merchant_map: +9 rules from May 2026 import learning loop

New merchants: Hobby Lobby, Pampurred Paws, Bojangles, Pittsburgh Zoo;
consolidations to existing BP/Exxon/Charlotte Motor Speedway; income
mappings for Emerson (paycheck) and Apple Cashback. Fix stale Red Robin
auto-tail rule (was spawning "Red Robin No" instead of consolidating).

Add retirement_projection.py + fire_projection.py: today's-dollars
compounding + FIRE/Coast-FIRE models with $50k spending floor.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

fire_projection.py

@@ -0,0 +1,199 @@
+#!/usr/bin/env python3
+"""
+FIRE / Coast-FIRE projection with a realistic rising-income career arc.
+
+Key upgrade over retirement_projection.py: income is NOT flat. It follows a
+career arc, and savings is a PERCENTAGE of income — so both your nest egg AND
+your lifestyle scale as you earn more. This is what lets us show the real
+trade: a high savings rate hits FIRE faster BOTH because you save more AND
+because you need less (lower spending => lower 25x FIRE number).
+
+All figures in TODAY'S DOLLARS (real returns).
+
+    /tmp/retire-venv/bin/python fire_projection.py
+"""
+
+import matplotlib
+matplotlib.use("Agg")
+import matplotlib.pyplot as plt
+from matplotlib.ticker import FuncFormatter
+
+# ----------------------------------------------------------------------------
+# ASSUMPTIONS
+# ----------------------------------------------------------------------------
+REAL_RETURN     = 0.07
+SWR             = 0.04          # 4% rule  ->  FIRE number = 25x annual spending
+START_AGE       = 25
+END_AGE         = 65
+START_PORTFOLIO = 48_000        # Roth IRA + Schwab Stocks today
+PIPELINE        = {25: 7_000, 26: 7_000, 27: 7_000, 28: 3_900}  # 529->Roth
+
+STIPEND         = 34_000        # grad years (age 25-26)
+START_SALARY    = 130_000       # industry, starting age 27
+TAX_GRAD        = 0.08          # effective all-in tax rate, grad years
+TAX_INDUSTRY    = 0.30          # effective all-in tax rate, industry (fed+PA+FICA+local)
+SPENDING_FLOOR  = 50_000        # Dane's rule: never live on less than $50k/yr (today's $)
+
+def gross_income(age):
+    """Real (today's $) gross income by age: stipend -> industry w/ tapering raises."""
+    if age <= 26:
+        return STIPEND
+    income = START_SALARY
+    for a in range(27, age):           # apply raises year over year up to `age`
+        if a < 40:   income *= 1.03    # 3% real ascent, early career
+        elif a < 50: income *= 1.01    # 1% real, mid career
+        else:        income *= 1.005   # ~flat, late career
+    return income
+
+def tax_rate(age):
+    return TAX_GRAD if age <= 26 else TAX_INDUSTRY
+
+# ----------------------------------------------------------------------------
+# SAVINGS STRATEGIES  — each returns SAVINGS DOLLARS given (age, gross, net).
+# The $50k spending floor is enforced: never save so much that you live on <$50k.
+# ----------------------------------------------------------------------------
+def hybrid_floor(age, gross, net):
+    """Front-load HARD ages 27-32 (live at the $50k floor, bank everything else),
+    then deliberately loosen to ~20% of gross — the Corvette-and-hobbies years."""
+    if age <= 26:
+        return 0.0
+    if age <= 32:                              # PHASE 1: live at floor, save the rest
+        return max(0.0, net - SPENDING_FLOOR)
+    target = gross * 0.20                       # PHASE 2: loosen to 20% of gross
+    return min(target, max(0.0, net - SPENDING_FLOOR))
+
+def steady_aggressive(age, gross, net):
+    if age <= 26:
+        return 0.0
+    return min(gross * 0.30, max(0.0, net - SPENDING_FLOOR))
+
+def balanced(age, gross, net):
+    if age <= 26:
+        return 0.0
+    return min(gross * 0.18, max(0.0, net - SPENDING_FLOOR))
+
+STRATEGIES = {
+    "Hybrid: hard 27-32 @ $50k floor, then loosen to 20%": (hybrid_floor, "#d93025"),
+    "Steady aggressive (30%)":                              (steady_aggressive, "#188038"),
+    "Balanced (18%)":                                       (balanced, "#1a73e8"),
+}
+
+# ----------------------------------------------------------------------------
+# SIMULATE
+# ----------------------------------------------------------------------------
+def simulate(save_fn):
+    ages, portfolio, spending, fire_num = [], [], [], []
+    bal = START_PORTFOLIO
+    fire_age = None
+    for age in range(START_AGE, END_AGE + 1):
+        gross = gross_income(age)
+        net = gross * (1 - tax_rate(age))
+        own_savings = save_fn(age, gross, net)
+        spend = net - own_savings                          # what you live on
+        fnum = 25 * spend                                   # FIRE target for THIS lifestyle
+
+        ages.append(age); portfolio.append(bal)
+        spending.append(spend); fire_num.append(fnum)
+        if fire_age is None and bal >= fnum and age >= 27:
+            fire_age = age
+
+        contrib = own_savings + PIPELINE.get(age, 0)
+        bal = bal * (1 + REAL_RETURN) + contrib
+    return dict(ages=ages, portfolio=portfolio, spending=spending,
+                fire_num=fire_num, fire_age=fire_age)
+
+results = {label: simulate(fn) for label, (fn, _) in STRATEGIES.items()}
+
+# ----------------------------------------------------------------------------
+# PLOT  (2 panels: wealth+FIRE, then lifestyle)
+# ----------------------------------------------------------------------------
+fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10.5), height_ratios=[1.6, 1])
+
+money_fmt = FuncFormatter(lambda v, _: f"${v/1e6:.1f}M" if v >= 1e6 else f"${v/1e3:.0f}k")
+
+# --- Panel 1: portfolio vs FIRE number ---
+for label, (fn, color) in STRATEGIES.items():
+    r = results[label]
+    ax1.plot(r["ages"], r["portfolio"], color=color, linewidth=2.6, label=label)
+    # FIRE number (dashed, same color, thin)
+    ax1.plot(r["ages"], r["fire_num"], color=color, linewidth=1.1,
+             linestyle=":", alpha=0.7)
+    # mark FIRE achievement
+    fa = r["fire_age"]
+    if fa:
+        idx = r["ages"].index(fa)
+        ax1.scatter([fa], [r["portfolio"][idx]], color=color, s=90, zorder=5,
+                    edgecolor="white", linewidth=1.5)
+        ax1.annotate(f"FI at {fa}", xy=(fa, r["portfolio"][idx]),
+                     xytext=(fa - 1.5, r["portfolio"][idx] + 0.35e6),
+                     fontsize=9.5, fontweight="bold", color=color)
+
+ax1.axvline(27, color="#5f6368", linestyle="--", alpha=0.5, linewidth=1.2)
+ax1.annotate("PhD done / industry income (~2028)", xy=(27, 0),
+             xytext=(27.3, 5.0e6), fontsize=9, color="#5f6368")
+# lean-FI floor: $50k lifestyle => $1.25M
+ax1.axhline(25 * SPENDING_FLOOR, color="#f9ab00", linestyle="-.", alpha=0.8, linewidth=1.4)
+ax1.annotate(f"Lean-FI floor: 25 x $50k = ${25*SPENDING_FLOOR/1e6:.2f}M",
+             xy=(START_AGE, 25*SPENDING_FLOOR), xytext=(START_AGE+0.3, 25*SPENDING_FLOOR+0.18e6),
+             fontsize=9, color="#b06000", fontweight="bold")
+ax1.set_title("FIRE trajectories — solid = portfolio, dotted = FIRE number (25x spend)",
+              fontsize=13, fontweight="bold")
+ax1.set_ylabel("Today's $")
+ax1.yaxis.set_major_formatter(money_fmt)
+ax1.set_xlim(START_AGE, END_AGE)
+ax1.set_ylim(bottom=0)
+ax1.grid(True, alpha=0.25)
+ax1.legend(loc="upper left", fontsize=9.5)
+
+# --- Panel 2: lifestyle (annual spending) ---
+for label, (fn, color) in STRATEGIES.items():
+    r = results[label]
+    ax2.plot(r["ages"], r["spending"], color=color, linewidth=2.4, label=label)
+ax2.set_title("What you actually live on each year (annual spending, today's $)",
+              fontsize=13, fontweight="bold")
+ax2.set_xlabel("Age")
+ax2.set_ylabel("Annual spending")
+ax2.yaxis.set_major_formatter(money_fmt)
+ax2.set_xlim(START_AGE, END_AGE)
+ax2.set_ylim(bottom=0)
+ax2.grid(True, alpha=0.25)
+ax2.axhline(STIPEND*(1-TAX_GRAD), color="#9aa0a6", linestyle="--", alpha=0.6)
+ax2.annotate("~current grad take-home", xy=(50, STIPEND*(1-TAX_GRAD)),
+             xytext=(50, STIPEND*(1-TAX_GRAD)+2500), fontsize=8.5, color="#9aa0a6")
+# the $50k floor — the line Dane won't cross
+ax2.axhline(SPENDING_FLOOR, color="#f9ab00", linestyle="-.", alpha=0.9, linewidth=1.6)
+ax2.annotate("$50k floor — won't live below this", xy=(40, SPENDING_FLOOR),
+             xytext=(40, SPENDING_FLOOR+3000), fontsize=9, color="#b06000", fontweight="bold")
+ax2.legend(loc="upper left", fontsize=9.5)
+
+fig.tight_layout()
+out = "/Users/danesabo/Documents/Finances/fire_projection.png"
+fig.savefig(out, dpi=140, bbox_inches="tight")
+
+# ----------------------------------------------------------------------------
+# TEXT SUMMARY
+# ----------------------------------------------------------------------------
+print("=" * 72)
+print("FIRE PROJECTION  (today's dollars, 7% real, rising-income career arc)")
+print("=" * 72)
+print(f"Income arc: ${STIPEND:,} stipend -> ${START_SALARY:,} at 27, "
+      f"3%/1%/0.5% real raises\n")
+LEAN_FI = 25 * SPENDING_FLOOR   # $1.25M -> permanently able to fall back to $50k
+for label in STRATEGIES:
+    r = results[label]
+    fa = r["fire_age"]
+    # spending + portfolio at a few ages
+    def at(age): return r["portfolio"][r["ages"].index(age)]
+    def sp(age): return r["spending"][r["ages"].index(age)]
+    # age at which portfolio first crosses the lean-FI safety net ($1.25M)
+    lean_age = next((a for a, p in zip(r["ages"], r["portfolio"])
+                     if p >= LEAN_FI and a >= 27), None)
+    print(label)
+    print(f"   $50k safety net funded (cross ${LEAN_FI/1e6:.2f}M): "
+          f"{'age '+str(lean_age) if lean_age else 'not by 65'}")
+    print(f"   Full FIRE on actual lifestyle: {'age '+str(fa) if fa else 'not by 65'}")
+    print(f"   Lifestyle (annual spend) at 30/40/50: "
+          f"${sp(30)/1e3:.0f}k / ${sp(40)/1e3:.0f}k / ${sp(50)/1e3:.0f}k")
+    print(f"   Portfolio at 50 / 65: ${at(50)/1e6:.2f}M / ${at(65)/1e6:.2f}M")
+    print()
+print(f"Chart saved: {out}")

migration/README.md

@@ -0,0 +1,107 @@
+# Firefly rebuild runbook
+
+One-time migration: wipe the CSV-era transactions and rebuild from
+FITID-stable QFX so every transaction has a permanent dedup key and a clean
+account taxonomy. Read this before running anything in this folder.
+
+## Why a rebuild (not in-place cleanup)
+
+Firefly history is young (everything ~Aug 2025+, ~950 txns, minimal manual
+data). Old CSV imports left ~343 fragmented junk expense accounts and no
+stable external_ids. A clean rebuild keyed on QFX `FITID` is a better
+foundation than reassigning junk in place. Decided 2026-05-17.
+
+## Hard prerequisites (do not skip)
+
+1. **Firefly DB backup.** Destructive, no undo. Do not run the wipe until a
+   DB dump/snapshot exists.
+2. **Exports** (in `../EXPORTS/`, gitignored): Apple/PNC/Costco QFX, Aug 2025
+   -> now, FITID on 100% of rows. Schwab/Coinbase/Cash (~35 txns) are
+   CSV-only/manual, handled separately.
+
+## Reconciliation (the trust gate)
+
+Per account: `opening_balance = QFX_ledger - sum(all that account's lines)`.
+Classification (transfer vs expense) never changes an account's own balance,
+so `opening + sum == ledger` must hold to the cent before trusting the wipe.
+Verified: PNC opening $6,866.10, Apple -$4,498.79, Costco -$2,541.57 (all
+tie). `rebuild_dryrun.py` recomputes this; re-run after any change.
+
+## Classification rules (PNC = the hub)
+
+- **Transfers** -- ALWAYS owned by the PNC leg: PNC's posting date and PNC's
+  FITID are authoritative, the card/brokerage counterpart line is paired by
+  amount (+/- a few days) and dropped. Every transfer lives under PNC, one
+  consistent date, never double-counted. Pairs: APPLECARD GSBANK -> Apple
+  Credit Card; CITI AUTOPAY -> Costco Visa Card; SCHWAB MONEYLINK -> Schwab
+  Stocks/Savings (disambiguate by amount); ATM WITHDRAWAL -> Cash; CARVANA
+  PAYOUT -> Illiquid Assets; big ATM DEPOSIT -> Coverdell; CAPITAL ONE ->
+  Capital One (closed). Codified in the skill's `references/transfers.md`.
+- **Income/expense**: Pitt salary -> Wages; Duquesne Light -> Utilities:
+  Electric; Compeer -> Rent; etc.
+- **Don't Know**: Venmo/CashApp/Zelle ("poker"), unrecallable checks, unknown
+  ATM deposits -> the `Don't Know` account, review later. Never guessed.
+- **Special accounts**: `Illiquid Assets` (cars; sale = transfer in),
+  `Don't Know` (catch-all). See the skill's memory / taxonomy notes.
+
+## Investment accounts
+
+Do NOT transaction-import Schwab/Roth/Coverdell/Coinbase (noise, and assets
+!= currency). Model as monthly-valued: opening balance + external MoneyLink
+transfers (from the PNC side) + one monthly valuation adjustment booked to
+`Investment Appreciation` / `Investment: Interest`. Dane supplies the current
+value at import; delta = the adjustment. Savings<->Stocks journals are
+transfers.
+
+## Execution order
+
+1. `python rebuild_dryrun.py` -> confirm all accounts still reconcile.
+2. Build the full normalized dataset (PNC + Apple + Costco, transfers typed,
+   payments paired/deduped, opening balances set).
+3. Drive review via the skill's browser workflow
+   (`references/review-workflow.md`): `--review-html`, resolve the ~190 tail
+   merchants in-situ (search-then-ask, <80% => ask), Export `decisions.json`.
+4. **Confirm DB backup exists.**
+5. Wipe transactions, prune empty junk expense accounts.
+6. `--decisions decisions.json --post`. Reconcile final balances against the
+   derived figures above.
+
+## Files here
+
+- `rebuild_pnc.py` -- PNC classifier + reconciliation (read-only)
+- `rebuild_dryrun.py` -- consolidated per-account reconciliation (read-only)
+- `pnc_classified.json` -- PNC classification output
+- `merchant_clusters.{json,md}` -- cluster proposal (taxonomy bootstrap)
+- `mock_firefly.py` -- stdlib mock used for skill eval/testing
+- `*review_preview*.html` -- review-UI previews on real data
+
+Nothing here writes to Firefly except the final `--post` in step 6.
+
+## Lessons from the first rebuild (2026-05-20)
+
+Captured here so a second rebuild doesn't re-discover them.
+
+- **Orphan paired transfers**: the PNC->Apple payment from 2025-08-01 has no
+  Apple-side line (Apple's QFX starts 08-02). Its effect was already in
+  Apple's derived opening; posting the transfer ALSO crediting Apple
+  double-counted by $3,218. Fix: `build_rebuild_dataset.py` now subtracts
+  orphan transfer amounts from the destination card's opening. See
+  `references/transfers.md` in the skill.
+- **Asset accounts require `account_role`** on POST /accounts. `defaultAsset`
+  works universally.
+- **Budgets do not auto-create.** If wiping to scratch, recreate Needs /
+  Wants / Savings via UI or POST before the import.
+- **Wipe via UI leaves stale revenue accounts / categories** (only
+  transaction-referenced asset accounts go). Prune manually if you want a
+  truly clean slate.
+- **Strip cached `account_id` from `merchant_map.json` before any rebuild.**
+  Pre-wipe ids are invalid post-wipe. The skill no longer caches to the map
+  (in-memory only) but old maps may still carry stale ids.
+- **Background Python with `nohup ... &` can lose stdout to buffering.** Use
+  `python -u` for the import step. The first rebuild's log was empty because
+  Python buffered everything and we mistook it for "ran but did nothing."
+- **`error_if_duplicate_hash` is now off** — Firefly's content-hash dedup
+  was too eager (rejected legit-distinct rows with same date+amt+desc, like
+  two parking sessions same garage). `external_id` precheck is the only dedup.
+- **Wipe by deleting transactions, not by deleting accounts.** Otherwise you
+  end up with stale ids referenced by merchant_map cache.

migration/build_rebuild_dataset.py

@@ -0,0 +1,134 @@
+"""Build the full rebuild dataset from the 3 QFX (READ-ONLY).
+
+Emits one normalized.json (the skill's schema) for ALL of PNC + Apple +
+Costco, with:
+  - transfers OWNED BY THE PNC LEG (PNC date + FITID authoritative); the
+    Apple PAYMENT lines and Costco positive AUTOPAY lines are the
+    counterparts and are DROPPED (paired by amount, +/- 6 days).
+  - PNC classified per the runbook (income / expense / Don't Know / special).
+  - Apple/Costco: negative = withdrawal (merchant), positive = deposit
+    (refund). merchant_map matching is left to firefly_import.py downstream.
+  - per-account reconciliation: opening + sum(its kept lines) must == QFX
+    ledger, else abort (no silent data loss).
+
+Nothing is posted. Output feeds `firefly_import.py --emit-plan/--review-html`.
+"""
+import re, json, hashlib, sys, html
+from collections import Counter
+
+D = "/Users/danesabo/Documents/Finances/EXPORTS/-MAY172026"
+SRC = {
+    "PNC Checking":      (f"{D}/PNC7552Aug012025-May152025.QFX",       "pnc"),
+    "Apple Credit Card": (f"{D}/Apple Card Transactions Aug 01 2025 - May 17 2026.qfx", "apple"),
+    "Costco Visa Card":  (f"{D}/CitiCostcoCard Aug012025-May172025.QFX","costco"),
+}
+
+def parse(path):
+    t = open(path, encoding="latin-1", errors="replace").read()
+    m = re.search(r"<LEDGERBAL>.*?<BALAMT>([^<\r\n]*)", t, re.S | re.I)
+    ledger = float(m.group(1))
+    blocks = re.findall(r"<STMTTRN>(.*?)(?=<STMTTRN>|</BANKTRANLIST>)", t, re.S | re.I)
+    def g(b, k):
+        mm = re.search(rf"<{k}>([^<\r\n]*)", b, re.I)
+        return mm.group(1).strip() if mm else ""
+    out = []
+    for b in blocks:
+        out.append({"date": g(b, "DTPOSTED")[:8], "amt": float(g(b, "TRNAMT")),
+                    "ttype": g(b, "TRNTYPE").upper(),
+                    "desc": html.unescape((g(b, "NAME") + " " + g(b, "MEMO")).strip()),
+                    "fitid": g(b, "FITID")})
+    return ledger, out
+
+def iso(d):  # YYYYMMDD -> YYYY-MM-DD
+    return f"{d[:4]}-{d[4:6]}-{d[6:8]}" if len(d) >= 8 else d
+
+# ---- PNC classification (runbook) ---------------------------------------
+def classify_pnc(desc, amt):
+    d = desc.upper()
+    if "APPLECARD GSBANK PAYMENT" in d:   return ("transfer", "Apple Credit Card")
+    if "CITI AUTOPAY PAYMENT" in d:       return ("transfer", "Costco Visa Card")
+    if "SCHWAB BROKERAGE MONEYLINK" in d:
+        # amount disambiguation per the Schwab JSONs
+        return ("transfer", "Schwab Savings" if abs(amt) in (5000.0, 3550.0)
+                              else "Schwab Stocks")
+    if "ATM WITHDRAWAL" in d:             return ("transfer", "Cash")
+    if "CARVANA PAYOUT" in d:             return ("transfer", "Illiquid Assets")
+    if "ATM DEPOSIT" in d and abs(amt) > 10000: return ("transfer", "Coverdell")
+    if "CAPITAL ONE TRANSFER" in d:       return ("transfer", "Capital One")
+    if "UNIV PITTSBURGH" in d and ("PAYROLL" in d or "SALARY" in d):
+        return ("deposit", "Pitt Salary")
+    if "INTEREST PAYMENT" in d:           return ("deposit", "Interest Income")
+    if "IRS TREAS 310" in d:              return ("deposit", "IRS Refund")
+    if "DUQUESNE LIGHT" in d:             return ("withdrawal", "Duquesne Light")
+    if "COMPEER" in d:                    return ("withdrawal", "Compeer Investments")
+    if "PITT TUITION" in d:               return ("withdrawal", "University of Pittsburgh")
+    if any(k in d for k in ("VENMO CASHOUT","CASH APP","ZEL FROM","ATM SURCHARGE","YARDI")):
+        return ("dontknow", "Don't Know")
+    return ("raw", None)   # leave to merchant_map / review downstream
+
+records, recon, dropped = [], {}, Counter()
+for acct, (path, tag) in SRC.items():
+    ledger, txns = parse(path)
+    s = round(sum(t["amt"] for t in txns), 2)
+    opening = round(ledger - s, 2)
+    recon[acct] = {"ledger": ledger, "sum": s, "opening": opening,
+                   "ties": abs(opening + s - ledger) < 0.01}
+    for t in txns:
+        amt, d = t["amt"], t["desc"]
+        ext = f"{tag}:{t['fitid'] or hashlib.sha1((iso(t['date'])+d+str(amt)).encode()).hexdigest()[:16]}"
+        if acct == "Apple Credit Card" and t["ttype"] == "PAYMENT":
+            dropped["apple_payment(paired->PNC)"] += 1; continue
+        if acct == "Costco Visa Card" and amt > 0 and "AUTOPAY" in d.upper():
+            dropped["costco_autopay(paired->PNC)"] += 1; continue
+        rec = {"date": iso(t["date"]), "amount": f"{abs(amt):.2f}",
+               "description": d, "asset_account": acct, "source_tag": tag,
+               "source_txn_id": t["fitid"] or None, "currency_code": "USD"}
+        if acct == "PNC Checking":
+            kind, target = classify_pnc(d, amt)
+            if kind == "transfer":
+                rec["type"] = "transfer"
+                if amt < 0: rec["destination_account"] = target
+                else:       rec["type"] = "transfer"; rec["asset_account"] = target; rec["destination_account"] = "PNC Checking"
+            elif kind in ("deposit", "withdrawal"):
+                rec["type"] = kind; rec["_canonical"] = target
+            elif kind == "dontknow":
+                rec["type"] = "withdrawal" if amt < 0 else "deposit"
+                rec["_canonical"] = "Don't Know"
+            else:
+                rec["type"] = "withdrawal" if amt < 0 else "deposit"
+        else:
+            rec["type"] = "withdrawal" if amt < 0 else "deposit"
+        records.append(rec)
+
+# --- Orphan adjustment: a PNC->Apple/Costco payment whose date predates the
+# card QFX window has its card-side effect already baked into the card's
+# DERIVED opening (because opening = ledger - sum_kept_card_lines, and the
+# orphan never appeared on the card side). If we ALSO post the PNC->card
+# transfer in the rebuild, the card account gets credited twice. So subtract
+# orphan transfer amounts from the card opening.
+APPLE_WINDOW_START = "2025-08-02"
+COSTCO_WINDOW_START = "2025-08-02"
+for r in records:
+    if r.get("type") == "transfer" and r["asset_account"] == "PNC Checking":
+        dest = r.get("destination_account")
+        if dest == "Apple Credit Card" and r["date"] < APPLE_WINDOW_START:
+            recon["Apple Credit Card"]["opening"] -= float(r["amount"])
+            recon["Apple Credit Card"]["opening"] = round(recon["Apple Credit Card"]["opening"], 2)
+        elif dest == "Costco Visa Card" and r["date"] < COSTCO_WINDOW_START:
+            recon["Costco Visa Card"]["opening"] -= float(r["amount"])
+            recon["Costco Visa Card"]["opening"] = round(recon["Costco Visa Card"]["opening"], 2)
+
+print("=== RECONCILIATION (must all tie) ===")
+ok = True
+for a, r in recon.items():
+    flag = "OK" if r["ties"] else "*** MISMATCH ***"
+    ok &= r["ties"]
+    print(f"  {a:20} ledger {r['ledger']:>11,.2f}  Σ {r['sum']:>11,.2f}  "
+          f"opening {r['opening']:>11,.2f}  {flag}")
+print("dropped (paired counterparts):", dict(dropped))
+print(f"normalized records: {len(records)}")
+if not ok:
+    print("ABORT: a reconciliation does not tie.", file=sys.stderr); sys.exit(1)
+json.dump(records, open("/tmp/rebuild_normalized.json", "w"), indent=1)
+json.dump(recon, open("/tmp/rebuild_recon.json", "w"), indent=1)
+print("wrote /tmp/rebuild_normalized.json")

migration/rebuild_clusters.md

@@ -0,0 +1,192 @@
+# Merchant cluster proposal
+
+- 386 clusters from 372 accounts + 1017 statement txns
+- **142** auto-proposable (>=0.80, clean canonical)
+- **244** NEED DANE (ambiguous / junky canonical / new merchant)
+
+## NEEDS DANE (top 40 by volume)
+_For each: what is the real merchant? You can type a name; it becomes a permanent rule._
+
+- **?** (conf 0.57, weight 75, 28 accts, 47 stmt) guess=`Amazon`
+    - desc: `AMAZON MARK* B00SF6VV0410 TERRY`
+    - desc: `AMAZON.COM*9R3UC0N93 440 TERRY A`
+    - desc: `AMAZON.COM*N428X9Q71 440 TERRY A`
+    - desc: `AMAZON MARK* B008Z3VV0410 TERRY`
+    - desc: `AMAZON MARK* B03Y156K1410 TERRY`
+    - desc: `AMAZON MARK* B204T9M31410 TERRY`
+    - accts: Amazon, Amazon Mark* B008z3vv0, Amazon Mark* B00sf6vv0, Amazon Mark* B00sf6vv0410 Terry Avenue North Seattle 98109 Wa Usa (return), Amazon Mark* B03y156k1, Amazon Mark* B204t9m31
+- **?** (conf 0.4, weight 56, 0 accts, 56 stmt) guess=`University Of Pittsburgh|Pitt Parking Pay Stati127 North`
+    - desc: `PITT PARKING PAY STATI127 NORTH`
+- **?** (conf 0.78, weight 37, 7 accts, 30 stmt) guess=`McDonald's`
+    - desc: `MCDONALDS 1862 3708 FORBES AVE P`
+    - desc: `MCDONALDS 1102 225 MOUNT LEBANON`
+    - desc: `MCDONALD'S F1862 3708 FORBES AVE`
+    - desc: `MCDONALD'S F1102 225 MT LEBANON`
+    - desc: `MCDONALDS 5834 2518 W LIBERTY RD`
+    - desc: `MCDONALD'S F27387 1412 B MAIN ST`
+    - accts: McDonald's, Mcdonald's F1102, Mcdonald's F1862, Mcdonald's F27387, Mcdonalds 1862, Mcdonalds 33234
+- **?** (conf 1.0, weight 30, 0 accts, 30 stmt) guess=`Castle Shannon Shop`
+    - desc: `CASTLE SHANNON SHOP' 799 CASTLE`
+- **?** (conf 0.71, weight 30, 2 accts, 28 stmt) guess=`Market District`
+    - desc: `MARKET DISTRICT #0014 7000 OXFOR`
+    - desc: `MARKET DISTRICT #0047 100 SETTLE`
+    - accts: Market District, Market District Supermarket
+- **?** (conf 0.4, weight 18, 0 accts, 18 stmt) guess=`Apple Com Bill One Apple`
+    - desc: `APPLE.COM/BILL ONE APPLE PARK WA`
+    - desc: `APPLE.COM/US ONE APPLE PARK WAY`
+    - desc: `APPLE.COM/BILL ONE APPLE PARK CU`
+- **?** (conf 0.47, weight 18, 8 accts, 10 stmt) guess=`Compeer`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB C5R6`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB MD64`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB 3Y6Q`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB R34S`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB D9FZ`
+    - desc: `COMPEER-COMP-CP WEB PMTS ACH WEB COMPEER-COMP-CP WEB PMTS ACH WEB F394`
+    - accts: COMPEER-COMP-CP WEB PMTS ACH WEB 3Y6QDL, COMPEER-COMP-CP WEB PMTS ACH WEB 7Y648K, COMPEER-COMP-CP WEB PMTS ACH WEB D9FZ0L, COMPEER-COMP-CP WEB PMTS ACH WEB F394TK, COMPEER-COMP-CP WEB PMTS ACH WEB JS0NNK, COMPEER-COMP-CP WEB PMTS ACH WEB K7TDFK
+- **?** (conf 0.4, weight 18, 0 accts, 18 stmt) guess=`Sq *La Gourmandine Oak116 Meyran`
+    - desc: `SQ *LA GOURMANDINE OAK116 MEYRAN`
+- **?** (conf 1.0, weight 17, 0 accts, 17 stmt) guess=`Kuhns Banksville`
+    - desc: `KUHNS BANKSVILLE 3125 BANKSVILLE`
+- **?** (conf 0.75, weight 13, 4 accts, 9 stmt) guess=`Starbucks`
+    - desc: `STARBUCKS STORE 27117 4022 FIFTH`
+    - desc: `STARBUCKS 27117 4022 5TH AVE PIT`
+    - desc: `STARBUCKS 8007827282 2401 UTAH A`
+    - accts: Starbucks, Starbucks 27117, Starbucks 8007827282, Starbucks Store 27117
+- **?** (conf 0.4, weight 11, 0 accts, 11 stmt) guess=`Claude Ai Subscription548 Market`
+    - desc: `CLAUDE.AI SUBSCRIPTION548 MARKET`
+- **?** (conf 0.61, weight 11, 2 accts, 9 stmt) guess=`Duquesne Light`
+    - desc: `DUQUESNE LIGHT PAYMENT ACH DEBIT DUQUESNE LIGHT PAYMENT ACH DEBIT xxxx`
+    - accts: DUQUESNE LIGHT PAYMENT ACH DEBIT xxxxxx5333, Duquesne Light
+- **?** (conf 0.4, weight 11, 1 accts, 10 stmt) guess=`T2`
+    - desc: `T2* MT LEBANON PA 8900 KEYSTONE`
+    - accts: T2* Mt Lebanon Pa
+- **?** (conf 1.0, weight 10, 1 accts, 9 stmt) guess=`Comcast / Xfinity`
+    - desc: `COMCAST / XFINITY 15 SUMMIT PARK`
+    - accts: Comcast / Xfinity
+- **?** (conf 1.0, weight 10, 0 accts, 10 stmt) guess=`Interest Payment Interest Payment`
+    - desc: `INTEREST PAYMENT INTEREST PAYMENT`
+- **?** (conf 0.4, weight 10, 0 accts, 10 stmt) guess=`Upmc Student Insurance600 Grant`
+    - desc: `UPMC STUDENT INSURANCE600 GRANT`
+- **?** (conf 0.4, weight 9, 0 accts, 9 stmt) guess=`Applecard Gsbank Payment Ach Web`
+    - desc: `APPLECARD GSBANK PAYMENT ACH WEB APPLECARD GSBANK PAYMENT ACH WEB-RECU`
+    - desc: `APPLECARD GSBANK PAYMENT ACH WEB APPLECARD GSBANK PAYMENT ACH WEB xxxx`
+- **?** (conf 0.4, weight 9, 0 accts, 9 stmt) guess=`Citi Autopay Payment Ach Web`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+    - desc: `CITI AUTOPAY PAYMENT ACH WEB-REC CITI AUTOPAY PAYMENT ACH WEB-RECUR xx`
+- **?** (conf 1.0, weight 9, 0 accts, 9 stmt) guess=`Daily Cash Adjustment`
+    - desc: `DAILY CASH ADJUSTMENT`
+- **?** (conf 1.0, weight 7, 0 accts, 7 stmt) guess=`Ebay O`
+    - desc: `EBAY O*07-14287-66191 2535 NORTH`
+    - desc: `EBAY O*07-14287-66190 2535 NORTH`
+    - desc: `EBAY O*07-14287-66189 2535 NORTH`
+    - desc: `EBAY O*07-14287-66188 2535 NORTH`
+    - desc: `EBAY O*07-14287-66187 2535 NORTH`
+    - desc: `EBAY O*07-14287-66186 2535 NORTH`
+- **?** (conf 0.9, weight 7, 1 accts, 6 stmt) guess=`Needle & Bean`
+    - desc: `SQ *NEEDLE & BEAN 320 CASTLE`
+    - accts: Needle & Bean
+- **?** (conf 0.4, weight 7, 0 accts, 7 stmt) guess=`University Of Pittsburgh|Univ Pittsburgh Salary Ach Credi`
+    - desc: `UNIV PITTSBURGH SALARY ACH CREDI UNIV PITTSBURGH SALARY ACH CREDIT xx0`
+- **?** (conf 1.0, weight 7, 0 accts, 7 stmt) guess=`Youtube Tv`
+    - desc: `GOOGLE *YOUTUBE TV 1600 AMPHITHE`
+- **?** (conf 0.62, weight 6, 2 accts, 4 stmt) guess=`Liberty Mutual`
+    - desc: `LIBERTY MUTUAL 175 BERKELEY ST 8`
+    - desc: `LIBERTY MUTUAL ATTN: COURTNEY MU`
+    - accts: Liberty Mutual
+- **?** (conf 0.53, weight 6, 2 accts, 4 stmt) guess=`Openai`
+    - desc: `OPENAI *CHATGPT SUBSCR548 MARKET`
+    - desc: `OPENAI 1455 3RD STREET SAN FRANC`
+    - accts: Openai, Openai *chatgpt Subscr
+- **?** (conf 0.4, weight 6, 0 accts, 6 stmt) guess=`Spo P&Amp Gspamelasdiner3703 F`
+    - desc: `SPO*P&G'SPAMELA'SDINER3703 F`
+- **?** (conf 1.0, weight 6, 2 accts, 4 stmt) guess=`Svdp Castle Shannon`
+    - desc: `SVDP CASTLE SHANNON 3423 LIBRARY`
+    - accts: SVDP Castle Shannon, Svdp Castle Shannon
+- **?** (conf 0.4, weight 5, 0 accts, 5 stmt) guess=`Bp 9604786Ukani Broqps2900 Banks`
+    - desc: `BP#9604786UKANI BROQPS2900 BANKS`
+- **?** (conf 0.4, weight 5, 2 accts, 3 stmt) guess=`Capital One Transfer Ach Web`
+    - desc: `CAPITAL ONE TRANSFER ACH WEB RT0 CAPITAL ONE TRANSFER ACH WEB RT0D854F`
+    - desc: `CAPITAL ONE TRANSFER ACH WEB PAY CAPITAL ONE TRANSFER ACH WEB PAYMENT `
+    - desc: `CAPITAL ONE TRANSFER ACH WEB PAY CAPITAL ONE TRANSFER ACH WEB PAYMENT `
+    - accts: CAPITAL ONE TRANSFER ACH WEB PAYMENT RT04E16C0EA8E68, CAPITAL ONE TRANSFER ACH WEB PAYMENT RT097FE1F911EB7
+- **?** (conf 0.69, weight 5, 1 accts, 4 stmt) guess=`Peacock`
+    - desc: `PEACOCK 75AE1 PREMIUM 30 ROCKEFE`
+    - desc: `PEACOCK 81D06 PREMIUM 30 ROCKEFE`
+    - desc: `PEACOCK EF701 PREMIUM 30 ROCKEFE`
+    - desc: `PEACOCK X6258 PREMIUM 30 ROCKEFE`
+    - accts: Peacock
+- **?** (conf 0.4, weight 5, 0 accts, 5 stmt) guess=`Spiegel Freedman Psych105 Braunl`
+    - desc: `SPIEGEL FREEDMAN PSYCH105 BRAUNL`
+- **?** (conf 0.4, weight 5, 0 accts, 5 stmt) guess=`University Of Pittsburgh|Rnk Pittsburgh P3610 Forbe`
+    - desc: `TST*RNK PITTSBURGH - P3610 FORBE`
+- **?** (conf 1.0, weight 5, 1 accts, 4 stmt) guess=`Www Costco Com`
+    - desc: `WWW COSTCO COM 800-955-2292`
+    - accts: WWW COSTCO COM 800-955-2292 WA
+- **?** (conf 0.4, weight 4, 0 accts, 4 stmt) guess=`Dave And Andy S Ho207`
+    - desc: `SQ *DAVE AND ANDY S HO207 ATWOOD`
+- **?** (conf 0.4, weight 4, 0 accts, 4 stmt) guess=`Enricos Tazza Do125 Lytton`
+    - desc: `SQ *ENRICO'S TAZZA D'O125 LYTTON`
+- **?** (conf 0.4, weight 4, 0 accts, 4 stmt) guess=`Hofbrauhaus Pittsburgh2705 S Wat`
+    - desc: `HOFBRAUHAUS PITTSBURGH2705 S WAT`
+- **?** (conf 0.65, weight 4, 1 accts, 3 stmt) guess=`Luis Benitez`
+    - desc: `ZEL FROM Luis Benitez ZEL FROM Luis Benitez`
+    - accts: Luis Benitez
+- **?** (conf 0.4, weight 4, 2 accts, 2 stmt) guess=`Pitt Tuition Pittpaymnt Ach Web`
+    - desc: `PITT TUITION PITTPAYMNT ACH WEB PITT TUITION PITTPAYMNT ACH WEB OPUxxx`
+    - desc: `PITT TUITION PITTPAYMNT ACH WEB PITT TUITION PITTPAYMNT ACH WEB OPUxxx`
+    - accts: PITT TUITION PITTPAYMNT ACH WEB OPUxxxx0412, PITT TUITION PITTPAYMNT ACH WEB OPUxxxx9683
+- **?** (conf 0.4, weight 4, 0 accts, 4 stmt) guess=`Schwab Brokerage Moneylink Ach W`
+    - desc: `SCHWAB BROKERAGE MONEYLINK ACH C SCHWAB BROKERAGE MONEYLINK ACH CREDIT`
+    - desc: `SCHWAB BROKERAGE MONEYLINK ACH C SCHWAB BROKERAGE MONEYLINK ACH CREDIT`
+    - desc: `SCHWAB BROKERAGE MONEYLINK ACH D SCHWAB BROKERAGE MONEYLINK ACH DEBIT `
+    - desc: `SCHWAB BROKERAGE MONEYLINK ACH W SCHWAB BROKERAGE MONEYLINK ACH WEB-RE`
+- **?** (conf 1.0, weight 4, 1 accts, 3 stmt) guess=`Subaru Of South Hills`
+    - desc: `SUBARU OF SOUTH HILLS 3260 WASHI`
+    - accts: Subaru Of South Hills
+
+## AUTO-PROPOSABLE (top 40 by volume)
+
+- `GomobilePGH`  (conf 1.0, weight 49, merges 4 accts) ids=[865, 642, 559, 781]
+- `Sheetz`  (conf 1.0, weight 43, merges 7 accts) ids=[566, 744, 739, 567, 774, 794, 738]
+- `Autozone`  (conf 1.0, weight 27, merges 6 accts) ids=[593, 812, 724, 714, 591, 806]
+- `Sunoco`  (conf 1.0, weight 27, merges 6 accts) ids=[599, 638, 827, 767, 820, 715]
+- `Costco Whse`  (conf 1.0, weight 22, merges 2 accts) ids=[842, 836]
+- `Harbor Freight Tools`  (conf 0.95, weight 18, merges 3 accts) ids=[878, 569, 737]
+- `Petco`  (conf 1.0, weight 15, merges 4 accts) ids=[546, 729, 797, 633]
+- `Chick-fil-A`  (conf 1.0, weight 14, merges 5 accts) ids=[630, 810, 832, 712, 702]
+- `Costco Gas`  (conf 1.0, weight 14, merges 2 accts) ids=[840, 837]
+- `D J*wsj`  (conf 1.0, weight 10, merges 1 accts) ids=[553]
+- `Rockauto`  (conf 0.94, weight 10, merges 1 accts) ids=[557]
+- `University Club`  (conf 0.86, weight 10, merges 2 accts) ids=[867, 637]
+- `Chikn Oakland`  (conf 1.0, weight 9, merges 1 accts) ids=[558]
+- `Raising Cane's`  (conf 1.0, weight 9, merges 3 accts) ids=[868, 561, 828]
+- `Barnes & Noble`  (conf 0.9, weight 7, merges 3 accts) ids=[603, 817, 658]
+- `Lowe's`  (conf 1.0, weight 7, merges 1 accts) ids=[673]
+- `PMUSA`  (conf 1.0, weight 7, merges 2 accts) ids=[885, 614]
+- `Home Depot`  (conf 0.83, weight 6, merges 1 accts) ids=[722]
+- `REI`  (conf 1.0, weight 6, merges 2 accts) ids=[684, 682]
+- `Target`  (conf 1.0, weight 6, merges 2 accts) ids=[605, 731]
+- `The Saloon Of`  (conf 0.82, weight 6, merges 2 accts) ids=[847, 801]
+- `Best Buy`  (conf 1.0, weight 5, merges 2 accts) ids=[751, 740]
+- `Check`  (conf 1.0, weight 5, merges 1 accts) ids=[524]
+- `Expedia`  (conf 1.0, weight 5, merges 2 accts) ids=[717, 711]
+- `Michaels Stores`  (conf 1.0, weight 5, merges 2 accts) ids=[587, 664]
+- `Rita's`  (conf 1.0, weight 5, merges 1 accts) ids=[882]
+- `Als Corner`  (conf 1.0, weight 4, merges 1 accts) ids=[762]
+- `CVS Pharmacy`  (conf 1.0, weight 4, merges 2 accts) ids=[783, 816]
+- `Dunkin`  (conf 1.0, weight 4, merges 2 accts) ids=[655, 846]
+- `Five Guys`  (conf 1.0, weight 4, merges 1 accts) ids=[723]
+- `Redhawk Coffee`  (conf 1.0, weight 4, merges 1 accts) ids=[721]
+- `Sportsmans Warehouse`  (conf 1.0, weight 4, merges 1 accts) ids=[568]
+- `Taco Bell`  (conf 1.0, weight 4, merges 2 accts) ids=[686, 691]
+- `TNT Pizza`  (conf 1.0, weight 4, merges 1 accts) ids=[624]
+- `Act Cntyalleghenyprk`  (conf 1.0, weight 3, merges 1 accts) ids=[776]
+- `Butterjoint`  (conf 1.0, weight 3, merges 1 accts) ids=[608]
+- `Ctlp*csc Serviceworks`  (conf 1.0, weight 3, merges 1 accts) ids=[650]
+- `Fiori's Pizzaria`  (conf 0.91, weight 3, merges 1 accts) ids=[551]
+- `Get Go`  (conf 1.0, weight 3, merges 1 accts) ids=[718]
+- `Giant Eagle`  (conf 1.0, weight 3, merges 1 accts) ids=[592]

retirement_projection.py

@@ -0,0 +1,144 @@
+#!/usr/bin/env python3
+"""
+Retirement trajectory projection for Dane.
+
+Models portfolio growth from age 25 -> 65 under three post-PhD savings
+scenarios, plus shows how negligible the one-time summer 401k contribution is.
+
+All figures are in TODAY'S DOLLARS (real returns), so they're directly
+comparable to current purchasing power. Edit the ASSUMPTIONS block and re-run.
+
+    /tmp/retire-venv/bin/python retirement_projection.py
+"""
+
+import matplotlib
+matplotlib.use("Agg")  # no display needed; save to file
+import matplotlib.pyplot as plt
+from matplotlib.ticker import FuncFormatter
+
+# ----------------------------------------------------------------------------
+# ASSUMPTIONS  (edit these and re-run)
+# ----------------------------------------------------------------------------
+REAL_RETURN      = 0.07     # 7% real (inflation-adjusted) annual return
+SWR              = 0.04     # 4% safe withdrawal rate
+START_AGE        = 25
+END_AGE          = 65
+START_PORTFOLIO  = 48_000   # Roth IRA ($16k) + Schwab Stocks ($32k) today
+
+# 529 -> Roth IRA pipeline (grandparent-funded). $24.9k total, $7k/yr cap.
+PIPELINE = {25: 7_000, 26: 7_000, 27: 7_000, 28: 3_900}   # sums to $24,900
+
+# Post-PhD own savings (401k + brokerage). PhD done end of 2027 => income 2028.
+POST_PHD_START_AGE = 27     # 2028 = age 27
+SCENARIOS = {
+    "Floor — no post-PhD saving":  0,
+    "Moderate — $20k/yr post-PhD": 20_000,
+    "Aggressive — $35k/yr post-PhD": 35_000,
+}
+
+# The contested decision: one summer of 10% Roth 401k = ~$2,240 at age 25.
+SUMMER_401K = 2_240
+
+# ----------------------------------------------------------------------------
+# MODEL
+# ----------------------------------------------------------------------------
+def project(post_phd_rate, summer_401k=0):
+    """Year-by-year: grow existing balance, then add year-end contributions.
+    Returns (ages, balances)."""
+    ages, balances = [], []
+    bal = START_PORTFOLIO + summer_401k  # summer contribution lands at age 25
+    for age in range(START_AGE, END_AGE + 1):
+        ages.append(age)
+        balances.append(bal)
+        # contribution for THIS year (added at year-end, after growth)
+        contrib = PIPELINE.get(age, 0)
+        if age >= POST_PHD_START_AGE:
+            contrib += post_phd_rate
+        bal = bal * (1 + REAL_RETURN) + contrib
+    return ages, balances
+
+# ----------------------------------------------------------------------------
+# PLOT
+# ----------------------------------------------------------------------------
+fig, ax = plt.subplots(figsize=(12, 7.5))
+colors = {"Floor — no post-PhD saving": "#9aa0a6",
+          "Moderate — $20k/yr post-PhD": "#1a73e8",
+          "Aggressive — $35k/yr post-PhD": "#188038"}
+
+finals = {}
+for label, rate in SCENARIOS.items():
+    ages, bals = project(rate)
+    finals[label] = bals[-1]
+    ax.plot(ages, bals, label=label, linewidth=2.6, color=colors[label])
+    # final value annotation
+    final = bals[-1]
+    income = final * SWR
+    ax.annotate(f"  ${final/1e6:.2f}M\n  (${income/1e3:.0f}k/yr @ 4%)",
+                xy=(65, final), xytext=(65.4, final),
+                va="center", fontsize=10, fontweight="bold",
+                color=colors[label])
+
+# The summer 401k question, framed HONESTLY.
+# $2,240 grows to the same gross amount in ANY account. The real question is
+# Roth-401k vs taxable-brokerage (Dane's actual alternative — he'd save it, not
+# spend it). The only delta is the tax treatment of the growth.
+LTCG_RATE = 0.15
+summer_gross = SUMMER_401K * (1 + REAL_RETURN) ** (END_AGE - START_AGE)  # grows either way
+summer_gain  = summer_gross - SUMMER_401K
+summer_taxable_net = summer_gross - summer_gain * LTCG_RATE   # taxable: pay LTCG on gain
+roth_advantage = summer_gross - summer_taxable_net            # Roth keeps the tax
+
+# vertical marker: PhD done / income starts
+ax.axvline(POST_PHD_START_AGE, color="#d93025", linestyle="--", alpha=0.55, linewidth=1.4)
+ax.annotate("PhD done /\nincome starts\n(~2028)", xy=(POST_PHD_START_AGE, ax.get_ylim()[1]*0.0),
+            xytext=(POST_PHD_START_AGE + 0.3, finals["Aggressive — $35k/yr post-PhD"]*0.62),
+            fontsize=9, color="#d93025")
+
+# annotation box: the summer 401k decision, framed honestly
+txt = (f"The contested decision (summer 10% Roth 401k, ~${SUMMER_401K:,}):\n"
+       f"• Grows to ~${summer_gross/1e3:.0f}k by 65 in ANY account\n"
+       f"• vs taxable brokerage (your real alt.), the Roth tax\n"
+       f"   benefit is only ~${roth_advantage/1e3:.1f}k — the rest happens\n"
+       f"   whether it's Roth or not.\n"
+       f"• On a ${finals['Aggressive — $35k/yr post-PhD']/1e6:.1f}M trajectory: a rounding error.")
+ax.text(0.03, 0.97, txt, transform=ax.transAxes, fontsize=9.5,
+        va="top", ha="left",
+        bbox=dict(boxstyle="round,pad=0.5", facecolor="#fef7e0", edgecolor="#f9ab00"))
+
+# formatting
+ax.set_title("Dane's Retirement Trajectory — today's dollars, 7% real return",
+             fontsize=14, fontweight="bold", pad=14)
+ax.set_xlabel("Age", fontsize=11)
+ax.set_ylabel("Portfolio value (today's $)", fontsize=11)
+ax.yaxis.set_major_formatter(FuncFormatter(lambda v, _: f"${v/1e6:.1f}M"))
+ax.set_xlim(START_AGE, 69.5)
+ax.set_ylim(bottom=0)
+ax.grid(True, alpha=0.25)
+ax.legend(loc="center left", bbox_to_anchor=(0.03, 0.55), fontsize=10, framealpha=0.9)
+
+fig.tight_layout()
+out = "/Users/danesabo/Documents/Finances/retirement_projection.png"
+fig.savefig(out, dpi=140, bbox_inches="tight")
+
+# ----------------------------------------------------------------------------
+# TEXT SUMMARY
+# ----------------------------------------------------------------------------
+print("=" * 64)
+print("RETIREMENT PROJECTION  (today's dollars, 7% real return)")
+print("=" * 64)
+print(f"Start: ${START_PORTFOLIO:,} at age {START_AGE}")
+print(f"529->Roth pipeline: ${sum(PIPELINE.values()):,} over ages "
+      f"{min(PIPELINE)}-{max(PIPELINE)}")
+print(f"Post-PhD saving starts at age {POST_PHD_START_AGE}\n")
+print(f"{'Scenario':<32}{'Age 65':>12}{'Income @4%':>14}")
+print("-" * 58)
+for label, final in finals.items():
+    print(f"{label:<32}{'$'+format(final/1e6, '.2f')+'M':>12}"
+          f"{'$'+format(final*SWR/1e3, '.0f')+'k/yr':>14}")
+print("-" * 58)
+print(f"\nSummer 10% Roth 401k (~${SUMMER_401K:,} one-time at age 25):")
+print(f"  grows to ~${summer_gross:,.0f} by age 65 in ANY account")
+print(f"  Roth-vs-taxable tax benefit only: ~${roth_advantage:,.0f}")
+print(f"  (= {roth_advantage/finals['Aggressive — $35k/yr post-PhD']*100:.3f}% "
+      f"of the aggressive total — a rounding error)")
+print(f"\nChart saved: {out}")

sam-bachelor-party-invoice.pdf

@@ -0,0 +1,86 @@
+%PDF-1.4
+%“Œ‹ž ReportLab Generated PDF document (opensource)
+1 0 obj
+<<
+/F1 2 0 R /F2 3 0 R /F3 4 0 R /F4 5 0 R
+>>
+endobj
+2 0 obj
+<<
+/BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font
+>>
+endobj
+3 0 obj
+<<
+/BaseFont /Helvetica-Bold /Encoding /WinAnsiEncoding /Name /F2 /Subtype /Type1 /Type /Font
+>>
+endobj
+4 0 obj
+<<
+/BaseFont /Helvetica-Oblique /Encoding /WinAnsiEncoding /Name /F3 /Subtype /Type1 /Type /Font
+>>
+endobj
+5 0 obj
+<<
+/BaseFont /Symbol /Name /F4 /Subtype /Type1 /Type /Font
+>>
+endobj
+6 0 obj
+<<
+/Contents 10 0 R /MediaBox [ 0 0 306 576 ] /Parent 9 0 R /Resources <<
+/Font 1 0 R /ProcSet [ /PDF /Text /ImageB /ImageC /ImageI ]
+>> /Rotate 0 /Trans <<
+
+>> 
+  /Type /Page
+>>
+endobj
+7 0 obj
+<<
+/PageMode /UseNone /Pages 9 0 R /Type /Catalog
+>>
+endobj
+8 0 obj
+<<
+/Author (Dane Sabo) /CreationDate (D:20260525192222-04'00') /Creator (\(unspecified\)) /Keywords () /ModDate (D:20260525192222-04'00') /Producer (ReportLab PDF Library - \(opensource\)) 
+  /Subject (\(unspecified\)) /Title (Sam's Bachelor Party - Settle Up) /Trapped /False
+>>
+endobj
+9 0 obj
+<<
+/Count 1 /Kids [ 6 0 R ] /Type /Pages
+>>
+endobj
+10 0 obj
+<<
+/Filter [ /ASCII85Decode /FlateDecode ] /Length 1214
+>>
+stream
+Gatm;>Ar4d'Ro4HS-PjQ<Ue;YA%PN0dY'MNGI*::<@_A[1BB/;a0eVis1[M4JIl1a,"PQ"R</8>o?@RGUgLY,m=bOA0)U!Q!>BFE!&K7YHOi('i5Yh!QGbO<R"QEe5+oYL0Z"=?nG4qD%^\%)pTH`7g,B&V$f?-qH1X)lr9b/os*XZi$BqZWDt#IVidSK>pO@/^a92IC?a#)l;?oFGOCs#Leq!<N\q9LD/9Ne;AUth>@dp.*ja.VPKNmqKm!p4q42N1Rd&L[k[eet=5j)"Y^J"l0@2o>'Oo>DJ!0;'/@5crR6&]8fl0RsO$3iqZNEAj>E7$^M[-m'h/>H1Am9)6jpYgEB7JGcf6+C-q(*0)jiOqRkTMqS]h&RfWncVDmRsErl4:$4?6T"g&Y7-1G-`VH"q4gLL1dP18iq*PRkNUF5bPZ!;o/gAHpcQ56W9Af51AL&<!SiPXKW,<M)B+BJK9H)\&76V#)$mom>U+6j4>FQNJH1]AAA5O==kX)tbHq`-mKSD,<Wb&*B'03WUY;MmZtr6dEr2@8T(pDiR^W>q1\d,2A2m>F<cQ^?.^LM!\lps+*TX'L3@fAn'&RK$e"3^Gq7F`E5t&#FBP8EpXmUq-R[^n@Qb2<<19MNe_'t8\73A^%*,E\_\:^VK_-;b,.3VWOXX#G2V6;&)6!b8L7\Ro9<Gn?d@s2;1M2a^4$c,cf1<CX#-PI@)5Ka!E>$.,!["Cdegcu]c^Kfd5EM3u9Y6hnp5Y`,R@g%c(S1C^P@NOUTb!aU,-?9Maf`'*V?CJqn&Es3E7qK]<:RG/6E]PW1GT<e#NRd;g,7aV[3,HPdI1Jb=069#iRTe;7iXs?Zs%-5?JY?[.43,Y9Feq??AI@1WaP(-57f(!U531KOn9G]k>fgO8GDp0G?F?_Gf$0r'd52HIKoFqbGjFBBSiX$7F?$/1YdtK`#p7Z/\]T)t?7Ym_YEaXoCMe[(9sU)Pf4V40m7u>@4Q/1qo3SH;/9:*^@72fgD[o+EA)hS@._.WEZ1E(o6V['-q!:Nfa47Z>C0:n;#SJR%4,AU*cqA[;S/qH\0fmWs4G[8kUQ2=XH>?fiS+!rt[*$W_E?_#cP/tW/ib]C7UBB!AfPG(12S7>h=C]Om*B2u-1(%$F>J50HQ/,dT:![f/k%`+eCt"n5Lo<nkUS7*J7-bOr5E&P>NuG92V:u1=#m>(@>g$gNRN!sU"Ug8ODu~>endstream
+endobj
+xref
+0 11
+0000000000 65535 f 
+0000000061 00000 n 
+0000000122 00000 n 
+0000000229 00000 n 
+0000000341 00000 n 
+0000000456 00000 n 
+0000000533 00000 n 
+0000000727 00000 n 
+0000000795 00000 n 
+0000001090 00000 n 
+0000001149 00000 n 
+trailer
+<<
+/ID 
+[<3d7dcb593c2c9ba6fed463683a4107b4><3d7dcb593c2c9ba6fed463683a4107b4>]
+% ReportLab generated PDF document -- digest (opensource)
+
+/Info 8 0 R
+/Root 7 0 R
+/Size 11
+>>
+startxref
+2455
+%%EOF