| 1 |
#!/usr/bin/env python3.8 |
| 2 |
|
| 3 |
"""pegen -- PEG Generator. |
| 4 |
|
| 5 |
Search the web for PEG Parsers for reference. |
| 6 |
""" |
| 7 |
|
| 8 |
import argparse |
| 9 |
import sys |
| 10 |
import time |
| 11 |
import token |
| 12 |
import traceback |
| 13 |
|
| 14 |
from pegen.grammar import Grammar |
| 15 |
from pegen.parser import Parser |
| 16 |
from pegen.parser_generator import ParserGenerator |
| 17 |
from pegen.tokenizer import Tokenizer |
| 18 |
from pegen.validator import validate_grammar |
| 19 |
|
| 20 |
|
| 21 |
def generate_c_code( |
| 22 |
args: argparse.Namespace, |
| 23 |
) -> tuple[Grammar, Parser, Tokenizer, ParserGenerator]: |
| 24 |
from pegen.build import build_c_parser_and_generator |
| 25 |
|
| 26 |
verbose = args.verbose |
| 27 |
verbose_tokenizer = verbose >= 3 |
| 28 |
verbose_parser = verbose == 2 or verbose >= 4 |
| 29 |
try: |
| 30 |
grammar, parser, tokenizer, gen = build_c_parser_and_generator( |
| 31 |
args.grammar_filename, |
| 32 |
args.tokens_filename, |
| 33 |
args.output, |
| 34 |
args.compile_extension, |
| 35 |
verbose_tokenizer, |
| 36 |
verbose_parser, |
| 37 |
args.verbose, |
| 38 |
keep_asserts_in_extension=False if args.optimized else True, |
| 39 |
skip_actions=args.skip_actions, |
| 40 |
) |
| 41 |
return grammar, parser, tokenizer, gen |
| 42 |
except Exception as err: |
| 43 |
if args.verbose: |
| 44 |
raise # Show traceback |
| 45 |
traceback.print_exception(err.__class__, err, None) |
| 46 |
sys.stderr.write("For full traceback, use -v\n") |
| 47 |
sys.exit(1) |
| 48 |
|
| 49 |
|
| 50 |
def generate_python_code( |
| 51 |
args: argparse.Namespace, |
| 52 |
) -> tuple[Grammar, Parser, Tokenizer, ParserGenerator]: |
| 53 |
from pegen.build import build_python_parser_and_generator |
| 54 |
|
| 55 |
verbose = args.verbose |
| 56 |
verbose_tokenizer = verbose >= 3 |
| 57 |
verbose_parser = verbose == 2 or verbose >= 4 |
| 58 |
try: |
| 59 |
grammar, parser, tokenizer, gen = build_python_parser_and_generator( |
| 60 |
args.grammar_filename, |
| 61 |
args.output, |
| 62 |
verbose_tokenizer, |
| 63 |
verbose_parser, |
| 64 |
skip_actions=args.skip_actions, |
| 65 |
) |
| 66 |
return grammar, parser, tokenizer, gen |
| 67 |
except Exception as err: |
| 68 |
if args.verbose: |
| 69 |
raise # Show traceback |
| 70 |
traceback.print_exception(err.__class__, err, None) |
| 71 |
sys.stderr.write("For full traceback, use -v\n") |
| 72 |
sys.exit(1) |
| 73 |
|
| 74 |
|
| 75 |
argparser = argparse.ArgumentParser( |
| 76 |
prog="pegen", description="Experimental PEG-like parser generator" |
| 77 |
) |
| 78 |
argparser.add_argument("-q", "--quiet", action="store_true", help="Don't print the parsed grammar") |
| 79 |
argparser.add_argument( |
| 80 |
"-v", |
| 81 |
"--verbose", |
| 82 |
action="count", |
| 83 |
default=0, |
| 84 |
help="Print timing stats; repeat for more debug output", |
| 85 |
) |
| 86 |
subparsers = argparser.add_subparsers(help="target language for the generated code") |
| 87 |
|
| 88 |
c_parser = subparsers.add_parser("c", help="Generate C code for inclusion into CPython") |
| 89 |
c_parser.set_defaults(func=generate_c_code) |
| 90 |
c_parser.add_argument("grammar_filename", help="Grammar description") |
| 91 |
c_parser.add_argument("tokens_filename", help="Tokens description") |
| 92 |
c_parser.add_argument( |
| 93 |
"-o", "--output", metavar="OUT", default="parse.c", help="Where to write the generated parser" |
| 94 |
) |
| 95 |
c_parser.add_argument( |
| 96 |
"--compile-extension", |
| 97 |
action="store_true", |
| 98 |
help="Compile generated C code into an extension module", |
| 99 |
) |
| 100 |
c_parser.add_argument( |
| 101 |
"--optimized", action="store_true", help="Compile the extension in optimized mode" |
| 102 |
) |
| 103 |
c_parser.add_argument( |
| 104 |
"--skip-actions", |
| 105 |
action="store_true", |
| 106 |
help="Suppress code emission for rule actions", |
| 107 |
) |
| 108 |
|
| 109 |
python_parser = subparsers.add_parser( |
| 110 |
"python", |
| 111 |
help="Generate Python code, needs grammar definition with Python actions", |
| 112 |
) |
| 113 |
python_parser.set_defaults(func=generate_python_code) |
| 114 |
python_parser.add_argument("grammar_filename", help="Grammar description") |
| 115 |
python_parser.add_argument( |
| 116 |
"-o", |
| 117 |
"--output", |
| 118 |
metavar="OUT", |
| 119 |
default="parse.py", |
| 120 |
help="Where to write the generated parser", |
| 121 |
) |
| 122 |
python_parser.add_argument( |
| 123 |
"--skip-actions", |
| 124 |
action="store_true", |
| 125 |
help="Suppress code emission for rule actions", |
| 126 |
) |
| 127 |
|
| 128 |
|
| 129 |
def main() -> None: |
| 130 |
from pegen.testutil import print_memstats |
| 131 |
|
| 132 |
args = argparser.parse_args() |
| 133 |
if "func" not in args: |
| 134 |
argparser.error("Must specify the target language mode ('c' or 'python')") |
| 135 |
|
| 136 |
t0 = time.time() |
| 137 |
grammar, parser, tokenizer, gen = args.func(args) |
| 138 |
t1 = time.time() |
| 139 |
|
| 140 |
validate_grammar(grammar) |
| 141 |
|
| 142 |
if not args.quiet: |
| 143 |
if args.verbose: |
| 144 |
print("Raw Grammar:") |
| 145 |
for line in repr(grammar).splitlines(): |
| 146 |
print(" ", line) |
| 147 |
|
| 148 |
print("Clean Grammar:") |
| 149 |
for line in str(grammar).splitlines(): |
| 150 |
print(" ", line) |
| 151 |
|
| 152 |
if args.verbose: |
| 153 |
print("First Graph:") |
| 154 |
for src, dsts in gen.first_graph.items(): |
| 155 |
print(f" {src} -> {', '.join(dsts)}") |
| 156 |
print("First SCCS:") |
| 157 |
for scc in gen.first_sccs: |
| 158 |
print(" ", scc, end="") |
| 159 |
if len(scc) > 1: |
| 160 |
print( |
| 161 |
" # Indirectly left-recursive; leaders:", |
| 162 |
{name for name in scc if grammar.rules[name].leader}, |
| 163 |
) |
| 164 |
else: |
| 165 |
name = next(iter(scc)) |
| 166 |
if name in gen.first_graph[name]: |
| 167 |
print(" # Left-recursive") |
| 168 |
else: |
| 169 |
print() |
| 170 |
|
| 171 |
if args.verbose: |
| 172 |
dt = t1 - t0 |
| 173 |
diag = tokenizer.diagnose() |
| 174 |
nlines = diag.end[0] |
| 175 |
if diag.type == token.ENDMARKER: |
| 176 |
nlines -= 1 |
| 177 |
print(f"Total time: {dt:.3f} sec; {nlines} lines", end="") |
| 178 |
if dt: |
| 179 |
print(f"; {nlines / dt:.0f} lines/sec") |
| 180 |
else: |
| 181 |
print() |
| 182 |
print("Caches sizes:") |
| 183 |
print(f" token array : {len(tokenizer._tokens):10}") |
| 184 |
print(f" cache : {len(parser._cache):10}") |
| 185 |
if not print_memstats(): |
| 186 |
print("(Can't find psutil; install it for memory stats.)") |
| 187 |
|
| 188 |
|
| 189 |
if __name__ == "__main__": |
| 190 |
if sys.version_info < (3, 8): # noqa: UP036 |
| 191 |
print("ERROR: using pegen requires at least Python 3.8!", file=sys.stderr) |
| 192 |
sys.exit(1) |
| 193 |
main() |