(token.size, 5)
def test_parse_int(self):
iter_stream = make_stream_iter(
b"130 oh oh 0.12 + some invalid text",
11
)
token = parse_number(11, iter_stream(10))
self.assertIsInstance(token, TokenInteger)
self.assertEqual(token.pos, 11)
self.assertEqual(token.value, 130)
self.assertEqual(token.size, 3)