QB EBOOT Insert v2

#v2:Support {} IDs for ascii encode fields
import binascii
import os
import shutil
import struct

INPUT_COL = 2
PTR_COL = 3
s1 = " .,'!?abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890~-&>():;/%"
s2 = "　．，’！？ａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺ１２３４５６７８９０～‐＆＞（）：；／％"
translate_table = str.maketrans(s1, s2)    #For converting between ASCII and fullwidth SHIFT-JIS characters

def is_hex(n):
    try:
        int(n, 16)
    except ValueError:
        return False
    return True

def text_decode(s, encoding='ascii', translate = True):
    out = b''
    while '"' in s:
        s = s.replace('"', '“', 1)
        s = s.replace('"', '”', 1)
    while '{' in s:
        if translate == True:
            out += s[:s.find('{')].translate(translate_table).encode(encoding)
        else:
            out += s[:s.find('{')].encode(encoding)
        code_pos = s.find('{')+1
        code_end = s.find('}')
        if code_end == -1:
            print('decode error')
            quit()
        out += s[code_pos:code_end].encode('ascii')
        s = s[code_end+1:]
    if translate == True:
        out += s.translate(translate_table).encode(encoding)
    else:
        out += s.encode(encoding)
    return out

def load_input(filename):
    inputdata = []
    with open(filename, 'r', encoding='utf-8') as f:
        for line in f:
            line = line.rstrip('\r\n').split('\t')
            if line[0] != '':               #New entry
                ptrs = []
                for ptr in line[PTR_COL:]:
                    if is_hex(ptr):
                        ptr = int(ptr, 16)
                    else:                   #End of pointer list
                        break
                    ptrs.append(ptr)
                inputdata.append([int(line[0], 16), line[INPUT_COL], ptrs]) #addr, text, ptrs
            else:                           #Additional line for same entry
                inputdata[-1][1] += '\x0a' + line[INPUT_COL]
    return inputdata

if os.path.isfile('EBOOT.orig'):    #This is to create a copy of the original (for running tool multiple times)
    pass
elif os.path.isfile('EBOOT.BIN'):
    shutil.copy('EBOOT.BIN', 'EBOOT.orig')
else:
    print('Missing EBOOT.bin.')
    quit()

with open('EBOOT.orig', 'rb') as f:
    filedata = bytearray(f.read())

#Insert SHIFT-JIS text (which is most of it)
inputdata = load_input('SJIS.tsv')
nextpos = 0x16CE00                          #Start of extra data (expanded EBOOT)
for addr, text, ptrs in sorted(inputdata, key=lambda x: x[0], reverse=True):
#Go backwards (reverse = True) // key is addr - we want to do later lines first
    endpos = filedata.find(b'\x00', addr)   #Find end of string
    while filedata[endpos] == 0:            #Find start of next string
        endpos += 1
    endpos -= 2                             #Need 2 nulls (at least) for SJIS terminator (I think)
    avail_space = endpos - addr
    text = text_decode(text, encoding='cp932')
    if len(text) > avail_space:                         #NOT enough space
        filedata[addr:endpos] = b'\x00' * avail_space   #Blank out the line
        text += b'\x00' * 2                             #Add terminator
        while len(text) % 4 != 0:                       #Align to word boundary
            text += b'\x00'
        filedata[nextpos:nextpos+len(text)] = text      #Insert text
#Compute virtual address (vaddr) - 0x89C4200 is the difference between file
#image position and virtual address for text in the EBOOT extra data area
        vaddr = nextpos + 0x89C4200
        for ptr in ptrs:
            if ptr > 0x08000000:                        #Virtual address specified in input
#0x8803F8C is the difference between file image position and virtual address
#for text in the original EBOOT area (not the extra data section)
                vaddr_orig = addr + 0x8803F8C
                if (vaddr & 0xFFFF) > 0x8000:           #Accomodate addiu instructions (all of them are addiu)
                    vaddr += 0x10000
                ptr = ptr - 0x8803F8C                   #Find file address of ptr
                filedata[ptr:ptr+2] = struct.pack('<H', vaddr >> 16)        #Write high bytes
                ptr = filedata.find(struct.pack('<H', vaddr_orig & 0xFFFF),
                                    ptr)                                    #Search for original low bytes
                if ptr % 4 != 0:                                            #Error on search misses (was really for debugging)
                    print('{} Error: ptr={}, searchedfor={}'.format( \
                        hex(addr), hex(ptr),  binascii.hexlify(struct.pack('<H', vaddr_orig & 0xFFFF))))
                filedata[ptr:ptr+2] = struct.pack('<H', vaddr & 0xFFFF)     #Write low bytes
            else:                                       #File address specified in input (this is a lot simpler)
                ptr += 0x20                             #Accomodate 0x20 added for new program header in EBOOT
                filedata[ptr:ptr+4] = struct.pack('<I', vaddr)              #Write the pointer
        nextpos += len(text)                            #Update state (position next text should go)
    else:                                               #THERE IS enough space
        text += b'\x00' * (avail_space - len(text))     #Pad the text out to the right length
        filedata[addr:endpos] = text                    #Replace the text
#Insert UTF-8 text
inputdata = load_input('UTF8.tsv')
for addr, text, ptrs in inputdata:
    endpos = filedata.find(b'\x00', addr)   #Find end of string
    avail_space = endpos - addr
    text = text_decode(text, encoding='utf-8', translate=False)
    if len(text) > avail_space:
        print('{} too long. Text len={}, avail space={}'.format(hex(addr), len(text), avail_space))
    else:
        text = text + b'\x00' * (avail_space - len(text))
        filedata[addr:endpos] = text
with open('EBOOT.BIN', 'wb') as f:
    f.write(filedata)