Parse byte string to numbers >>> from dask.utils import parse_bytes >>> parse_bytes('100') 100 >>> parse_bytes('100 MB') 100000000 >>> parse_bytes('100M') 100000000 >>> parse_bytes('5kB') 5000 >>> parse_bytes('5.4 kB') 5400 >>> parse_bytes('1kiB')
(s: float | str)
| 1583 | |
| 1584 | |
| 1585 | def parse_bytes(s: float | str) -> int: |
| 1586 | """Parse byte string to numbers |
| 1587 | |
| 1588 | >>> from dask.utils import parse_bytes |
| 1589 | >>> parse_bytes('100') |
| 1590 | 100 |
| 1591 | >>> parse_bytes('100 MB') |
| 1592 | 100000000 |
| 1593 | >>> parse_bytes('100M') |
| 1594 | 100000000 |
| 1595 | >>> parse_bytes('5kB') |
| 1596 | 5000 |
| 1597 | >>> parse_bytes('5.4 kB') |
| 1598 | 5400 |
| 1599 | >>> parse_bytes('1kiB') |
| 1600 | 1024 |
| 1601 | >>> parse_bytes('1e6') |
| 1602 | 1000000 |
| 1603 | >>> parse_bytes('1e6 kB') |
| 1604 | 1000000000 |
| 1605 | >>> parse_bytes('MB') |
| 1606 | 1000000 |
| 1607 | >>> parse_bytes(123) |
| 1608 | 123 |
| 1609 | >>> parse_bytes('5 foos') |
| 1610 | Traceback (most recent call last): |
| 1611 | ... |
| 1612 | ValueError: Could not interpret 'foos' as a byte unit |
| 1613 | """ |
| 1614 | if isinstance(s, (int, float)): |
| 1615 | return int(s) |
| 1616 | s = s.replace(" ", "") |
| 1617 | if not any(char.isdigit() for char in s): |
| 1618 | s = "1" + s |
| 1619 | |
| 1620 | for i in range(len(s) - 1, -1, -1): |
| 1621 | if not s[i].isalpha(): |
| 1622 | break |
| 1623 | index = i + 1 |
| 1624 | |
| 1625 | prefix = s[:index] |
| 1626 | suffix = s[index:] |
| 1627 | |
| 1628 | try: |
| 1629 | n = float(prefix) |
| 1630 | except ValueError as e: |
| 1631 | raise ValueError("Could not interpret '%s' as a number" % prefix) from e |
| 1632 | |
| 1633 | try: |
| 1634 | multiplier = byte_sizes[suffix.lower()] |
| 1635 | except KeyError as e: |
| 1636 | raise ValueError("Could not interpret '%s' as a byte unit" % suffix) from e |
| 1637 | |
| 1638 | result = n * multiplier |
| 1639 | return int(result) |
| 1640 | |
| 1641 | |
| 1642 | byte_sizes = { |