#!/usr/bin/env python3 """X account monitor deep pass 2026-11-03: disjoint buckets, organic core top10, watchlist diff, freshness, originals-only dormancy, film STRONG_TERMS audit.""" import json, re from datetime import datetime, timezone, timedelta from collections import Counter D = "/home/kiro/trend-engine/trends/x/raw/" d = json.load(open(D + "scan_results_2026-11-03_FULL.json")) y = json.load(open(D + "scan_results_2026-11-02_FULL.json")) q = d["qualifying"] IRRELEVANT = {"AIBuzzNews","AIGuideHQ","AITECHio","AIFrontliner","TheAIColony","AITOPIAai", "GoogleLabs","GoogleAI","DonutAI","cladlabs","FaithJaiye2187","KammelaBala","MyStageAI", "AlexStudio44","AI_with_jasmin","Ai_Vaidehi","ChrisLaubAI","CINEMA505","AIScottAdams", "AISuperDomain","AI_Jasonyu","AIByLeo"} FILM = {"DiscussingFilm","nofilmschool"} LABS = {"OpenAI","AnthropicAI"} OFFTOPIC = {"interesting_aIl"} HEALTH = {"hubermanlab"} NEW = {"AIsaOneHQ","AiWithTariq"} AI_TERMS = re.compile(r"\b(ai\b|a\.i\.|seedance|kling|runway|veo|sora|minimax|hailuo|higgsfield|" r"midjourney|nano ?banana|gpt-?[0-9.]*|chatgpt|gemini|generative|render|vfx|ai film|aifilm|" r"grokbot|motion capture|deepfake|davinci|blender|autocad|storyboard|(ai|3d|generated) ?animation|" r"diffusion|comfyui|(?= 0.15%)") organic = [x for x in buckets["core"] if rate(x) >= 0.15] organic.sort(key=lambda z: z["score"], reverse=True) for i,x in enumerate(organic[:12],1): print(f"{i:2d}|@{x['handle']:16s}|{x['likes']:6d}L|{x['rt']:5d}RT|{x['replies']:4d}R|{x['views']:9d}V|{x['score']:7.0f}|lk%{rate(x):5.2f}|{x['date']}") print(f" {x['text'][:180]}") print(f" {x['url']}") print("\nORGANIC cutoff:", f"{organic[9]['score']:.0f}" if len(organic)>=10 else "n/a", "| rank11:", f"{organic[10]['score']:.0f}" if len(organic)>10 else "n/a") print("RAW cutoff:", f"{sorted(buckets['core'],key=lambda z:z['score'],reverse=True)[9]['score']:.0f}", "| rank11:", f"{sorted(buckets['core'],key=lambda z:z['score'],reverse=True)[10]['score']:.0f}") print("\n"+"="*70) print("LOW LIKE-RATE (<0.15%) inside core -- paid/pushed reach") for x in sorted([z for z in buckets["core"] if rate(z)<0.15],key=lambda z:z["score"],reverse=True): print(f" @{x['handle']:16s}|{x['likes']:5d}L|{x['views']:9d}V|lk%{rate(x):6.3f}|{x['date']}|{x['text'][:90]}") print("\n"+"="*70) print("NEW / UNCLASSIFIED handles:") for x in sorted(buckets["new"],key=lambda z:z["score"],reverse=True): print(f" @{x['handle']}|{x['likes']}L|{x['rt']}RT|{x['views']}V|lk%{rate(x):.2f}|{x['date']}") print(f" {x['text'][:300]}") print(f" {x['url']}") print("\n"+"="*70) print("FILM-INDUSTRY bucket:", len(buckets["film"]), "total,", sum(1 for x in buckets["film"] if x["is_ai"]), "AI-relevant") farm = Counter(x["handle"] for x in q if x["handle"] in FILM) print(" counts:", dict(farm)) for h in FILM: posts=[t for t in d["raw"].get(h,[]) if not t.get("retweeted_tweet")] s=sum(1 for t in posts if STRONG.search(t.get("text") or "")) a=sum(1 for t in posts if re.search(r"\bai\b", (t.get("text") or ""), re.I)) print(f" @{h}: {len(posts)} originals fetched, STRONG hits={s}, bare-ai hits={a}") print("\n"+"="*70) print("HANDLE COVERAGE:", len(set(x['handle'] for x in q)), "of", len(d['handles'])) c = Counter(x["handle"] for x in q) print(" ", ", ".join(f"@{h}({n})" for h,n in c.most_common())) print("\n"+"="*70) print("FRESHNESS (newest fetched tweet per handle, hours ago)") cutoff = datetime(2026,10,31,0,0,0) def pd(s): try: return datetime.strptime(s,"%a %b %d %H:%M:%S %z %Y").replace(tzinfo=None) except Exception: return None now = datetime.utcnow() rows=[] for h,tw in d["raw"].items(): dts=[pd(t.get("createdAt","")) for t in tw] dts=[x for x in dts if x] if not dts: continue mx=max(dts) origs=[t for t in tw if not t.get("retweeted_tweet") and not (t.get("text") or "").startswith("RT @")] inw=sum(1 for t in origs if (pd(t.get("createdAt","")) or datetime.min)>=cutoff) rows.append((h,(now-mx).total_seconds()/3600,len(tw),inw,mx)) rows.sort(key=lambda r:r[1]) print(" FRESH 8:", ", ".join(f"@{h} {a:.1f}h" for h,a,_,_,_ in rows[:8])) print(" KEY ACCOUNTS:") for h in ["NVIDIAAI","venturetwins","pika_labs","CharaspowerAI","c_valenzuelab","runwayml", "Kling_ai","AIWarper","higgsfield","Flovaai","higgsfield_ai","EHuanglu","azed_ai", "AIsaOneHQ","AiWithTariq","openart_ai","ElsaSofia__AI","ItsMaryamAi","WuxiaRocks","AI__TSUBAKI"]: for hh,a,n,i2,mx in rows: if hh==h: print(f" @{h:15s} n={n:2d} newest {a:7.1f}h ({mx}) in-window originals={i2}") print(" STALEST 12:", ", ".join(f"@{h} {a:.0f}h" for h,a,_,_,_ in rows[-12:])) stale=[r for r in rows if r[1]>48] print(f" handles >48h stale: {len(stale)}") print("\n"+"="*70) print("DORMANCY (originals only, RT-filtered) -- zero-qualifying handles") zero=[h for h in d["handles"] if h not in c] for h in zero: origs=[t for t in d["raw"].get(h,[]) if not t.get("retweeted_tweet") and not (t.get("text") or "").startswith("RT @")] dts=[pd(t.get("createdAt","")) for t in origs]; dts=[x for x in dts if x] inw=[t for t in origs if (pd(t.get("createdAt","")) or datetime.min)>=cutoff] best=max([(t.get("likeCount") or 0) for t in inw], default=0) if not dts: print(f" @{h}: NO DATA"); continue age=(datetime.utcnow()-max(dts)).total_seconds()/3600 verdict="DORMANT" if not inw else f"THRESHOLD ARTEFACT (inwin={len(inw)}, best={best}L)" print(f" @{h:18s} newest original {max(dts)} ({age:.1f}h) {verdict}") print("\n"+"="*70) print("TREND (qualifying totals)") for dt in ["2026-10-26","2026-10-27","2026-10-28","2026-10-29","2026-10-30","2026-10-31","2026-11-01","2026-11-02","2026-11-03"]: try: z=json.load(open(D+f"scan_results_{dt}_FULL.json")) print(f" {dt}: {len(z['qualifying']):4d} qual / {len(z['handles'])} scanned") except Exception as e: print(f" {dt}: {e}") print("\n"+"="*70) print("WATCHLIST DIFF vs yesterday") print(" yesterday handles:", len(y["handles"]), " today:", len(d["handles"])) print(" added:", sorted(set(d["handles"])-set(y["handles"]))) print(" removed:", sorted(set(y["handles"])-set(d["handles"])))