Start working on the harness

This commit is contained in:
John Doty 2024-05-29 07:48:55 -07:00
parent 7c4705714e
commit 4f8aef3f89

View file

@ -1,4 +1,9 @@
import bisect
import enum
import select
import sys
import termios
import tty
import typing
import grammar
@ -88,46 +93,139 @@ def parse(table, tokens, trace=None):
)
def harness(lexer_func, grammar_func, start_rule, source_path):
# generator = parser.GenerateLR1
generator = parser.GenerateLALR
# https://en.wikipedia.org/wiki/ANSI_escape_code
# https://gist.github.com/fnky/458719343aabd01cfb17a3a4f7296797
trace = None
# trace = trace_state
table = grammar_func().build_table(start=start_rule, generator=generator)
print(f"{len(table)} states")
class CharColor(enum.IntEnum):
CHAR_COLOR_DEFAULT = 0
CHAR_COLOR_BLACK = 30
CHAR_COLOR_RED = enum.auto()
CHAR_COLOR_GREEN = enum.auto()
CHAR_COLOR_YELLOW = enum.auto()
CHAR_COLOR_BLUE = enum.auto()
CHAR_COLOR_MAGENTA = enum.auto()
CHAR_COLOR_CYAN = enum.auto()
CHAR_COLOR_WHITE = enum.auto() # Really light gray
CHAR_COLOR_BRIGHT_BLACK = 90 # Really dark gray
CHAR_COLOR_BRIGHT_RED = enum.auto()
CHAR_COLOR_BRIGHT_GREEN = enum.auto()
CHAR_COLOR_BRIGHT_YELLOW = enum.auto()
CHAR_COLOR_BRIGHT_BLUE = enum.auto()
CHAR_COLOR_BRIGHT_MAGENTA = enum.auto()
CHAR_COLOR_BRIGHT_CYAN = enum.auto()
CHAR_COLOR_BRIGHT_WHITE = enum.auto()
average_entries = sum(len(row) for row in table) / len(table)
max_entries = max(len(row) for row in table)
print(f"{average_entries} average, {max_entries} max")
if source_path:
with open(source_path, "r", encoding="utf-8") as f:
src = f.read()
tokens = lexer_func(src)
def ESC(x: bytes) -> bytes:
return b"\033" + x
def CSI(x: bytes) -> bytes:
return ESC(b"[" + x)
CLEAR = CSI(b"H") + CSI(b"0m")
def enter_alt_screen():
sys.stdout.buffer.write(CSI(b"?1049h"))
def leave_alt_screen():
sys.stdout.buffer.write(CSI(b"?1049l"))
class Harness:
source: str | None
def __init__(self, lexer_func, grammar_func, start_rule, source_path):
# self.generator = parser.GenerateLR1
self.generator = parser.GenerateLALR
self.lexer_func = lexer_func
self.grammar_func = grammar_func
self.start_rule = start_rule
self.source_path = source_path
self.source = None
self.table = None
self.tokens = None
self.tree = None
self.errors = None
def run(self):
while True:
i, _, _ = select.select([sys.stdin], [], [], 1)
if i:
k = sys.stdin.read(1)
print(f"Key {k}\r")
return
self.update()
def update(self):
if self.table is None:
self.table = self.grammar_func().build_table(
start=self.start_rule, generator=self.generator
)
if self.tokens is None:
with open(self.source_path, "r", encoding="utf-8") as f:
self.source = f.read()
self.tokens = self.lexer_func(self.source)
# print(f"{tokens.lines}")
# tokens.dump(end=5)
(_, errors) = parse(table, tokens, trace=trace)
if len(errors) > 0:
print(f"{len(errors)} errors:")
for error in errors:
print(f" {error}")
if self.tree is None and self.errors is None:
(tree, errors) = parse(self.table, self.tokens, trace=None)
self.tree = tree
self.errors = errors
sys.stdout.buffer.write(CLEAR)
rows, cols = termios.tcgetwinsize(sys.stdout.fileno())
average_entries = sum(len(row) for row in self.table) / len(self.table)
max_entries = max(len(row) for row in self.table)
print(f"{len(self.table)} states - {average_entries} average, {max_entries} max\r")
if self.tree is not None:
lines = []
self.format_node(lines, self.tree)
for line in lines[: rows - 2]:
print(line[:cols] + "\r")
sys.stdout.flush()
sys.stdout.buffer.flush()
def format_node(self, lines, node, indent=0):
"""Print out an indented concrete syntax tree, from parse()."""
lines.append((" " * indent) + node[0])
for child in node[1]:
self.format_node(lines, child, indent + 2)
if __name__ == "__main__":
import sys
source_path = None
if len(sys.argv) == 2:
source_path = sys.argv[1]
harness(
lexer_func=grammar.FineTokens,
grammar_func=grammar.FineGrammar,
start_rule="file",
source_path=source_path,
)
fd = sys.stdin.fileno()
old_settings = termios.tcgetattr(fd)
try:
tty.setraw(fd)
enter_alt_screen()
h = Harness(
lexer_func=grammar.FineTokens,
grammar_func=grammar.FineGrammar,
start_rule="file",
source_path=source_path,
)
h.run()
finally:
leave_alt_screen()
termios.tcsetattr(fd, termios.TCSADRAIN, old_settings)
# print(parser_faster.format_table(gen, table))
# print()