1
0
mirror of https://github.com/zeldaret/oot.git synced 2024-09-21 12:54:51 +00:00
oot/tools/compress.py
cadmic d674dad3da
Match compression for gc-eu-mq (#1704)
* Improve compression

* Format

* Typo

* Use Python assignment expression in tools/dmadata.py

Co-authored-by: Dragorn421 <Dragorn421@users.noreply.github.com>

* Review decompress_baserom.py

* Replace DEFINE_DEBUG_SCENE with CPP defines

* Pass is_zlib_compressed instead of version

* Reword NOLOAD comment in write_compress_ranges

* Remove redundant comment

---------

Co-authored-by: Dragorn421 <Dragorn421@users.noreply.github.com>
2024-02-03 23:59:19 +01:00

279 lines
8.9 KiB
Python
Executable File

#!/usr/bin/env python3
# SPDX-FileCopyrightText: 2024 zeldaret
# SPDX-License-Identifier: CC0-1.0
from __future__ import annotations
import argparse
from pathlib import Path
import dataclasses
import struct
import time
import multiprocessing
import multiprocessing.pool
import crunch64
import dmadata
def align(v: int):
v += 0xF
return v // 0x10 * 0x10
@dataclasses.dataclass
class RomSegment:
vrom_start: int
vrom_end: int
is_compressed: bool
data: memoryview | None
data_async: multiprocessing.pool.AsyncResult | None
@property
def uncompressed_size(self):
return self.vrom_end - self.vrom_start
# Make interrupting the compression with ^C less jank
# https://stackoverflow.com/questions/72967793/keyboardinterrupt-with-python-multiprocessing-pool
def set_sigint_ignored():
import signal
signal.signal(signal.SIGINT, signal.SIG_IGN)
def compress_rom(
rom_data: memoryview,
dmadata_offset: int,
compress_entries_indices: set[int],
n_threads: int = None,
):
"""
rom_data: the uncompressed rom data
dmadata_offset: the offset in the rom where the dmadata starts
compress_entries_indices: the indices in the dmadata of the segments that should be compressed
n_threads: how many cores to use for compression
"""
# Segments of the compressed rom (not all are compressed)
compressed_rom_segments: list[RomSegment] = []
dma_entries = dmadata.read_dmadata(rom_data, dmadata_offset)
# We sort the DMA entries by ROM start because `compress_entries_indices`
# refers to indices in ROM order, but the uncompressed dmadata might not be
# in ROM order.
dma_entries.sort(key=lambda dma_entry: dma_entry.vrom_start)
with multiprocessing.Pool(n_threads, initializer=set_sigint_ignored) as p:
# Extract each segment from the input rom
for entry_index, dma_entry in enumerate(dma_entries):
segment_rom_start = dma_entry.rom_start
segment_rom_end = dma_entry.rom_start + (
dma_entry.vrom_end - dma_entry.vrom_start
)
segment_data_uncompressed = rom_data[segment_rom_start:segment_rom_end]
is_compressed = entry_index in compress_entries_indices
if is_compressed:
segment_data = None
segment_data_async = p.apply_async(
crunch64.yaz0.compress,
(bytes(segment_data_uncompressed),),
)
else:
segment_data = segment_data_uncompressed
segment_data_async = None
compressed_rom_segments.append(
RomSegment(
dma_entry.vrom_start,
dma_entry.vrom_end,
is_compressed,
segment_data,
segment_data_async,
)
)
# Wait on compression of all compressed segments
waiting_on_segments = [
segment for segment in compressed_rom_segments if segment.is_compressed
]
total_uncompressed_size_of_data_to_compress = sum(
segment.uncompressed_size for segment in waiting_on_segments
)
uncompressed_size_of_data_compressed_so_far = 0
while waiting_on_segments:
# Show progress
progress = (
uncompressed_size_of_data_compressed_so_far
/ total_uncompressed_size_of_data_to_compress
)
print(f"Compressing... {progress * 100:.1f}%", end="\r")
# The segments for which the compression is not finished yet are
# added to this list
still_waiting_on_segments = []
got_some_results = False
for segment in waiting_on_segments:
assert segment.data is None
assert segment.data_async is not None
try:
compressed_data = segment.data_async.get(0)
except multiprocessing.TimeoutError:
# Compression not finished yet
still_waiting_on_segments.append(segment)
else:
# Compression finished!
assert isinstance(compressed_data, bytes)
segment.data = memoryview(compressed_data)
uncompressed_size_of_data_compressed_so_far += (
segment.uncompressed_size
)
got_some_results = True
segment.data_async = None
if not got_some_results and still_waiting_on_segments:
# Nothing happened this wait iteration, idle a bit
time.sleep(0.010)
waiting_on_segments = still_waiting_on_segments
print("Putting together the compressed rom...")
# Put together the compressed rom
compressed_rom_size = sum(
align(len(segment.data)) for segment in compressed_rom_segments
)
pad_to_multiple_of = 8 * 2**20 # 8 MiB
compressed_rom_size_padded = (
(compressed_rom_size + pad_to_multiple_of - 1)
// pad_to_multiple_of
* pad_to_multiple_of
)
compressed_rom_data = memoryview(bytearray(compressed_rom_size_padded))
compressed_rom_dma_entries: list[dmadata.DmaEntry] = []
rom_offset = 0
for segment in compressed_rom_segments:
assert segment.data is not None
segment_rom_start = rom_offset
segment_rom_end = align(segment_rom_start + len(segment.data))
i = segment_rom_start + len(segment.data)
assert i <= len(compressed_rom_data)
compressed_rom_data[segment_rom_start:i] = segment.data
compressed_rom_dma_entries.append(
dmadata.DmaEntry(
segment.vrom_start,
segment.vrom_end,
segment_rom_start,
segment_rom_end if segment.is_compressed else 0,
)
)
rom_offset = segment_rom_end
assert rom_offset == compressed_rom_size
# Pad the compressed rom with the pattern matching the baseroms
for i in range(compressed_rom_size, compressed_rom_size_padded):
compressed_rom_data[i] = i % 256
# Write the new dmadata
offset = dmadata_offset
for dma_entry in compressed_rom_dma_entries:
dma_entry.to_bin(compressed_rom_data[offset:])
offset += dmadata.DmaEntry.SIZE_BYTES
return compressed_rom_data
def main():
parser = argparse.ArgumentParser()
parser.add_argument(
"--in",
dest="in_rom",
required=True,
help="path to an uncompressed rom to be compressed",
)
parser.add_argument(
"--out",
dest="out_rom",
required=True,
help="path of the compressed rom to write out",
)
parser.add_argument(
"--dma-start",
dest="dma_start",
type=lambda s: int(s, 16),
required=True,
help=(
"The dmadata location in the rom, as a hexadecimal offset (e.g. 0x12f70)."
),
)
parser.add_argument(
"--compress",
dest="compress_ranges",
required=True,
help=(
"The indices in the dmadata of the entries to be compressed,"
" where 0 is the first entry."
" It is a comma-separated list of individual indices and inclusive ranges."
" e.g. '0-1,3,5,6-9' is all indices from 0 to 9 (included) except 2 and 4."
),
)
parser.add_argument(
"--threads",
dest="n_threads",
type=int,
default=1,
help="how many cores to use for parallel compression",
)
args = parser.parse_args()
in_rom_p = Path(args.in_rom)
if not in_rom_p.exists():
parser.error(f"Input rom file {in_rom_p} doesn't exist.")
out_rom_p = Path(args.out_rom)
dmadata_offset = args.dma_start
compress_ranges_str: str = args.compress_ranges
compress_entries_indices = set()
for compress_range_str in compress_ranges_str.split(","):
compress_range_ends_str = compress_range_str.split("-")
assert len(compress_range_ends_str) <= 2, (
compress_range_ends_str,
compress_range_str,
compress_ranges_str,
)
compress_range_ends = [int(v_str) for v_str in compress_range_ends_str]
if len(compress_range_ends) == 1:
compress_entries_indices.add(compress_range_ends[0])
else:
assert len(compress_range_ends) == 2
compress_range_first, compress_range_last = compress_range_ends
compress_entries_indices.update(
range(compress_range_first, compress_range_last + 1)
)
n_threads = args.n_threads
in_rom_data = in_rom_p.read_bytes()
out_rom_data = compress_rom(
memoryview(in_rom_data),
dmadata_offset,
compress_entries_indices,
n_threads,
)
out_rom_p.write_bytes(out_rom_data)
if __name__ == "__main__":
main()