"""Project 101: what kind of #buildinpublic post gets answered by OTHER people?
Read-only, public AppView, 0.7 s between calls. Writes posts.json.
Window: posts created 6 h to 54 h ago (so replies had time to arrive)."""
import json, time, urllib.request, urllib.parse
from datetime import datetime, timezone, timedelta

UA = {"User-Agent": "Vonk (AI character) reading the #buildinpublic feed to learn how people talk, project 101"}
API = "https://api.bsky.app/xrpc/"
FEEDS = {"bip240": "at://did:plc:oio4hkxaop4ao4wz2pp3f4cr/app.bsky.feed.generator/build",
         "bip115": "at://did:plc:txqgp5l2dxwqivbwppd3mh2e/app.bsky.feed.generator/aaafflelemvyy"}
ME = "did:plc:e46yjn4y2jktr6riufqzx5fc"
NOW = datetime.now(timezone.utc)
NEW, OLD = NOW - timedelta(hours=6), NOW - timedelta(hours=54)

def get(path, **q):
    url = API + path + "?" + urllib.parse.urlencode(q, doseq=True)
    err = None
    for _ in range(3):
        try:
            time.sleep(0.7)
            return json.load(urllib.request.urlopen(urllib.request.Request(url, headers=UA), timeout=25))
        except Exception as e:
            err = e; time.sleep(2)
    return {"error": str(err)}

def ts(s): return datetime.fromisoformat(s.replace("Z", "+00:00"))

posts = {}
def take(p, src):
    rec = p.get("record", {})
    if rec.get("reply"):  # only top-level posts
        return
    t = ts(rec.get("createdAt", p.get("indexedAt")))
    if not (OLD <= t <= NEW) or p["author"]["did"] == ME:
        return
    e = p.get("embed") or {}
    et = e.get("$type", "")
    facets = rec.get("facets", [])
    links = any(f["features"][0]["$type"].endswith("#link") for f in facets)
    d = posts.setdefault(p["uri"], {
        "uri": p["uri"], "url": f"https://bsky.app/profile/{p['author']['handle']}/post/{p['uri'].split('/')[-1]}",
        "did": p["author"]["did"], "handle": p["author"]["handle"], "time": rec["createdAt"], "text": rec.get("text", ""),
        "image": "images" in et or "video" in et or "images" in json.dumps(e.get("media", {}))[:200],
        "link": links or "external" in et, "quote": "record" in et,
        "bot": any(l.get("val") == "bot" for l in p["author"].get("labels", [])),
        "likes": p.get("likeCount", 0), "replies": p.get("replyCount", 0), "reposts": p.get("repostCount", 0),
        "src": []})
    d["src"].append(src)

for name, uri in FEEDS.items():
    cursor = None
    for _ in range(20):
        q = {"feed": uri, "limit": 100}
        if cursor: q["cursor"] = cursor
        r = get("app.bsky.feed.getFeed", **q)
        if "error" in r: print(name, r["error"]); break
        items = r.get("feed", [])
        for it in items: take(it["post"], name)
        cursor = r.get("cursor")
        if not items or not cursor or ts(items[-1]["post"]["record"].get("createdAt", items[-1]["post"]["indexedAt"])) < OLD:
            break
cursor = None
for _ in range(15):
    q = {"q": "#buildinpublic", "tag": "buildinpublic", "sort": "latest", "limit": 100}
    if cursor: q["cursor"] = cursor
    r = get("app.bsky.feed.searchPosts", **q)
    if "error" in r: print("search", r["error"]); break
    items = r.get("posts", [])
    for p in items: take(p, "tag")
    cursor = r.get("cursor")
    if not items or not cursor or ts(items[-1]["record"]["createdAt"]) < OLD: break
print("posts in window:", len(posts))

# follower counts
dids = sorted({p["did"] for p in posts.values()})
prof = {}
for i in range(0, len(dids), 25):
    r = get("app.bsky.actor.getProfiles", actors=dids[i:i+25])
    for a in r.get("profiles", []): prof[a["did"]] = a.get("followersCount", 0)
for p in posts.values(): p["followers"] = prof.get(p["did"], 0)

# who replied: others vs the author themself (threads count as replies too)
n = 0
for p in sorted(posts.values(), key=lambda x: -x["replies"]):
    if p["replies"] == 0 or n >= 260: p["others"] = 0 if p["replies"] == 0 else None; continue
    r = get("app.bsky.feed.getPostThread", uri=p["uri"], depth=1, parentHeight=0); n += 1
    reps = (r.get("thread") or {}).get("replies", []) or []
    o = [x["post"] for x in reps if "post" in x and x["post"]["author"]["did"] != p["did"]]
    p["others"] = len(o)
    p["other_replies"] = [{"handle": x["author"]["handle"], "bot": any(l.get("val") == "bot" for l in x["author"].get("labels", [])),
                           "text": x["record"].get("text", "")[:200]} for x in o[:6]]
json.dump({"fetched": NOW.isoformat(), "window": [OLD.isoformat(), NEW.isoformat()], "posts": list(posts.values())},
          open("posts.json", "w", encoding="utf-8"), ensure_ascii=False, indent=1)
print("threads read:", n, "authors:", len(dids))