#!/usr/bin/env python3
"""Generate the four embeddable charts for field study No. 04 (Harris County, TX).

Every plotted value is READ FROM harris-commercial-record-2026-09.csv at run time.
Nothing is typed in by hand: if a metric is not in the CSV, this script raises
rather than draw it. The CSV is produced by count_harris.py, published beside it.

    python3 make-charts.py [path/to/harris-commercial-record-2026-09.csv]

Same drawing code as page/field-study-03/make-charts.py, so the four studies
share one visual language."""

import csv, os, sys

HERE = os.path.dirname(os.path.abspath(__file__))
CSV_PATH = sys.argv[1] if len(sys.argv) > 1 else os.path.join(
    HERE, "harris-commercial-record-2026-09.csv")

def load(path):
    vals = {}
    with open(path, newline="", encoding="utf-8") as fh:
        rows = [ln for ln in fh if not ln.startswith("#")]
    for r in csv.DictReader(rows):
        vals[r["metric"]] = (r["value"], r["share"])
    if not vals:
        raise SystemExit("no metrics parsed from " + path)
    return vals

V = load(CSV_PATH)

def num(metric):
    """A figure, straight from the CSV. KeyError if it was never counted."""
    if metric not in V:
        raise KeyError("not in the CSV, so it was not counted: " + metric)
    s = V[metric][0]
    return float(s) if "." in s else int(s)

def pct(metric):
    """The share as the CSV prints it, e.g. '50.2%'. Never recomputed here."""
    if metric not in V or not V[metric][1]:
        raise KeyError("no share published for: " + metric)
    return V[metric][1]

SURFACE="#f7f5ed"; INK="#173e32"; MUTED="#7b9384"
ACCENT="#0d6b47"; GRAY="#b6c4b5"; GRID="#d9e0d4"
FONT="system-ui,-apple-system,'Segoe UI',Helvetica,Arial,sans-serif"
CREDIT="Sterling Digital Partners · glvtl.com/field-study"

def bar(x,y,w,h,r=4):
    if w<=r: return f'<rect x="{x}" y="{y}" width="{max(w,1):.1f}" height="{h}" fill="currentColor"/>'
    return (f'<path d="M{x},{y} H{x+w-r:.1f} A{r},{r} 0 0 1 {x+w:.1f},{y+r} '
            f'V{y+h-r} A{r},{r} 0 0 1 {x+w-r:.1f},{y+h} H{x} Z" fill="currentColor"/>')

def wrap(t,n=112):
    out,cur=[],""
    for w in t.split():
        if len(cur)+len(w)+1>n: out.append(cur); cur=w
        else: cur=(cur+" "+w).strip()
    if cur: out.append(cur)
    return out

def chart(fn,title,sub,rows,axis_max,ticks,legend,note,W=720,L=248,fmt="{:,}",tickfmt=None):
    n=len(rows); band=38; top=92
    tickfmt = tickfmt or fmt
    nl=wrap(note)
    H=top+n*band+34+len(legend)*19+len(nl)*16+32
    R=66; pw=W-L-R
    s=[f'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 {W} {H}" width="{W}" height="{H}" '
       f'role="img" aria-label="{title}. {sub}" font-family="{FONT}">',
       f'<title>{title}</title><desc>{sub} {note}</desc>',
       f'<rect width="{W}" height="{H}" fill="{SURFACE}"/>',
       f'<text x="0" y="30" font-size="19" font-weight="600" fill="{INK}">{title}</text>',
       f'<text x="0" y="54" font-size="13.5" fill="{MUTED}">{sub}</text>']
    for t in ticks:
        x=L+pw*t/axis_max
        s.append(f'<line x1="{x:.1f}" y1="{top-12}" x2="{x:.1f}" y2="{top+n*band-10}" stroke="{GRID}" stroke-width="1"/>')
        s.append(f'<text x="{x:.1f}" y="{top+n*band+8}" font-size="11.5" fill="{MUTED}" text-anchor="middle">{tickfmt.format(t)}</text>')
    for i,(lab,val,hot) in enumerate(rows):
        y=top+i*band; w=pw*val/axis_max; c=ACCENT if hot else GRAY
        s.append(f'<text x="{L-14}" y="{y+14}" font-size="13" fill="{INK}" text-anchor="end">{lab}</text>')
        s.append(f'<g color="{c}">{bar(L,y,w,20)}</g>')
        s.append(f'<text x="{L+w+9:.1f}" y="{y+14}" font-size="13" font-weight="600" fill="{INK}">{fmt.format(val)}</text>')
    ly=top+n*band+32
    for j,(sw,txt) in enumerate(legend):
        yy=ly+j*19
        s.append(f'<rect x="0" y="{yy-9}" width="11" height="11" rx="2" fill="{sw}"/>')
        s.append(f'<text x="18" y="{yy}" font-size="12.5" fill="{MUTED}">{txt}</text>')
    ny=ly+len(legend)*19+10
    for k,line in enumerate(nl):
        s.append(f'<text x="0" y="{ny+k*16}" font-size="11.5" fill="{MUTED}">{line}</text>')
    s.append(f'<text x="{W}" y="{H-9}" font-size="11.5" font-weight="600" fill="{ACCENT}" '
             f'text-anchor="end">{CREDIT}</text>')
    s.append('</svg>')
    open(os.path.join(HERE,fn),'w').write('\n'.join(s))
    print("wrote",fn,f"({W}x{H})")

def stacked(fn,title,sub,rows,legend,note,W=720,L=206):
    nl=wrap(note); band=54; top=98
    H=top+len(rows)*band+18+len(legend)*19+len(nl)*16+32
    R=26; pw=W-L-R
    s=[f'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 {W} {H}" width="{W}" height="{H}" '
       f'role="img" aria-label="{title}. {sub}" font-family="{FONT}">',
       f'<title>{title}</title><desc>{sub} {note}</desc>',
       f'<rect width="{W}" height="{H}" fill="{SURFACE}"/>',
       f'<text x="0" y="30" font-size="19" font-weight="600" fill="{INK}">{title}</text>',
       f'<text x="0" y="54" font-size="13.5" fill="{MUTED}">{sub}</text>']
    for i,(lab,a,b,ntxt) in enumerate(rows):
        y=top+i*band; tot=a+b
        wa=pw*a/tot; wb=pw*b/tot-2
        s.append(f'<text x="{L-14}" y="{y+13}" font-size="13.5" fill="{INK}" text-anchor="end">{lab}</text>')
        s.append(f'<text x="{L-14}" y="{y+31}" font-size="12" fill="{MUTED}" text-anchor="end">{ntxt}</text>')
        s.append(f'<g color="{ACCENT}">{bar(L,y,wa,24,0)}</g>')
        s.append(f'<g color="{GRAY}">{bar(L+wa+2,y,wb,24)}</g>')
        s.append(f'<text x="{L+10}" y="{y+16.5}" font-size="13" font-weight="600" fill="#f7f5ed">'
                 f'{a*100.0/tot:.1f}%</text>')
        s.append(f'<text x="{L+wa+12}" y="{y+16.5}" font-size="13" font-weight="600" fill="{INK}">'
                 f'{b*100.0/tot:.1f}%</text>')
    ly=top+len(rows)*band+6
    for j,(sw,txt) in enumerate(legend):
        yy=ly+j*19
        s.append(f'<rect x="0" y="{yy-9}" width="11" height="11" rx="2" fill="{sw}"/>')
        s.append(f'<text x="18" y="{yy}" font-size="12.5" fill="{MUTED}">{txt}</text>')
    ny=ly+len(legend)*19+10
    for k,line in enumerate(nl):
        s.append(f'<text x="0" y="{ny+k*16}" font-size="11.5" fill="{MUTED}">{line}</text>')
    s.append(f'<text x="{W}" y="{H-9}" font-size="11.5" font-weight="600" fill="{ACCENT}" '
             f'text-anchor="end">{CREDIT}</text>')
    s.append('</svg>')
    open(os.path.join(HERE,fn),'w').write('\n'.join(s))
    print("wrote",fn,f"({W}x{H})")


# --- Finding 01: who taxes it -------------------------------------------------
no_city   = num("commercial_in_no_city")
in_city   = num("commercial_in_a_city")
val_no    = num("value_no_city_usd")
val_in    = num("value_in_city_usd")
total_act = num("commercial_accounts_f1_f2")
total_val = num("commercial_appraised_value_usd")

stacked("field-study-04-jurisdiction-split.svg",
  "A third of Harris County's commercial property is taxed by no city",
  "Harris County commercial and industrial accounts (state class F1 and F2) by whether any "
  "municipality taxes them.",
  [("Commercial accounts", no_city, in_city, f"n = {total_act:,}"),
   ("Commercial appraised value", val_no, val_in,
    f"n = ${total_val/1e9:,.1f}bn")],
  [(ACCENT,"No municipality on the account's taxing-jurisdiction list"),
   (GRAY,"Taxed by at least one of the county's 34 municipalities")],
  "The City of Houston itself taxes {houston:,} of these accounts ({hsh}) and ${hv:.1f}bn of the value ({hvs}). "
  "This counts property-tax jurisdiction, which is what the roll records; it is not a count of "
  "extraterritorial jurisdiction, and a property outside every city's tax jurisdiction may still sit "
  "inside a city's limited-purpose annexation. Source: Harris Central Appraisal District 2026 CAMA "
  "files, dated 13 Sep 2026, counted 20 Sep 2026.".format(
      houston=num("houston_accounts"), hsh=pct("houston_accounts"),
      hv=num("houston_value_usd")/1e9, hvs=pct("houston_value_usd")))

# --- Finding 02: leaving the city does not simplify the bill ------------------
chart("field-study-04-jurisdiction-stack.svg",
  "Leaving the city does not mean fewer taxing jurisdictions",
  "Mean number of distinct taxing jurisdictions per Harris County commercial account.",
  [("All commercial accounts", num("mean_taxing_jurisdictions"), False),
   ("Inside a city", num("mean_taxing_jurisdictions_in_city"), False),
   ("Outside every city", num("mean_taxing_jurisdictions_no_city"), True),
   ("Outside every city, in a utility district", num("mean_taxing_jurisdictions_no_city_in_utility"), True)],
  10.0, [0,2,4,6,8,10],
  [(ACCENT,"Outside every city's property-tax jurisdiction"),
   (GRAY,"Inside a city, or the whole county")],
  "The city is one layer of roughly nine, and removing it usually adds a utility district in its place. "
  "{a:,} of {t:,} commercial accounts ({as_}) are taxed by seven or more jurisdictions and {b:,} ({bs}) "
  "by ten or more; the maximum on the roll is {m:.0f}, out of 1,078 districts. Source: Harris Central "
  "Appraisal District 2026 CAMA files, dated 13 Sep 2026, counted 20 Sep 2026.".format(
      a=num("commercial_taxed_by_7_or_more"), t=total_act, as_=pct("commercial_taxed_by_7_or_more"),
      b=num("commercial_taxed_by_10_or_more"), bs=pct("commercial_taxed_by_10_or_more"),
      m=num("max_taxing_jurisdictions")),
  L=306, fmt="{:.2f}", tickfmt="{:.0f}")

# --- Finding 03: the absence of price -----------------------------------------
chart("field-study-04-price-absence.svg",
  "The Harris County conveyance record has no price in it",
  "deeds.txt, the county appraisal district's complete recorded-conveyance file. Five columns: "
  "acct, dos, clerk_yr, clerk_id, deed_id.",
  [("Recorded conveyances", num("recorded_conveyances_in_deeds_txt"), False),
   ("... carrying a date", num("conveyances_with_a_date_and_no_price"), False),
   ("... on a commercial account", num("commercial_conveyance_rows"), False),
   ("... carrying a price", num("deeds_txt_price_columns"), True)],
  2600000, [0,650000,1300000,1950000,2600000],
  [(ACCENT,"Columns that could hold a price, consideration or instrument type: 0 of 5"),
   (GRAY,"What the file does record")],
  "Texas is a non-disclosure state: there is no statute requiring a sale price to be reported, so the "
  "appraisal district does not publish one. This is an absence and is never expressed as a share of "
  "hidden sales, because how many of these conveyances were arm's-length sales, and for how much, is "
  "exactly what non-disclosure hides. Source: Harris Central Appraisal District 2026 CAMA files, dated "
  "13 Sep 2026, counted 20 Sep 2026.",
  L=248)

# --- Finding 04: address collisions, as an upper bound ------------------------
chart("field-study-04-address-collisions.svg",
  "One site address, more than one commercial account",
  "Harris County commercial accounts sharing a site address with another account. "
  "n = {:,} commercial accounts.".format(total_act),
  [("Share an address", num("commercial_sharing_an_address"), False),
   ("... excluding placeholder addresses", num("commercial_sharing_an_address_excl_placeholder"), False),
   ("... and on 2+ Key Map grid pages", num("commercial_sharing_address_across_key_maps"), True),
   ("... both filters applied", num("commercial_sharing_address_across_key_maps_excl_placeholder"), True)],
  11000, [0,2750,5500,8250,11000],
  [(ACCENT,"Upper bound on genuinely dispersed collisions"),
   (GRAY,"Every shared address, dispersed or stacked")],
  "HCAD publishes no parcel coordinate in this extract, so study No. 03's distance test cannot be run "
  "here. The Key Map grid page is the finest locator the district publishes, and it is a cell rather "
  "than a point: a subdivision straddling a grid line reads as dispersed, which the largest group on "
  "the roll does. Treat the bottom two bars as an upper bound, not a measurement. {p:,} accounts "
  "({ps}) carry no house number at all. Source: Harris Central Appraisal District 2026 CAMA files, "
  "dated 13 Sep 2026, counted 20 Sep 2026.".format(p=num("commercial_address_no_house_number"),
          ps=pct("commercial_address_no_house_number")),
  L=248)