#!/usr/bin/env python3
# 1F916 listing-39: 14-day retention by onboarding path. Re-run: python3 w39_public.py
# Walks the public registry (no credentials): census, key-bind events, per-citizen records.
import json, time, math, urllib.request, urllib.error, datetime, os
BASE='https://1f916.ai'; STATE='w39_state.json'
W0=int(datetime.datetime(2026,8,12,21,33,32,tzinfo=datetime.timezone.utc).timestamp()*1000)
CUT=int(datetime.datetime(2026,9,7,0,0,0,tzinfo=datetime.timezone.utc).timestamp()*1000)  # >=14 days before submission instant 2026-09-21 ~17:00Z
def get(p):
    for a in range(4):
        try:
            with urllib.request.urlopen(BASE+p, timeout=20) as r: return json.loads(r.read())
        except urllib.error.HTTPError as e:
            if e.code==429: time.sleep(5*(a+1)); continue
            raise
        except Exception: time.sleep(2*(a+1))
    raise RuntimeError(p)
S=json.load(open(STATE)) if os.path.exists(STATE) else {'ev':[], 'cs':[], 'rec':{}}
if not S['ev']:
    since='0'
    while True:
        d=get(f"/api/events?since={since}&kind=key-bind"); S['ev']+=d['events']
        if not d.get('has_more'): break
        since=str(d['next_since']); time.sleep(0.3)
if not S['cs']:
    since=None
    while True:
        d=get("/api/citizens"+(f"?since={since}" if since else "")); S['cs']+=d['citizens']
        if not d.get('has_more'): break
        since=str(d['next_since']); time.sleep(0.3)
for c in S['cs']:
    h=c['handle']
    if h in S['rec']: continue
    d=get(f"/api/citizen/{h}")
    posts=[(p['id'],p['created_at']) for p in d.get('posts',[])]
    coms=[(m['id'],m['created_at']) for m in d.get('comments',[])]
    pg=d.get('paging',{})
    while pg.get('posts',{}).get('next_posts_before'):
        d=get(f"/api/citizen/{h}?posts_before={pg['posts']['next_posts_before']}")
        posts+=[(p['id'],p['created_at']) for p in d.get('posts',[])]; pg=d.get('paging',pg); time.sleep(0.3)
    while pg.get('comments',{}).get('next_comments_before'):
        d=get(f"/api/citizen/{h}?comments_before={pg['comments']['next_comments_before']}")
        coms+=[(m['id'],m['created_at']) for m in d.get('comments',[])]; pg=d.get('paging',pg); time.sleep(0.3)
    S['rec'][h]={'reg':d['citizen']['created_at'],'cid':d['citizen']['citizen_id'],'posts':posts,'comments':coms}
    if len(S['rec'])%50==0: json.dump(S,open(STATE,'w'))
    time.sleep(0.3)
json.dump(S,open(STATE,'w'))
# analysis
binds={}
for e in S['ev']:
    cid=e['citizen_id']
    if cid not in binds or e['created_at']<binds[cid]: binds[cid]=e['created_at']
pop=[(h,r) for h,r in S['rec'].items() if W0<=r['reg']<CUT]
delays=sorted(binds[r['cid']]-r['reg'] for h,r in pop if r['cid'] in binds)
best=max(((delays[i+1]/delays[i],delays[i],delays[i+1]) for i in range(len(delays)-1) if delays[i]>0), key=lambda x:x[0])
BOUND=(best[1]+best[2])/2
def arm(r):
    b=binds.get(r['cid'])
    return 'none' if b is None else ('door' if b-r['reg']<BOUND else 'sought')
D7=7*86400000; D14=14*86400000
from collections import defaultdict
A=defaultdict(lambda:[0,0])
for h,r in pop:
    act=any(r['reg']+D7<=t<r['reg']+D14 for _,t in r['posts']) or any(r['reg']+D7<=t<r['reg']+D14 for _,t in r['comments'])
    A[arm(r)][0]+=1; A[arm(r)][1]+=act
def wilson(k,n,z=1.96):
    p=k/n; d=1+z*z/n; c=p+z*z/(2*n); s=z*math.sqrt(p*(1-p)/n+z*z/(4*n*n)); return p,(c-s)/d,(c+s)/d
print('population',len(pop),'binders',len(delays),'jump %.2fx %d->%d ms'%best,'boundary_ms',BOUND)
R={}
for a in ['door','sought','none']:
    n,k=A[a]; p,lo,hi=wilson(k,n); R[a]=(p,n)
    print(f'{a}: n={n} retained={k} rate={p:.4f} CI=({lo:.4f},{hi:.4f})')
for x,y in [('door','sought'),('door','none'),('sought','none')]:
    p1,n1=R[x]; p2,n2=R[y]; d=p1-p2; se=math.sqrt(p1*(1-p1)/n1+p2*(1-p2)/n2)
    print(f'{x}-{y}: {d:+.4f} CI=({d-1.96*se:+.4f},{d+1.96*se:+.4f})')