import json, urllib.request, urllib.parse, time, os

UA = {"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36"}
OUT = "/Users/drjones/astraea-books/.case-verify/snippets"

def search(q, court=None, after=None, before=None, n=6):
    qq = urllib.parse.quote(q)
    url = f"https://www.courtlistener.com/api/rest/v4/search/?q={qq}&type=o&format=json"
    if court: url += f"&court={court}"
    if after: url += f"&filed_after={after}"
    if before: url += f"&filed_before={before}"
    for i in range(4):
        try:
            req = urllib.request.Request(url, headers=UA)
            with urllib.request.urlopen(req, timeout=40) as r:
                return json.loads(r.read())
        except Exception as e:
            if i == 3: return {"error": str(e)}
            time.sleep(5 + i * 5)

# (name, court, after, before, queries, target_substring)
CASES = {
 "Littlefield_own": ("wash", "1997-01-01", "1997-12-31", ['"burden of proving" commingling', '"trace" "separate property" commingled', '"clear and satisfactory" evidence'], "littlefield"),
 "Littlefield_cited": ("washctapp", "2010-01-01", "2026-12-31", ['"Littlefield" commingling tracing separate'], "littlefield"),
 "Kovacs_own": ("wash", "1993-01-01", "1993-12-31", ['transmutation "joint tenancy"', '"clear and convincing" transmutation', '"source of funds" acquisition'], "kovacs"),
 "Kovacs_cited": ("wash", "2000-01-01", "2026-12-31", ['"Kovacs" transmutation'], "kovacs"),
 "Short_own": ("wash", "1995-01-01", "1995-12-31", ['"transmutation" Short 61176', '"separate property" commingled account'], "short"),
 "Short_cited": ("wash", "2000-01-01", "2026-12-31", ['"In re Marriage of Short"'], "short"),
 "Elam_own": ("wash", "1982-01-01", "1982-12-31", ['"gift" "community property" presumption acquisition', '"clear and convincing evidence" rebut'], "elam"),
 "Konzen_own": ("wash", "1985-01-01", "1985-12-31", ['"date of acquisition" separate community'], "konzen"),
 "Mueller_own": ("washctapp", "2007-01-01", "2007-12-31", ['commingling "reasonable certainty"', '"clear and satisfactory" trace separate'], "mueller"),
 "Glorfield_own": ("washctapp", "1980-01-01", "1980-12-31", ['transmutation community intent'], "glorfield"),
 "PearsonM_own": ("washctapp", "1993-01-01", "1993-12-31", ['"transmutation" separate property community'], "pearson-maines"),
 "Olivares_own": ("washctapp", "1993-01-01", "1993-12-31", ['"transmutation"', '"joint tenancy" separate'], "olivares"),
 "Sedlock_own": ("washctapp", "1993-01-01", "1993-12-31", ['"transmutation"', '"changed its character"'], "sedlock"),
 "Hadley_cited": ("wash", "1990-01-01", "2026-12-31", ['"Hadley" separate property division'], "hadley"),
 "Kraft_cited": ("washctapp", "2010-01-01", "2026-12-31", ['"Kraft" separate property "economic circumstances"'], "kraft"),
 "Berol_own": ("wash", "1950-01-01", "1950-12-31", ['"presumption" community deed Berol'], "berol"),
}

allout = {}
for name, (court, after, before, queries, target) in CASES.items():
    snips = []
    for q in queries:
        d = search(q, court, after, before)
        if "error" in d:
            snips.append({"q": q, "err": d["error"]}); continue
        for res in d.get("results", []):
            cn = (res.get("caseName") or "").lower()
            sn = res.get("opinions", [{}])[0].get("snippet", "")
            if len(sn) > 200:
                snips.append({"q": q, "case": res.get("caseName"), "date": res.get("dateFiled"),
                              "cit": res.get("citation", [])[:2], "snip": sn})
                break
        time.sleep(2)
    allout[name] = snips
    print(f"=== {name}: {len(snips)}")
    for s in snips:
        if "err" in s: print("  ERR", s["err"]); continue
        print("   Q:", s["q"], "|", s["case"], "|", s["date"])
    time.sleep(2)

json.dump(allout, open(f"{OUT}/harvest3.json", "w"), indent=1)
print("saved harvest3.json")