data_cleaning/make_messy.py

"""
make_messy.py — Reproducibly generate a realistically messy 500-row CSV.

Simulates the kind of "customer orders" export you actually receive from a
legacy CRM: mixed date formats, duplicate rows, casing/whitespace noise,
sentinel nulls, and amounts that are text, not numbers.

Every row is produced from seed=42, so re-running the script always yields
the exact same file. The output is deliberately written with `csv` (no
pandas) so the mess is preserved verbatim — e.g. "NEW YORK" stays uppercase.

Usage:
    python make_messy.py [output.csv]
"""

from __future__ import annotations

import csv
import datetime
import random
from pathlib import Path

SEED = 42
N_ROWS = 500

CITIES = [
    "New York", "Chicago", "Houston", "Phoenix", "Philadelphia",
    "San Antonio", "San Diego", "Dallas", "San Jose", "Austin",
    "Jacksonville", "Columbus", "Charlotte", "San Francisco", "Seattle",
    "Denver",
]

FIRST = ["James", "Mary", "John", "Patricia", "Robert", "Jennifer", "Michael",
         "Linda", "David", "Elizabeth", "William", "Barbara", "Richard",
         "Susan", "Joseph", "Jessica", "Thomas", "Sarah", "Charles", "Karen"]
LAST = ["Smith", "Johnson", "Williams", "Brown", "Jones", "Garcia", "Miller",
        "Davis", "Rodriguez", "Martinez", "Hernandez", "Lopez", "Gonzalez",
        "Wilson", "Anderson", "Thomas", "Taylor", "Moore", "Jackson", "Martin"]

PAYMENTS = ["visa", "mastercard", "amex", "paypal", "bank_transfer"]
STATUSES = ["pending", "shipped", "delivered", "cancelled", "refunded"]

DATE_FORMATS = [
    lambda d: d.strftime("%Y-%m-%d"),        # 2024-03-01
    lambda d: d.strftime("%m/%d/%Y"),        # 03/01/2024
    lambda d: d.strftime("%d/%m/%Y"),        # 01/03/2024  (ambiguous!)
    lambda d: d.strftime("%b %d, %Y"),       # Mar 01, 2024
    lambda d: d.strftime("%d-%b-%Y"),        # 01-Mar-2024
    lambda d: d.strftime("%Y/%m/%d"),        # 2024/03/01
]


def main(out_path: Path) -> None:
    rng = random.Random(SEED)
    base_days = datetime.date(2024, 1, 1).toordinal()

    rows: list[list[str]] = []
    for i in range(1, N_ROWS + 1):
        first, last = rng.choice(FIRST), rng.choice(LAST)
        name = f"{first} {last}"
        # casing + whitespace noise in names (20%)
        r = rng.random()
        if r < 0.06:
            name = name.upper()
        elif r < 0.12:
            name = name.lower()
        elif r < 0.16:
            name = f"  {name} "
        elif r < 0.20:
            name = name.replace(" ", "  ")

        email_base = f"{first}.{last}".lower()
        domain = rng.choice(["gmail.com", "example.com", "outlook.com",
                             "yahoo.com", "company.org"])
        email = f"{email_base}{rng.randint(1, 999)}@{domain}"
        r = rng.random()
        if r < 0.05:
            email = email.upper()
        elif r < 0.09:
            email = f" {email} "

        city = rng.choice(CITIES)
        r = rng.random()
        if r < 0.08:
            city = city.upper()
        elif r < 0.14:
            city = city.lower()
        elif r < 0.20:
            city = f" {city} "

        day = rng.randint(0, 364)
        d = datetime.date.fromordinal(base_days + day)
        fmt = rng.choice(DATE_FORMATS)
        order_date = fmt(d)

        qty = rng.randint(1, 12)
        quantity = str(qty)
        if rng.random() < 0.05:
            quantity = f"{qty}.0"              # float-string where int wanted
        if rng.random() < 0.04:
            quantity = f" {quantity}"          # stray space

        unit_price = round(rng.uniform(5.0, 500.0), 2)
        if rng.random() < 0.10:
            unit_price_s = f"${unit_price:,.2f}"   # $1,234.56 as text
        else:
            unit_price_s = f"{unit_price:.2f}"

        total = round(unit_price * qty, 2)
        r = rng.random()
        if r < 0.08:
            total_s = f"${total:,.2f}"
        elif r < 0.12:
            total_s = f"({total:,.2f})"          # accounting negatives-as-parens
            # (kept positive here but formatted like a credit)
        else:
            total_s = f"{total:.2f}"

        payment = rng.choice(PAYMENTS)
        if rng.random() < 0.06:
            payment = payment.upper()
        if rng.random() < 0.05:
            payment = f" {payment} "

        status = rng.choice(STATUSES)
        r = rng.random()
        if r < 0.07:
            status = status.upper()
        elif r < 0.12:
            status = status.title() + " "       # trailing space

        # Nulls: use a mix of true empty strings and sentinel tokens.
        def maybe_null(v: str, p: float = 0.04) -> str:
            if rng.random() < p:
                return rng.choice(["", "", "N/A", "n/a", "NA", "?"])
            return v

        oid = f"ORD-{10000 + i}"
        if rng.random() < 0.03:
            oid = oid.replace("-", " ")        # ORD 10001
        if rng.random() < 0.03:
            oid = f" {oid}"

        rows.append([
            maybe_null(oid, 0.01),
            name,
            maybe_null(email, 0.05),
            maybe_null(city, 0.05),
            maybe_null(order_date, 0.03),
            quantity,
            unit_price_s,
            total_s,
            maybe_null(payment, 0.04),
            status,
        ])

    # --- Inject duplicates -------------------------------------------------
    # Exact duplicates (5%) and near-duplicates (3%: same order_id, messy
    # casing on a field).
    exact_idx = rng.sample(range(len(rows)), 25)
    near_idx = rng.sample([j for j in range(len(rows)) if j not in exact_idx],
                          15)
    for j in exact_idx:
        rows.append(list(rows[j]))
    for j in near_idx:
        dup = list(rows[j])
        dup[1] = dup[1].strip().lower()        # name casing differs
        rows.append(dup)

    rng.shuffle(rows)

    header = ["order_id", "customer_name", "email", "city", "order_date",
              "quantity", "unit_price", "total_amount", "payment_method",
              "status"]
    with out_path.open("w", newline="", encoding="utf-8") as fh:
        w = csv.writer(fh)
        w.writerow(header)
        w.writerows(rows)

    print(f"Wrote {len(rows)} data rows -> {out_path}")


if __name__ == "__main__":
    import sys
    target = Path(sys.argv[1]) if len(sys.argv) > 1 else Path(
        "messy_orders.csv")
    main(target)