feat: update wallet handling to support separate TON and USDT balances

- Refactor `get_wallet()` to return `ton_balance` and `usdt_balance` in `WalletInfo`.
- Update transaction processing to validate balances for both TON and USDT payment methods.
- Introduce `parse_required_payment_amount` utility to extract payment amounts from responses.
- Modify tests to cover new balance checks and payment method handling.
- Upgrade GitHub Actions artifact upload action to v7.
- Enhance documentation and examples to reflect changes in wallet balance handling.

tests/001_test_decode.py

@@ -1,37 +1,62 @@
 """Tests for clean_decode() — TON BOC payload decoding."""
 
+import base64
 import re
+from unittest.mock import patch
 
 import pytest
+from ton_core import Cell
 
 from pyfragment.types import ParseError
 from pyfragment.utils.decoder import clean_decode
 
-PAYLOADS = [
+PAYLOAD_CASES = [
     pytest.param(
         "te6ccgEBAgEALwABTgAAAAAxMDAwMDAwIFRlbGVncmFtIFN0YXJzIAoKUmVmI1RQb01wegEABkM3ZQ",
+        True,
         id="stars",
     ),
     pytest.param(
         "te6ccgEBAgEANAABTgAAAABUZWxlZ3JhbSBQcmVtaXVtIGZvciAxIHllYXIgCgpSZWYjcgEAEE9OQnM2cmNt",
+        True,
         id="premium",
     ),
     pytest.param(
         "te6ccgEBAgEAMAABTgAAAABUZWxlZ3JhbSBhY2NvdW50IHRvcCB1cCAKClJlZiNrMXpDRQEACFkxd3g",
+        True,
         id="topup",
     ),
+    pytest.param(
+        "te6ccgEBAgEAfgABqA-KfqVP885dhccidjC3GwgBCkiH8LM_zUu0afyGCTWJwX1mDjdlf2rMa9UoQlD4UHUAF1jLlcMomlo5RJTwl8jnDDdfdhc7EgQQWPqFQ9IjyLPCAwEASgAAAAA1MCBUZWxlZ3JhbSBTdGFycyAKClJlZiNtOUpoWndBcFE",
+        False,
+        id="real_stars_50",
+    ),
+    pytest.param(
+        "te6ccgEBAgEANgABTgAAAABUZWxlZ3JhbSBQcmVtaXVtIGZvciAzIG1vbnRocyAKClJlZgEAFCMzcFdKdGJkYnU",
+        False,
+        id="real_premium_3m",
+    ),
+    pytest.param(
+        "te6ccgEBAwEAhgABqg-KfqWibdDaYaJCPUWWgvAIAQpIh_CzP81LtGn8hgk1icF9Zg43ZX9qzGvVKEJQ-FB1ABdYy5XDKJpaOUSU8JfI5ww3X3YXOxIEEFj6hUPSI8izwgMBAU4AAAAAMTAwMDAwIFRlbGVncmFtIFN0YXJzIAoKUmVmIzBoZ0RmNEYCAAQ5VA",
+        False,
+        id="real_stars_100k",
+    ),
 ]
 
 
 # Decode valid payload tests
 
 
-@pytest.mark.parametrize("payload", PAYLOADS)
-def test_decode_payload(payload: str) -> None:
+@pytest.mark.parametrize(("payload", "strict_ref"), PAYLOAD_CASES)
+def test_decode_payload(payload: str, strict_ref: bool) -> None:
     result = clean_decode(payload)
-    assert "Telegram" in result
-    assert re.search(r"Ref#[A-Za-z0-9]+", result), f"no Ref# in {result!r}"
-    assert all(ord(c) < 128 for c in result), f"non-ASCII chars in {result!r}"
+    if isinstance(result, str):
+        assert "Telegram" in result
+        if strict_ref:
+            assert re.search(r"Ref#[A-Za-z0-9]+", result), f"no Ref# in {result!r}"
+        assert all(ord(c) < 128 for c in result), f"non-ASCII chars in {result!r}"
+    else:
+        assert isinstance(result, Cell)
 
 
 # Edge case tests
@@ -44,3 +69,64 @@ def test_empty_payload_returns_empty_string() -> None:
 def test_invalid_payload_raises_parse_error() -> None:
     with pytest.raises(ParseError):
         clean_decode("!!!not-valid-base64!!!")
+
+
+def test_decode_payload_accepts_base64url_alphabet() -> None:
+    class _FakeSlice:
+        def load_uint(self, _: int) -> int:
+            return 0
+
+        def load_snake_string(self) -> str:
+            return "Telegram Stars Ref#abc"
+
+    class _FakeCell:
+        def begin_parse(self) -> _FakeSlice:
+            return _FakeSlice()
+
+    raw = b"\xfb\xef\xff\x00"
+    payload = base64.urlsafe_b64encode(raw).decode().rstrip("=")
+
+    with patch("pyfragment.utils.decoder.Cell.one_from_boc", return_value=_FakeCell()) as mocked:
+        result = clean_decode(payload)
+
+    mocked.assert_called_once_with(raw)
+    assert result == "Telegram Stars Ref#abc"
+
+
+def test_clean_decode_returns_text_comment_when_utf8() -> None:
+    class _FakeSlice:
+        def load_uint(self, _: int) -> int:
+            return 0
+
+        def load_snake_string(self) -> str:
+            return "Telegram Premium Ref#abc"
+
+    class _FakeCell:
+        def begin_parse(self) -> _FakeSlice:
+            return _FakeSlice()
+
+    payload = base64.urlsafe_b64encode(b"\x00\x01").decode().rstrip("=")
+    with patch("pyfragment.utils.decoder.Cell.one_from_boc", return_value=_FakeCell()):
+        parsed = clean_decode(payload)
+
+    assert parsed == "Telegram Premium Ref#abc"
+
+
+def test_clean_decode_returns_cell_for_binary_payload() -> None:
+    class _FakeSlice:
+        def load_uint(self, _: int) -> int:
+            return 0
+
+        def load_snake_string(self) -> str:
+            raise UnicodeDecodeError("utf-8", b"\xff", 0, 1, "invalid start byte")
+
+    class _FakeCell:
+        def begin_parse(self) -> _FakeSlice:
+            return _FakeSlice()
+
+    payload = base64.urlsafe_b64encode(b"\x00\x01").decode().rstrip("=")
+    fake_cell: object = _FakeCell()
+    with patch("pyfragment.utils.decoder.Cell.one_from_boc", return_value=fake_cell):
+        parsed = clean_decode(payload)
+
+    assert parsed is fake_cell