Files
bal-electrum-plugin/bal/core/animated_qr.py

1181 lines
42 KiB
Python

"""
bal.core.animated_qr
====================
GUI-free implementation of the interoperable animated-QR transfer formats
used to move BAL will data between devices.
Supported wire formats (each self-describing and order-independent on
receive):
* **BALQR** (native): ``BAL1<TTT><iii><flag><payload>`` compact fixed-width
header (11 chars, 3-digit base36 count fields, no separators); legacy
``BALQR1|total|index|flags|payload`` still imported.
* **BC-UR v1** (BCR-2020-005 rev1 draft, May 2020)::
ur:bytes/1of7/<bc32-digest>/<bc32-fragment>
Fragments partition the BC32 rendering of the CBOR byte string; the
SHA-256 digest of the wrapped payload ties the parts together.
* **BC-UR v2** (BCR-2020-005 rev 2 / BCR-2020-012)::
ur:bytes/2-9/<bytewords-minimal fragment>
Fountain-coded parts; each part is a CBOR array
``[seq_num, seq_len, message_len, checksum, data]`` whose CBOR bytes are
bytewords-minimal encoded with a trailing per-part CRC-32. The
``checksum`` field holds the CRC-32 of the whole wrapped message, so the
parts are mixable and order-independent.
* **BBQR** (Coinkite)::
B$<encoding><type><2 base36 total><2 base36 index><payload>
Equal-length text frames; the payload is uppercase hex, RFC-4648
base32, or raw-deflate (``wbits=-10``) zlib plus base32.
Everything is implemented from scratch on top of the Python standard library
only (``zlib``, ``hashlib``, ``base64``), so the shipped plugin zip stays a
self-contained bundle with no third-party dependencies (house rule).
This module never imports Qt or any Electrum GUI code (house rule).
"""
from __future__ import annotations
import base64
import hashlib
import zlib
from typing import Dict, FrozenSet, List, Optional, Sequence, Set, Tuple
# --------------------------------------------------------------------------- #
# Errors & safety caps
# --------------------------------------------------------------------------- #
class AnimatedQrError(ValueError):
"""Base error for all animated-QR codec failures."""
class FormatNotDetectedError(AnimatedQrError):
"""The scanned text does not look like any known animated-QR format."""
class TransferConflictError(AnimatedQrError):
"""An incoming frame belongs to a different transfer than the open one."""
class SessionLimitError(AnimatedQrError):
"""A receive session exceeded its safety caps."""
class ChecksumError(AnimatedQrError):
"""A part failed its checksum / digest validation."""
# Safety caps for untrusted scanner input.
_MAX_SESSION_PARTS = 20000
_MAX_MESSAGE_BYTES = 32 * 1024 * 1024
# --------------------------------------------------------------------------- #
# CBOR minimals (byte-string envelope + the fountain part header)
# --------------------------------------------------------------------------- #
_BYTE_STR_RES = 0x40 # byte string, length < 24
_BYTE_STR_1 = 0x58 # byte string, 1-byte length
_BYTE_STR_2 = 0x59 # byte string, 2-byte length
_BYTE_STR_4 = 0x60 # byte string, 4-byte length
_ARRAY_RES = 0x80
_UNSIGNED_RES = 0x00
def cbor_byte_string(data: bytes) -> bytes:
"""Wrap ``data`` in the minimal CBOR byte-string envelope (0x40..0x60)."""
n = len(data)
if n < 24:
head = bytes([_BYTE_STR_RES + n])
elif n <= 0xFF:
head = bytes([_BYTE_STR_1, n])
elif n <= 0xFFFF:
head = bytes([_BYTE_STR_2]) + n.to_bytes(2, "big")
elif n <= 0xFFFFFFFF:
head = bytes([_BYTE_STR_4]) + n.to_bytes(4, "big")
else:
raise AnimatedQrError("payload too large for the UR byte-string envelope")
return head + data
def unwrap_ur_cbor(message: bytes) -> bytes:
"""Strip the CBOR byte-string envelope, falling back to the raw bytes.
Receivers keep working even when the emitter embedded the payload without
any CBOR wrapping (some third-party ``ur:bytes`` emitters do).
"""
if not message:
raise AnimatedQrError("empty decoded message")
b0 = message[0]
if _BYTE_STR_RES <= b0 <= 0x57:
header_len, n = 1, b0 - _BYTE_STR_RES
elif b0 == _BYTE_STR_1 and len(message) >= 2:
header_len, n = 2, message[1]
elif b0 == _BYTE_STR_2 and len(message) >= 3:
header_len, n = 3, int.from_bytes(message[1:3], "big")
elif b0 == _BYTE_STR_4 and len(message) >= 5:
header_len, n = 5, int.from_bytes(message[1:5], "big")
else:
return message
if header_len + n != len(message):
raise AnimatedQrError("decoded message has an inconsistent CBOR length")
return message[header_len:]
def _cbor_unsigned(value: int) -> bytes:
if value < 24:
return bytes([_UNSIGNED_RES + value])
if value <= 0xFF:
return bytes([0x18, value])
if value <= 0xFFFF:
return bytes([0x19]) + value.to_bytes(2, "big")
if value <= 0xFFFFFFFF:
return bytes([0x1A]) + value.to_bytes(4, "big")
return bytes([0x1B]) + value.to_bytes(8, "big")
def cbor_part(seq_num: int, seq_len: int, message_len: int, checksum: int, data: bytes) -> bytes:
"""The CBOR body of a BC-UR v2 fountain part (``[seq, seq_len, message_len, checksum, data]``)."""
out = bytearray([_ARRAY_RES + 5])
out += _cbor_unsigned(seq_num)
out += _cbor_unsigned(seq_len)
out += _cbor_unsigned(message_len)
out += _cbor_unsigned(checksum)
out += cbor_byte_string(data)
return bytes(out)
def _need(buf: bytes, pos: int, count: int) -> None:
if pos + count > len(buf):
raise AnimatedQrError("truncated CBOR part header")
def _cbor_read_unsigned(buf: bytes, pos: int) -> Tuple[int, int]:
if pos >= len(buf):
raise AnimatedQrError("truncated CBOR part header")
octet = buf[pos]
if octet & 0xE0 != _UNSIGNED_RES:
raise AnimatedQrError("unexpected CBOR type in part header")
pos += 1
additional = octet & 0x1F
if additional < 24:
return additional, pos
if additional == 24:
_need(buf, pos, 1)
return buf[pos], pos + 1
if additional == 25:
_need(buf, pos, 2)
return int.from_bytes(buf[pos : pos + 2], "big"), pos + 2
if additional == 26:
_need(buf, pos, 4)
return int.from_bytes(buf[pos : pos + 4], "big"), pos + 4
if additional == 27:
_need(buf, pos, 8)
return int.from_bytes(buf[pos : pos + 8], "big"), pos + 8
raise AnimatedQrError("unsupported CBOR integer width in part header")
def _cbor_read_bytes(buf: bytes, pos: int) -> Tuple[bytes, int]:
if pos >= len(buf):
raise AnimatedQrError("truncated CBOR part header")
octet = buf[pos]
pos += 1
if octet & 0xE0 != _BYTE_STR_RES:
raise AnimatedQrError("expected a CBOR byte string in part header")
additional = octet & 0x1F
if additional < 24:
n = additional
elif additional == 24:
_need(buf, pos, 1)
n, pos = buf[pos], pos + 1
elif additional == 25:
_need(buf, pos, 2)
n, pos = int.from_bytes(buf[pos : pos + 2], "big"), pos + 2
elif additional == 26:
_need(buf, pos, 4)
n, pos = int.from_bytes(buf[pos : pos + 4], "big"), pos + 4
else:
raise AnimatedQrError("unsupported CBOR byte-string width in part header")
_need(buf, pos, n)
return buf[pos : pos + n], pos + n
def _cbor_read_array(buf: bytes, pos: int) -> Tuple[int, int]:
if pos >= len(buf):
raise AnimatedQrError("truncated CBOR part header")
octet = buf[pos]
pos += 1
if octet & 0xE0 != _ARRAY_RES:
raise AnimatedQrError("expected a CBOR array in part header")
additional = octet & 0x1F
if additional < 24:
return additional, pos
if additional == 24:
_need(buf, pos, 1)
return buf[pos], pos + 1
if additional == 25:
_need(buf, pos, 2)
return int.from_bytes(buf[pos : pos + 2], "big"), pos + 2
raise AnimatedQrError("unsupported CBOR array header in part")
# --------------------------------------------------------------------------- #
# CRC-32 (same polynomial as ``zlib.crc32``, network byte order)
# --------------------------------------------------------------------------- #
def crc32_int(data: bytes) -> int:
"""CRC-32 over ``data`` as an unsigned 32-bit integer."""
return zlib.crc32(data) & 0xFFFFFFFF
def crc32_bytes(data: bytes) -> bytes:
"""CRC-32 over ``data`` as 4 network-order (big-endian) bytes."""
return crc32_int(data).to_bytes(4, "big")
# --------------------------------------------------------------------------- #
# Bytewords (BCR-2020-012)
# --------------------------------------------------------------------------- #
_BYTEWORDS = (
"ableacidalsoapexaquaarchatomauntawayaxisbackbaldbarnbeltbetabiasbluebodybragbr"
"ewbulbbuzzcalmcashcatschefcityclawcodecolacookcostcruxcurlcuspcyandarkdatadays"
"delidicedietdoordowndrawdropdrumdulldutyeacheasyechoedgeepicevenexamexiteyesfa"
"ctfairfernfigsfilmfishfizzflapflewfluxfoxyfreefrogfuelfundgalagamegeargemsgift"
"girlglowgoodgraygrimgurugushgyrohalfhanghardhawkheathelphighhillholyhopehornhu"
"tsicedideaidleinchinkyintoirisironitemjadejazzjoinjoltjowljudojugsjumpjunkjury"
"keepkenokeptkeyskickkilnkingkitekiwiknoblamblavalazyleaflegsliarlimplionlistlo"
"goloudloveluaulucklungmainmanymathmazememomenumeowmildmintmissmonknailnavyneed"
"newsnextnoonnotenumbobeyoboeomitonyxopenovalowlspaidpartpeckplaypluspoempoolpo"
"sepuffpumapurrquadquizraceramprealredorichroadrockroofrubyruinrunsrustsafesaga"
"scarsetssilkskewslotsoapsolosongstubsurfswantacotasktaxitenttiedtimetinytoilto"
"mbtoystriptunatwinuglyundouniturgeuservastveryvetovialvibeviewvisavoidvowswall"
"wandwarmwaspwavewaxywebswhatwhenwhizwolfworkyankyawnyellyogayurtzapszerozestzi"
"nczonezoom"
)
_WORDS = [_BYTEWORDS[i : i + 4] for i in range(0, 1024, 4)]
_DIM = 26
_WORD_LOOKUP: Optional[List[int]] = None
def _word_lookup() -> List[int]:
"""First/last-letter lookup table (built lazily, mirrors Bytewords)."""
global _WORD_LOOKUP
if _WORD_LOOKUP is None:
table = [-1] * (_DIM * _DIM)
for i, word in enumerate(_WORDS):
x = ord(word[0]) - ord("a")
y = ord(word[3]) - ord("a")
table[y * _DIM + x] = i
_WORD_LOOKUP = table
return _WORD_LOOKUP
def _decode_word(word: str, word_len: int) -> int:
if len(word) != word_len:
raise AnimatedQrError("invalid bytewords word length")
x = ord(word[0]) - ord("a")
y = ord(word[3] if word_len == 4 else word[1]) - ord("a")
if not (0 <= x < _DIM and 0 <= y < _DIM):
raise AnimatedQrError("invalid bytewords characters")
value = _word_lookup()[y * _DIM + x]
if value == -1:
raise AnimatedQrError("invalid bytewords first/last pair")
if word_len == 4:
full = _WORDS[value]
if word[1] != full[1] or word[2] != full[2]:
raise AnimatedQrError("invalid bytewords middle letters")
return value
def bytewords_minimal_encode(data: bytes) -> str:
"""BCR-2020-012 bytewords-minimal: one two-letter word per byte, then CRC."""
crc = data + crc32_bytes(data)
return "".join(_WORDS[b][0] + _WORDS[b][3] for b in crc)
def bytewords_minimal_decode(text: str) -> bytes:
"""Inverse of :func:`bytewords_minimal_encode` (validates the CRC-32)."""
if len(text) % 2:
raise AnimatedQrError("invalid bytewords length (odd)")
values = [_decode_word(text[i : i + 2], 2) for i in range(0, len(text), 2)]
payload = bytes(values)
if len(payload) < 5:
raise AnimatedQrError("bytewords payload too short")
body, checksum = payload[:-4], payload[-4:]
if crc32_bytes(body) != checksum:
raise AnimatedQrError("bytewords CRC-32 mismatch")
return body
# --------------------------------------------------------------------------- #
# BC32 (the deprecated bech32-derived codec used by BC-UR v1)
# --------------------------------------------------------------------------- #
_BC32_ALPHABET = "qpzry9x8gf2tvdw0s3jn54khce6mua7l"
_BC32_REV = {ch: i for i, ch in enumerate(_BC32_ALPHABET)}
_BECH32_GENERATOR = [0x3B6A57B2, 0x26508E6D, 0x1EA119FA, 0x3D4233DD, 0x2A1462B3]
def bech32_polymod(values: Sequence[int]) -> int:
chk = 1
for value in values:
top = chk >> 25
chk = (chk & 0x1FFFFFF) << 5 ^ value
for i in range(5):
if (top >> i) & 1:
chk ^= _BECH32_GENERATOR[i]
return chk
def _bc32_checksum(values: List[int]) -> List[int]:
polymod = bech32_polymod([0] + values + [0] * 6) ^ 0x3FFFFFFF
return [(polymod >> 5 * (5 - i)) & 31 for i in range(6)]
def _bc32_verify(values: List[int]) -> bool:
return bech32_polymod([0] + values) == 0x3FFFFFFF
def bc32_encode(data: bytes) -> str:
"""BCR-2020-005 BC32: bech32 without the human-readable part and divider."""
acc = 0
bits = 0
values: List[int] = []
for byte in data:
acc = (acc << 8) | byte
bits += 8
while bits >= 5:
bits -= 5
values.append((acc >> bits) & 31)
if bits:
values.append((acc << (5 - bits)) & 31)
values += _bc32_checksum(values)
return "".join(_BC32_ALPHABET[v] for v in values)
def bc32_decode(text: str) -> bytes:
"""Inverse of :func:`bc32_encode` (validates the 6-char checksum)."""
lowered = text.lower()
try:
values = [_BC32_REV[ch] for ch in lowered]
except KeyError:
raise AnimatedQrError("invalid BC-UR v1 character") from None
if len(values) < 6 or not _bc32_verify(values):
raise AnimatedQrError("invalid BC-UR v1 checksum")
data = values[:-6]
acc = 0
bits = 0
out = bytearray()
for value in data:
acc = (acc << 5) | value
bits += 5
if bits >= 8:
bits -= 8
out.append((acc >> bits) & 0xFF)
return bytes(out)
# --------------------------------------------------------------------------- #
# xoshiro256** + alias sampler (exact ports of the reference RNG chain)
# --------------------------------------------------------------------------- #
_MASK64 = (1 << 64) - 1
class _Xoshiro256:
"""xoshiro256** 1.0, seeded via SHA-256 of a byte sequence."""
def __init__(self, seed: bytes):
digest = hashlib.sha256(seed).digest()
self._s = [
int.from_bytes(digest[offset : offset + 8], "big")
for offset in range(0, 32, 8)
]
@staticmethod
def _rotl(x: int, k: int) -> int:
return ((x << k) | (x >> (64 - k))) & _MASK64
def next(self) -> int:
result = (self._rotl((self._s[1] * 5) & _MASK64, 7) * 9) & _MASK64
t = (self._s[1] << 17) & _MASK64
s = self._s
s[2] ^= s[0]
s[3] ^= s[1]
s[1] ^= s[2]
s[0] ^= s[3]
s[2] ^= t
s[3] = self._rotl(s[3], 45)
return result
def next_double(self) -> float:
return self.next() / float(1 << 64)
def next_int(self, low: int, high: int) -> int:
return int(self.next_double() * (high - low + 1)) + low
class _RandomAliasSampler:
"""Vose's alias method, built in the exact order of the reference code."""
def __init__(self, probs: Sequence[float]):
total = sum(probs)
assert total > 0
n = len(probs)
normalized = [p * float(n) / total for p in probs]
small: List[int] = []
large: List[int] = []
for i in range(n - 1, -1, -1):
(small if normalized[i] < 1 else large).append(i)
self._probs = [0] * n
self._aliases = [0] * n
while small and large:
a = small.pop()
g = large.pop()
self._probs[a] = normalized[a]
self._aliases[a] = g
normalized[g] += normalized[a] - 1
(small if normalized[g] < 1 else large).append(g)
while large:
self._probs[large.pop()] = 1
while small:
self._probs[small.pop()] = 1
def next(self, rng: _Xoshiro256) -> int:
r1 = rng.next_double()
r2 = rng.next_double()
n = len(self._probs)
i = int(float(n) * r1)
return i if r2 < self._probs[i] else self._aliases[i]
def choose_fragments(seq_num: int, seq_len: int, checksum: int) -> Set[int]:
"""The fragments mixed into a BC-UR v2 fountain part (reference seed math).
Sequence numbers ``1..seq_len`` emit the pure fragment ``{seq_num - 1}``;
every larger sequence number deterministically mixes a pseudo-random
subset of fragments seeded by ``SHA256(seq ‖ checksum)``.
"""
if seq_num <= seq_len:
return {seq_num - 1}
seed = seq_num.to_bytes(4, "big") + checksum.to_bytes(4, "big")
rng = _Xoshiro256(seed)
probs: List[float] = [1.0 / i for i in range(1, seq_len + 1)]
degree = _RandomAliasSampler(probs).next(rng) + 1
remaining = list(range(seq_len))
shuffled: List[int] = []
while remaining:
index = rng.next_int(0, len(remaining) - 1)
shuffled.append(remaining.pop(index))
return set(shuffled[:degree])
def _partition_message(message: bytes, fragment_len: int) -> List[bytes]:
fragments: List[bytes] = []
for offset in range(0, len(message), fragment_len):
fragment = message[offset : offset + fragment_len]
if len(fragment) < fragment_len:
fragment += b"\x00" * (fragment_len - len(fragment))
fragments.append(fragment)
return fragments
def _mix_fragments(fragments: Sequence[bytes], indexes: Set[int], fragment_len: int) -> bytes:
result = bytearray(fragment_len)
for index in indexes:
for i, byte in enumerate(fragments[index]):
result[i] ^= byte
return bytes(result)
# --------------------------------------------------------------------------- #
# BC-UR v2 (bytewords-minimal + fountain)
# --------------------------------------------------------------------------- #
def _ur2_header(seq_num: int, seq_len: int) -> str:
return "ur:bytes/{}-{}/".format(seq_num, seq_len)
def _ur2_part_string(seq_num: int, seq_len: int, message_len: int, checksum: int, data: bytes) -> str:
body = cbor_part(seq_num, seq_len, message_len, checksum, data)
return _ur2_header(seq_num, seq_len) + bytewords_minimal_encode(body)
def _ur2_part_cost(seq_num: int, seq_len: int, message_len: int, checksum: int, data_len: int) -> int:
body_len = len(cbor_part(seq_num, seq_len, message_len, checksum, b"\x00" * data_len))
# bytewords_minimal_encode appends a 4-byte CRC over the body.
return len(_ur2_header(seq_num, seq_len)) + 2 * (body_len + 4)
def ur2_frames(payload: bytes, budget_chars: int) -> List[str]:
"""Encode ``payload`` into BC-UR v2 fountain frames.
``budget_chars`` is the largest frame string the carrying QR code may
hold. The first ``seq_len`` frames are pure (one fragment each); a second
wave of ``seq_len`` mixed (fountain) frames follows so the receiver can
recover with a few parts still missing.
"""
message = cbor_byte_string(payload)
message_len = len(message)
checksum = crc32_int(message)
single_cost = len("ur:bytes/") + len(bytewords_minimal_encode(message))
if single_cost <= budget_chars:
return ["ur:bytes/" + bytewords_minimal_encode(message)]
fragment_len = message_len
fragment_count = 1
while True:
seq_len = fragment_count
worst_seq = 2 * seq_len # the export loop emits up to 2*seq_len parts
cost = _ur2_part_cost(worst_seq, seq_len, message_len, checksum, fragment_len)
if cost <= budget_chars:
break
fragment_count += 1
fragment_len = -(-message_len // fragment_count)
if fragment_count > message_len:
raise AnimatedQrError("QR budget too small for a BC-UR v2 part")
fragments = _partition_message(message, fragment_len)
frames: List[str] = []
for seq_num in range(1, seq_len + 1):
frames.append(_ur2_part_string(seq_num, seq_len, message_len, checksum, fragments[seq_num - 1]))
for seq_num in range(seq_len + 1, 2 * seq_len + 1):
data = _mix_fragments(fragments, choose_fragments(seq_num, seq_len, checksum), fragment_len)
frames.append(_ur2_part_string(seq_num, seq_len, message_len, checksum, data))
return frames
def ur2_parse_part(frame_text: str) -> Tuple[int, int, int, int, bytes]:
"""Parse a BC-UR v2 part into ``(seq, seq_len, message_len, checksum, data)``."""
frame_text = frame_text.strip().lower()
prefix = "ur:bytes/"
if not frame_text.startswith(prefix):
raise AnimatedQrError("not a BC-UR v2 part")
tail = frame_text[len(prefix) :]
if "/" not in tail:
body = bytewords_minimal_decode(tail)
return 1, 1, len(body), crc32_int(body), body
seq_head, words = tail.split("/", 1)
try:
seq_num_s, seq_len_s = seq_head.split("-", 1)
seq_num, seq_len = int(seq_num_s), int(seq_len_s)
except ValueError:
raise AnimatedQrError("bad BC-UR v2 sequence header") from None
if seq_len < 1 or not 1 <= seq_num <= 2**32 - 1:
raise AnimatedQrError("bad BC-UR v2 sequence numbers")
body = bytewords_minimal_decode(words)
arr, pos = _cbor_read_array(body, 0)
if arr != 5:
raise AnimatedQrError("bad BC-UR v2 part header arity")
seq_again, pos = _cbor_read_unsigned(body, pos)
seq_len_again, pos = _cbor_read_unsigned(body, pos)
message_len, pos = _cbor_read_unsigned(body, pos)
checksum, pos = _cbor_read_unsigned(body, pos)
data, pos = _cbor_read_bytes(body, pos)
if pos != len(body):
raise AnimatedQrError("trailing garbage in BC-UR v2 part header")
if seq_again != seq_num or seq_len_again != seq_len:
raise AnimatedQrError("BC-UR v2 part header mismatch")
return seq_num, seq_len, message_len, checksum, bytes(data)
# --------------------------------------------------------------------------- #
# BC-UR v1 (BCR-2020-005 rev1: BC32 fragments + SHA-256 digest)
# --------------------------------------------------------------------------- #
def _ur1_digest(message: bytes) -> str:
return bc32_encode(hashlib.sha256(message).digest())
def _ur1_prefix(index: int, total: int, digest: str) -> str:
return "ur:bytes/{}{}/{}/".format(
index, "of{}".format(total), digest
)
def ur1_frames(payload: bytes, budget_chars: int) -> List[str]:
"""Encode ``payload`` into BC-UR v1 fragments (``NofM`` + BC32 + digest)."""
message = cbor_byte_string(payload)
digest = _ur1_digest(message)
full = bc32_encode(message)
total = 1
while True:
longest = _ur1_prefix(total, total, digest)
capacity = budget_chars - len(longest)
if capacity < 1:
raise AnimatedQrError("QR budget too small for BC-UR v1")
if len(full) <= capacity * total:
break
total += 1
if total > _MAX_SESSION_PARTS:
raise AnimatedQrError("BC-UR v1 transfer demands too many parts")
frames: List[str] = []
pos = 0
for index in range(1, total + 1):
prefix = _ur1_prefix(index, total, digest)
capacity = budget_chars - len(prefix)
frames.append(prefix + full[pos : pos + capacity])
pos += capacity
return frames
def ur1_parse_part(frame_text: str) -> Tuple[int, int, str, str]:
"""Parse a BC-UR v1 part into ``(index, total, digest, fragment)``.
Accepts both the multipart form (``ur:bytes/NofM/<digest>/<frag>``) and
the single-part form (``ur:bytes/<frag>``, no sequence header or digest).
"""
frame_text = frame_text.strip().lower()
prefix = "ur:bytes/"
if not frame_text.startswith(prefix):
raise AnimatedQrError("not a BC-UR v1 part")
tail = frame_text[len(prefix) :]
parts = tail.split("/")
if len(parts) == 1:
return 1, 1, "", parts[0]
if len(parts) != 3:
raise AnimatedQrError("bad BC-UR v1 part structure")
seq_head, digest, fragment = parts
if "of" not in seq_head:
raise AnimatedQrError("BC-UR v1 part misses the sequence header")
try:
index_s, total_s = seq_head.split("of", 1)
index, total = int(index_s), int(total_s)
except ValueError:
raise AnimatedQrError("bad BC-UR v1 sequence header") from None
if total < 1 or not 1 <= index <= total:
raise AnimatedQrError("bad BC-UR v1 sequence numbers")
if len(digest) != 58:
raise AnimatedQrError("bad BC-UR v1 digest")
return index, total, digest, fragment
# --------------------------------------------------------------------------- #
# BBQR (Coinkite)
# --------------------------------------------------------------------------- #
_BBQR_PREFIX = "B$"
def _bbqr_base36(n: int) -> str:
if not 0 <= n <= 1295:
raise AnimatedQrError("BBQR part count out of range")
def digit(x: int) -> str:
return chr(48 + x) if x < 10 else chr(65 + x - 10)
return digit(n // 36) + digit(n % 36)
def _bbqr_base32(data: bytes) -> str:
return base64.b32encode(data).decode("ascii").rstrip("=")
def _bbqr_encode(raw: bytes, encoding: str) -> Tuple[str, str, int]:
"""Return ``(encoding, encoded_text, split_mod)`` honouring the reference."""
if encoding == "H":
return "H", raw.hex().upper(), 2
if encoding == "Z":
compressor = zlib.compressobj(wbits=-10)
compressed = compressor.compress(raw) + compressor.flush()
if len(compressed) < len(raw):
return "Z", _bbqr_base32(compressed), 8
encoding = "2"
if encoding != "2":
raise AnimatedQrError("unknown BBQR encoding")
return "2", _bbqr_base32(raw), 8
def bbqr_frames(payload: bytes, budget_chars: int, encoding: str = "Z", type_code: str = "B") -> List[str]:
"""Encode ``payload`` into BBQR frames (``B$<enc><type><N><n>…``)."""
if len(type_code) != 1 or not type_code.isalnum():
raise AnimatedQrError("bad BBQR type code")
encoding, encoded, split_mod = _bbqr_encode(payload, encoding)
chunk = budget_chars - 8
if chunk < split_mod:
raise AnimatedQrError("QR budget too small for a BBQR frame")
chunk -= chunk % split_mod
if chunk < 1:
raise AnimatedQrError("QR budget too small for a BBQR frame")
if len(payload) > _MAX_MESSAGE_BYTES:
raise AnimatedQrError("BBQR payload exceeds the size cap")
total = -(-len(encoded) // chunk)
if total > 1295:
raise AnimatedQrError("BBQR transfer demands too many parts")
header = _BBQR_PREFIX + encoding + type_code + _bbqr_base36(total)
frames: List[str] = []
pos = 0
for index in range(total):
frames.append(header + _bbqr_base36(index) + encoded[pos : pos + chunk])
pos += chunk
return frames
def bbqr_parse_part(frame_text: str) -> Tuple[str, str, int, int, str]:
"""Parse a BBQR frame into ``(encoding, type_code, total, index, payload)``."""
frame_text = frame_text.strip()
if len(frame_text) < 10 or not frame_text.startswith(_BBQR_PREFIX):
raise AnimatedQrError("not a BBQR frame")
encoding = frame_text[2]
type_code = frame_text[3]
if encoding not in ("H", "2", "Z"):
raise AnimatedQrError("unknown BBQR encoding")
try:
total = int(frame_text[4:6], 36)
index = int(frame_text[6:8], 36)
except ValueError:
raise AnimatedQrError("bad BBQR part numbers") from None
if total < 1 or not 0 <= index < total:
raise AnimatedQrError("bad BBQR part numbers")
if index >= _MAX_SESSION_PARTS:
raise AnimatedQrError("BBQR part number out of range")
return encoding, type_code, total, index, frame_text[8:]
def _bbqr_decode(encoded_parts: Sequence[str], encoding: str) -> bytes:
pieces: List[bytes] = []
for part in encoded_parts:
if encoding == "H":
try:
pieces.append(bytes.fromhex(part))
except ValueError:
raise AnimatedQrError("invalid BBQR hex payload") from None
continue
padding = (8 - (len(part) % 8)) % 8
try:
pieces.append(base64.b32decode(part + "=" * padding))
except (ValueError, TypeError):
raise AnimatedQrError("invalid BBQR base32 payload") from None
raw = b"".join(pieces)
if encoding == "Z":
try:
inflater = zlib.decompressobj(wbits=-10)
out = inflater.decompress(raw, _MAX_MESSAGE_BYTES + 1)
except zlib.error:
raise AnimatedQrError("invalid BBQR zlib payload") from None
if len(out) > _MAX_MESSAGE_BYTES or inflater.unconsumed_tail:
raise AnimatedQrError("BBQR payload exceeds the size cap")
return out
return raw
# --------------------------------------------------------------------------- #
# Format detection & per-frame identity for the shared debounce
# --------------------------------------------------------------------------- #
FORMAT_LABELS = {
"balqr": "BAL QR",
"ur1": "BC-UR v1",
"ur2": "BC-UR v2",
"bbqr": "BBQR",
}
def format_name(fmt: str) -> str:
"""Human-readable name of a wire format for UI labels."""
return FORMAT_LABELS.get(fmt, fmt)
def detect_format(text: str) -> Optional[str]:
"""Return the wire format of a scanned string, or ``None``."""
text = text.strip()
if not text:
return None
lowered = text.lower()
if lowered.startswith(("balqr", "bal1")):
return "balqr"
if text.startswith(_BBQR_PREFIX):
return "bbqr"
if not lowered.startswith("ur:"):
return None
if lowered.startswith("ur:bytes/"):
remainder = lowered[len("ur:bytes/") :]
first = remainder.split("/", 1)[0]
if "of" in first:
return "ur1"
if "-" in first:
return "ur2"
# Single-part: the whole remainder is the body. Prefer a bytewords v2
# body (CBOR byte-string head 0x40..0x60), then BC32 v1.
try:
body = bytewords_minimal_decode(remainder)
except AnimatedQrError:
pass
else:
if body and 0x40 <= body[0] <= 0x60:
return "ur2"
try:
bc32_decode(remainder)
except AnimatedQrError:
return None
return "ur1"
return None
def parse_for_detection(text: str) -> Tuple[str, str, int, int]:
"""Parse a frame and return ``(format, session_key, frame_total, index)``.
``session_key`` identifies the transfer the frame belongs to and drives
the shared reset/ignore/accept debounce. Raises
:class:`AnimatedQrError` when the text cannot be parsed.
"""
fmt = detect_format(text)
if fmt == "balqr":
total, index, _compressed, _payload = _parse_balqr(text)
return "balqr", "balqr:{}".format(total), total, index
if fmt == "ur1":
index, total, digest, _frag = ur1_parse_part(text)
return "ur1", "ur1:{}".format(digest), total, index
if fmt == "ur2":
seq, seq_len, message_len, checksum, _data = ur2_parse_part(text)
return "ur2", "ur2:{}-{}-{}".format(seq_len, message_len, checksum), seq_len, seq
if fmt == "bbqr":
encoding, type_code, total, index, _payload = bbqr_parse_part(text)
return "bbqr", "bbqr:{}{}:{}".format(encoding, type_code, total), total, index
raise FormatNotDetectedError("Not a supported QR transfer format")
def _parse_balqr(text: str) -> Tuple[int, int, bool, str]:
from bal.core.qrtransfer import parse_frame
return parse_frame(text)
# --------------------------------------------------------------------------- #
# Receive sessions (order-independent assembly per format)
# --------------------------------------------------------------------------- #
class _BalQrSession:
def __init__(self):
self._frames: Dict[int, str] = {}
self._total = 0
self._compressed = False
@property
def total(self) -> int:
return self._total
@property
def received(self) -> int:
return len(self._frames)
@property
def done(self) -> bool:
return bool(self._total) and len(self._frames) >= self._total
def add(self, text: str) -> str:
total, index, compressed, payload = _parse_balqr(text)
if self._total and total != self._total:
raise TransferConflictError("BAL QR transfer total changed")
if len(self._frames) >= _MAX_SESSION_PARTS:
raise SessionLimitError("too many BAL QR frames")
if not self._total:
self._total = total
self._compressed = compressed
if index in self._frames:
return "dup"
self._frames[index] = payload
return "ok"
def resolve(self) -> Tuple[str, bool]:
from bal.core.qrtransfer import assemble
text = assemble(self._frames, self._total)
return text, self._compressed
class _Ur1Session:
def __init__(self):
self._total = 0
self._digest = ""
self._fragments: Dict[int, str] = {}
@property
def total(self) -> int:
return self._total
@property
def received(self) -> int:
return len(self._fragments)
@property
def done(self) -> bool:
return bool(self._total) and len(self._fragments) >= self._total
def add(self, text: str) -> str:
index, total, digest, fragment = ur1_parse_part(text)
if self._total:
if total != self._total or digest != self._digest:
raise TransferConflictError("BC-UR v1 transfer digest changed")
else:
self._total = total
self._digest = digest
if total > _MAX_SESSION_PARTS:
raise SessionLimitError("BC-UR v1 demands too many parts")
if index in self._fragments:
return "dup"
self._fragments[index] = fragment
return "ok"
def resolve(self) -> Tuple[str, bool]:
full = "".join(self._fragments[i] for i in range(1, self._total + 1))
try:
message = bc32_decode(full)
except AnimatedQrError:
raise ChecksumError("BC-UR v1 checksum mismatch") from None
if self._digest and _ur1_digest(message) != self._digest:
raise ChecksumError("BC-UR v1 digest mismatch")
return _transfer_text(unwrap_ur_cbor(message)), False
class _Ur2Session:
"""Fountain decoder mirroring the reference (C++/python) semantics."""
def __init__(self):
self._seq_len = 0
self._message_len = 0
self._checksum = 0
self._fragment_len = 0
self._received: Set[int] = set()
self._simple: Dict[FrozenSet[int], bytes] = {}
self._mixed: Dict[FrozenSet[int], bytes] = {}
self._queue: List[Tuple[FrozenSet[int], bytes]] = []
self._processed = 0
self._result: Optional[bytes] = None
self._bad = False
@property
def total(self) -> int:
return self._seq_len
@property
def received(self) -> int:
return self._processed
@property
def done(self) -> bool:
return self._result is not None
def add(self, text: str) -> str:
seq, seq_len, message_len, checksum, data = ur2_parse_part(text)
if self._seq_len:
if not self._validate(seq_len, message_len, checksum, len(data)):
raise TransferConflictError("BC-UR v2 transfer header changed")
else:
self._seq_len = seq_len
self._message_len = message_len
self._checksum = checksum
self._fragment_len = len(data)
if seq_len > _MAX_SESSION_PARTS or message_len > _MAX_MESSAGE_BYTES:
raise SessionLimitError("BC-UR v2 session exceeds safety caps")
indexes = frozenset(choose_fragments(seq, self._seq_len, self._checksum))
self._receive(indexes, bytes(data))
return "ok"
def _validate(self, seq_len: int, message_len: int, checksum: int, data_len: int) -> bool:
return (
seq_len == self._seq_len
and message_len == self._message_len
and checksum == self._checksum
and data_len == self._fragment_len
)
def _receive(self, indexes: FrozenSet[int], data: bytes) -> None:
if self._result is not None or self._bad:
return
self._queue.append((indexes, data))
while self._result is None and not self._bad and self._queue:
self._process(self._queue.pop(0))
self._processed += 1
def _process(self, item: Tuple[FrozenSet[int], bytes]) -> None:
indexes, data = item
if len(indexes) == 1:
self._process_simple(indexes, data)
else:
self._process_mixed(indexes, data)
def _process_simple(self, indexes: FrozenSet[int], data: bytes) -> None:
fragment_index = next(iter(indexes))
if fragment_index in self._received:
return
self._simple[indexes] = data
self._received.add(fragment_index)
if self._received == set(range(self._seq_len)):
self._finish()
return
self._reduce_mixed_by(indexes, data)
def _reduce_mixed_by(self, indexes: FrozenSet[int], data: bytes) -> None:
new_mixed: Dict[FrozenSet[int], bytes] = {}
for other_indexes, other_data in self._mixed.items():
reduced = self._reduce_part(other_indexes, other_data, indexes, data)
if len(reduced[0]) == 1:
self._queue.append(reduced)
else:
new_mixed[reduced[0]] = reduced[1]
self._mixed = new_mixed
def _process_mixed(self, indexes: FrozenSet[int], data: bytes) -> None:
if indexes in self._mixed:
return
reduced_indexes, reduced_data = indexes, data
for simple_indexes, simple_data in self._simple.items():
reduced_indexes, reduced_data = self._reduce_part(
reduced_indexes, reduced_data, simple_indexes, simple_data
)
for other_indexes, other_data in list(self._mixed.items()):
reduced_indexes, reduced_data = self._reduce_part(
reduced_indexes, reduced_data, other_indexes, other_data
)
if len(reduced_indexes) == 1:
self._queue.append((reduced_indexes, reduced_data))
else:
self._reduce_mixed_by(reduced_indexes, reduced_data)
if reduced_indexes not in self._mixed:
self._mixed[reduced_indexes] = reduced_data
@staticmethod
def _reduce_part(
a_indexes: FrozenSet[int], a_data: bytes, b_indexes: FrozenSet[int], b_data: bytes
) -> Tuple[FrozenSet[int], bytes]:
if b_indexes == a_indexes or not b_indexes.issubset(a_indexes):
return a_indexes, a_data
new_indexes = a_indexes - b_indexes
new_data = bytes(x ^ y for x, y in zip(a_data, b_data, strict=True))
return new_indexes, new_data
def _finish(self) -> None:
fragments = []
for index in range(self._seq_len):
key = frozenset([index])
if key not in self._simple:
self._bad = True
return
fragments.append(self._simple[key])
message = b"".join(fragments)[: self._message_len]
if crc32_int(message) != self._checksum:
self._bad = True
return
self._result = message
def resolve(self) -> Tuple[str, bool]:
if self._bad:
raise ChecksumError("BC-UR v2 message checksum mismatch")
if self._result is None:
raise AnimatedQrError("BC-UR v2 session is not complete")
return _transfer_text(unwrap_ur_cbor(self._result)), False
class _BbqrSession:
def __init__(self):
self._total = 0
self._encoding = ""
self._type_code = ""
self._parts: Dict[int, str] = {}
@property
def total(self) -> int:
return self._total
@property
def received(self) -> int:
return len(self._parts)
@property
def done(self) -> bool:
return bool(self._total) and len(self._parts) >= self._total
def add(self, text: str) -> str:
encoding, type_code, total, index, payload = bbqr_parse_part(text)
if self._total:
if (encoding, type_code, total) != (self._encoding, self._type_code, self._total):
raise TransferConflictError("BBQR frame header changed")
else:
self._total = total
self._encoding = encoding
self._type_code = type_code
if index in self._parts:
return "dup"
self._parts[index] = payload
return "ok"
def resolve(self) -> Tuple[str, bool]:
ordered = [self._parts[i] for i in range(self._total)]
raw = _bbqr_decode(ordered, self._encoding)
return _transfer_text(raw), False
def _transfer_text(raw: bytes) -> str:
try:
return raw.decode("utf-8")
except UnicodeDecodeError:
raise AnimatedQrError("decoded transfer is not valid UTF-8") from None
class AnimatedQrSession:
"""Facade over the per-format receive sessions used by the QR import page."""
_DIALECTS = (("balqr", "_BalQrSession"), ("ur1", "_Ur1Session"), ("ur2", "_Ur2Session"), ("bbqr", "_BbqrSession"))
def __init__(self):
self._inner: Optional[object] = None
self._fmt: Optional[str] = None
@property
def format(self) -> Optional[str]:
return self._fmt
def add_part(self, text: str) -> str:
"""Feed one scanned frame; returns ``"ok"``/``"dup"``, raises on bad input."""
fmt = detect_format(text)
if fmt is None:
raise FormatNotDetectedError("Not a supported QR transfer format")
if self._inner is None:
self._fmt = fmt
self._inner = self._make(fmt)
elif fmt != self._fmt:
raise TransferConflictError(
"Switched QR format mid-import ({} -> {})".format(self.format, fmt)
)
return self._inner.add(text) # type: ignore[no-any-return]
@staticmethod
def _make(fmt: str) -> object:
if fmt == "balqr":
return _BalQrSession()
if fmt == "ur1":
return _Ur1Session()
if fmt == "ur2":
return _Ur2Session()
if fmt == "bbqr":
return _BbqrSession()
raise AssertionError("unknown animated-QR format {}".format(fmt))
@property
def total(self) -> int:
return self._inner.total if self._inner is not None else 0
@property
def received(self) -> int:
return self._inner.received if self._inner is not None else 0
@property
def done(self) -> bool:
return bool(self._inner is not None and self._inner.done)
def resolve(self) -> Tuple[str, bool]:
if self._inner is None:
raise AnimatedQrError("no transfer has been received")
return self._inner.resolve() # type: ignore[no-any-return]