Tasmota/tools/unishox/compress-html-uncompressed.py

217 lines
6.8 KiB
Python

###############################################################
# compresses all files found in ..\..\tasmota\html_uncompressed
# write compressed C code to ..\..\tasmota\html_compressed
# Instructions:
# open a console, e.g. in vscode, open a 'terminal'
# cd .\tools\unishox
# run:
# python compress-html-uncompressed.py
#
# The intent it to commit both uncompressed and compressed to the repo
# else this script would need to be run at build.
#
# Example Tasmota code:
# #ifdef USE_UNISHOX_COMPRESSION
# #include "./html_compressed/HTTP_SCRIPT_CONSOL.h"
# #else
# #include "./html_uncompressed/HTTP_SCRIPT_CONSOL.h"
# #endif
#
###############################################################
import unishox
from sys import argv
from datetime import datetime
from pathlib import Path
from hashlib import sha256
self_dir = Path(__file__).absolute().parent
base_dir = self_dir.parent.parent
def extract_c_string(s: str) -> str:
state = 0
escape = False
out = ""
for c in s:
if state == 0: # before string
if c == '"': # entering string
out = '"'
state = 1
elif c == '/': # start of comment before entering string
state = 99 # we're done
elif state == 1: # in string
if escape: # escaped char
out += '\\' + c
escape = False
elif c == '\\': # escaped char
escape = True
elif c == '"': # end of string
out += '"'
state = 99 # we're done
else:
out += c
return out
def compress_html(source, target, argv=None, verbose=False):
if argv is None: argv = []
with open(source, "r") as f:
text = f.read()
src_sha, old_sha = sha256(text.encode()).hexdigest(), None
if not ('--force' in argv):
with open(target, "r") as f:
for line in f:
prefix = line[:17]
if prefix == '// input sha256: ':
old_sha = line[17:17+64]
break
if src_sha == old_sha:
return (0, 0)
#text = Tk().clipboard_get()
# print(text)
# parsing and cleaning
text_list = text.splitlines()
text = '' #just reuse the string
const_name = '<var>' #default if no name will be found
line_number = 0
for line in text_list:
pos = line.find("const char")
# print(pos, line)
if pos > -1:
line_list = line.rsplit(" ")
for el in line_list:
if el.find('[]') > -1:
const_name = el[:-2] #extract the "const char" variable name
line_list.pop(line_number)
else: # remove line comments
line_el = extract_c_string(line)
# print(line_el)
text = text + line_el
line_number = line_number +1
# print const_name
# print text
#remove unwanted quotation marks
qm = []
pos =0
last_char = ""
for char in text:
if char == "\"":
if last_char != "\\":
qm.append(pos) #find all quotation marks without preceding backslash
last_char = char
pos = pos + 1
# print(qm)
lastel = 0
input = ""
for pos in qm:
sub = text[lastel+1:pos:]
if not sub.isspace() and pos-lastel > 1:
# print(lastel, pos)
input = input + sub #only copy substrings that are not whitespace
# print(text[lastel+1:pos:])
lastel = pos
if verbose:
print("####### Parsing input from " + str(source.relative_to(base_dir)))
print(" Const char name: "+const_name)
#print('####### Cleaned input:')
#print(input)
#construct output (taken from shadinger)
input = input.replace("\\t", "\t")
input = input.replace("\\n", "\n")
input = input.replace("\\r", "\r")
input = input.replace("\\f", "\f")
input = input.replace("\\b", "\b")
input = input.replace("\\\"", u"\u0022")
in_bytes = bytearray(input, 'utf-8')
in_len = len(in_bytes)
out_bytes = bytearray(in_len * 2)
UNISHOX = unishox.Unishox()
out_len = UNISHOX.compress(in_bytes, len(in_bytes), out_bytes, len(out_bytes))
if verbose:
print(" ####### Compression result:")
reduction = 100-(float(out_len)/float(in_len)*100)
print(f" Compressed from {in_len} to {out_len}, -{reduction:.1f}%")
out_bytes = out_bytes[:out_len] # truncate to right size
#PROGMEM is growing in steps 0,8,24,40,56,... bytes of data resulting in size of 0,16,32,48,64,... bytes
for in_real in range(8,in_len+16,16):
if in_real>=in_len:
if verbose:
print(f" Old real PROGMEM-size:{in_real+8}(unused bytes:{in_real-in_len})")
break
for out_real in range(8,out_len+16,16):
if out_real>=out_len:
if verbose:
print(f" New real PROGMEM-size:{out_real+8}(unused bytes:{out_real-out_len})")
break
if verbose:
print(f" the optimal case would be raw bytes + 8, real difference: {in_real - out_real}bytes")
# https://www.geeksforgeeks.org/break-list-chunks-size-n-python/
def chunked(my_list, n):
return [my_list[i * n:(i + 1) * n] for i in range((len(my_list) + n - 1) // n )]
# split in chunks of 20 characters
chunks = chunked(out_bytes, 20)
lines_raw = [ "\"\\x" + "\\x".join( [ '{:02X}'.format(b) for b in chunk ] ) + "\"" for chunk in chunks ]
line_complete = f"const char {const_name}_COMPRESSED[] PROGMEM = " + ("\n" + " "*29).join(lines_raw) + ";"
lines = f"\nconst size_t {const_name}_SIZE = {in_len};\n{line_complete}\n\n"
#print('####### Final output:')
#print(lines)
definition = f"#define {const_name} Decompress({const_name}_COMPRESSED,{const_name}_SIZE).c_str()"
#print(definition)
now = datetime.now() # current date and time
percent = int((float(out_real)/float(in_real))*100.0)
saving = in_real - out_real
#totalIn = totalIn + in_real
#totalSaved = totalSaved + saving
comment = "/////////////////////////////////////////////////////////////////////\n"
comment += "// compressed by tools/unishox/compress-html-uncompressed.py\n"
comment += f"// input sha256: {src_sha}\n"
comment += "/////////////////////////////////////////////////////////////////////\n"
with open(target, "w") as f:
f.write(comment + lines + definition)
return (in_real, saving)
if verbose:
print("####### Wrote output to " + str(target.relative_to(base_dir)))
def compress_dir(source_dir, target_dir, argv=None, verbose=False):
totalIn, totalSaved = 0, 0
for source in source_dir.iterdir():
target = Path(target_dir, source.stem + ".h")
bytesIn, bytesSaved = compress_html(source, target, argv, verbose)
totalIn += bytesIn
totalSaved += bytesSaved
return (totalIn, totalSaved)
if __name__ == '__main__':
path_uncompressed = Path(base_dir, 'tasmota', 'html_uncompressed')
path_compressed = Path(base_dir, 'tasmota', 'html_compressed')
totalIn, totalSaved = compress_dir(path_uncompressed, path_compressed, argv, True)
if totalSaved > 0:
print(f"If all files are in use, total saving was {totalSaved} out of {totalIn}")