doshomik 0.1.1

Parse and format South Asian numeric strings (Bangla/English, lakh-crore system).

doshomik

Parse and format South Asian numeric strings — Bangla and English digits, lakh-crore system, currency markers.

from doshomik import parse, format

parse("১ কোটি ২৫ লক্ষ")   # → 12_500_000
parse("2.5 crore")          # → 25_000_000
parse("৳ 1,25,000")         # → 125_000

format(12_500_000, script="bangla",   grouping="lakh")    # → "১,২৫,০০,০০০"
format(12_500_000, script="english",  grouping="lakh")    # → "1,25,00,000"
format(12_500_000, script="english",  grouping="western") # → "12,500,000"

Installation

pip

pip install doshomik

uv

uv add doshomik

Requires Python 3.10 or later. No runtime dependencies.

Usage

Parsing

parse(s) accepts a string containing South Asian numeric notation and returns a plain int.

from doshomik import parse

# Bangla digits
parse("১২৩")           # → 123
parse("০")             # → 0

# English digits
parse("125000")        # → 125_000

# Multiplier words (Bangla and English)
parse("১ হাজার")       # → 1_000
parse("1 hazar")       # → 1_000
parse("1 thousand")    # → 1_000

parse("১ লাখ")         # → 100_000
parse("1 lakh")        # → 100_000
parse("1 lac")         # → 100_000

parse("১ কোটি")        # → 10_000_000
parse("1 crore")       # → 10_000_000

# Compound expressions
parse("১ কোটি ২৫ লক্ষ")         # → 12_500_000
parse("1 crore 25 lakh")         # → 12_500_000
parse("1 crore 50 lakh 25 hazar") # → 15_025_000

# Decimal multipliers
parse("2.5 crore")    # → 25_000_000
parse("1.5 lakh")     # → 150_000

# Grouped numbers (lakh and western styles)
parse("1,25,000")     # → 125_000
parse("1,000,000")    # → 1_000_000

# Currency markers are stripped automatically
parse("৳ 1,25,000")   # → 125_000
parse("500 taka")     # → 500
parse("500 tk.")      # → 500
parse("bdt 250")      # → 250

# International (as used in English-language BD press)
parse("1 million")    # → 1_000_000
parse("1 billion")    # → 1_000_000_000

# Mixed script
parse("১০ lakh")      # → 1_000_000
parse("10 লাখ")       # → 1_000_000

What parse accepts:

Category	Examples
Bangla digits	০১২৩৪৫৬৭৮৯
English digits	0123456789
Thousands	hazar, hajar, thousand, k, হাজার
Lakhs	lakh, lac, lakhs, লাখ, লক্ষ
Crores	crore, crores, cr, কোটি
International	million, mn, m, billion, bn, b
Grouping commas	lakh-style (1,25,000) and western (1,000,000)
Currency markers	৳, tk, tk., taka, টাকা, rs, rs., bdt
Decimal multipliers	2.5 crore, 1.5 lakh

What parse does NOT accept:

• Bare decimals without a multiplier ("1.5" raises ParseError — result is not an integer)
• Strings with no numeric content ("abc", "")
• Standalone multipliers ("crore", "লাখ")

Formatting

format(n, *, script, grouping) converts a plain int to a formatted string.

from doshomik import format

# script="bangla" | "english"
# grouping="lakh"  | "western" | "none"

format(125_000, script="english", grouping="lakh")    # → "1,25,000"
format(125_000, script="bangla",  grouping="lakh")    # → "১,২৫,০০০"
format(125_000, script="english", grouping="western") # → "125,000"
format(125_000, script="bangla",  grouping="western") # → "১২৫,০০০"
format(125_000, script="english", grouping="none")    # → "125000"
format(125_000, script="bangla",  grouping="none")    # → "১২৫০০০"

# Negative numbers
format(-1_000_000, script="english", grouping="lakh")    # → "-10,00,000"
format(-1_000_000, script="bangla",  grouping="lakh")    # → "-১০,০০,০০০"

Lakh grouping follows the South Asian convention: the rightmost group has 3 digits, every group to the left has 2 digits.

12,50,00,000  →  "12 crore 50 lakh"
 1,25,000     →  "1 lakh 25 thousand"

Default arguments: script="bangla", grouping="lakh" — i.e., format(n) produces a Bangla-script lakh-grouped string.

Error handling

from doshomik import parse, format
from doshomik import ParseError, FormatError, DoshomikError

# DoshomikError is the base class for both ParseError and FormatError
try:
    value = parse("not a number")
except ParseError as e:
    print(e)  # no numeric content found in 'not a number'

try:
    result = format(1.5, script="english", grouping="lakh")
except FormatError as e:
    print(e)  # n must be int, got float

# Catch either with the base class
try:
    parse("crore")  # multiplier without preceding number
except DoshomikError as e:
    print(type(e).__name__, e)

ParseError is raised when:

• The input contains no numeric content
• A multiplier appears without a preceding number
• The computed result is not a whole integer (e.g. "1.5" with no multiplier)

FormatError is raised when:

• n is not a plain int (floats, strings, None, and bool are all rejected)
• script is not "bangla" or "english" (case-sensitive)
• grouping is not "lakh", "western", or "none" (case-sensitive)

API reference

parse(s: str) -> int

Parse a South Asian numeric string and return an integer.

Parameter	Type	Description
s	str	Input string to parse

Raises ParseError on invalid input.

format(n: int, *, script: str = "bangla", grouping: str = "lakh") -> str

Format an integer as a South Asian numeric string.

Parameter	Type	Default	Description
n	int	—	Integer to format (booleans rejected)
script	str	"bangla"	"bangla" or "english"
grouping	str	"lakh"	"lakh", "western", or "none"

Raises FormatError on invalid arguments.

Exceptions

Exception	Base	When
DoshomikError	Exception	Base class
ParseError	DoshomikError	Input cannot be parsed
FormatError	DoshomikError	Arguments are invalid

Contributing

git clone https://github.com/rayhan-mahmuud/doshomik
cd doshomik
uv sync --all-groups

# Run the full test suite (pytest + coverage)
uv run pytest

# Type check
uv run mypy

# Lint and format
uv run ruff check src tests
uv run ruff format src tests

Tests target Python 3.10–3.13. Property-based tests use Hypothesis.

Changelog

See CHANGELOG.md.

License

MIT — see LICENSE.