@danhachuel/thunderbolt 0.6.11 → 0.6.13

package/app/main.py

@@ -110,6 +110,8 @@ from hermes_ui.creative_generation import CreativeGenerationError, generate_crea
 from hermes_ui.media_generation import MediaGenerationError, format_media_generation_error, generate_image_for_card, generate_video_for_card
 from hermes_ui.growth_youtube import list_analyses, run_audit
 from hermes_ui.growth_tiktok import render_growth_tiktok
+from hermes_ui.growth_instagram import render_growth_instagram
+from hermes_ui.growth_facebook_pages import render_growth_facebook_pages
 from hermes_ui.canva_auth import authorization_url, create_pkce_pair, create_state, exchange_code
 from integrations.platforms import IntegrationResult, TikTokAdapter, YouTubeAdapter, fetch_channel_videos_public
 from integrations.tiktok_public import fetch_public_tiktok_profile, normalize_tiktok_reference
@@ -2198,6 +2200,8 @@ def classify_channel_platform(channel: Any) -> str:
         return "youtube"
     if value in {"tiktok", "tik-tok", "tt"}:
         return "tiktok"
+    if value in {"instagram", "ig"}:
+        return "instagram"
     def has_tiktok_marker(item: Any) -> bool:
         if isinstance(item, dict):
             return any(has_tiktok_marker(key) or has_tiktok_marker(value) for key, value in item.items())
@@ -9070,8 +9074,8 @@ def main():
         "Redes Sociais": lambda: render_edit_placeholder("Redes Sociais", "Área reservada para a futura funcionalidade de redes sociais."),
         "Analista Growth Youtube": render_growth_youtube,
         "Analista Growth Tiktok": render_growth_tiktok,
-        "Analista Growth Instagram": lambda: render_edit_placeholder("Analista Growth Instagram", ""),
-        "Analista Facebook Pages": lambda: render_edit_placeholder("Analista Facebook Pages", ""),
+        "Analista Growth Instagram": render_growth_instagram,
+        "Analista Facebook Pages": render_growth_facebook_pages,
         "Analista Bilibili": lambda: render_edit_placeholder("Analista Bilibili", ""),
         "Documentação": lambda: render_edit_placeholder("Documentação", "Seleccione um tutorial no menu expansível."),
         "Tutorial Meta": render_models_ai_tutorial,

package/hermes_ui/growth_facebook_pages.py

@@ -0,0 +1,200 @@
+from __future__ import annotations
+
+import re
+import uuid
+from datetime import datetime, timezone
+from html import escape
+from pathlib import Path
+from typing import Any, Callable, Mapping
+
+import streamlit as st
+
+from .storage import STORAGE, read_json, write_json
+
+ANALYSES_FILENAME = "facebook_pages_growth_analyses.json"
+ANALYSES_DIR = STORAGE / "growth" / "facebook_pages"
+AGENT_METRICS = (
+    "Demanda validada", "Qualidade da thumbnail", "Qualidade do título",
+    "Hook — retenção inicial", "Ritmo e edição", "Origem Recommended", "Conversão em seguidores", "Cadência",
+)
+
+
+def analysis_code(now: datetime | None = None) -> str:
+    moment = now or datetime.now(timezone.utc)
+    return f"GFA-{moment.strftime('%Y%m%d-%H%M%S')}-{uuid.uuid4().hex[:6].upper()}"
+
+
+def _score_color(score: int | float) -> str:
+    return "red" if score <= 30 else "yellow" if score < 70 else "green"
+
+
+def _safe_int(value: Any, default: int = 0) -> int:
+    try:
+        return int(value or default)
+    except (TypeError, ValueError):
+        return default
+
+
+def _public_posts(page: Mapping[str, Any], limit: int = 5) -> list[dict[str, Any]]:
+    posts = page.get("posts") or page.get("recent_posts") or page.get("public_posts") or []
+    if isinstance(posts, list):
+        return [dict(item) for item in posts[:limit] if isinstance(item, Mapping)]
+    return []
+
+
+def _pages() -> list[dict[str, Any]]:
+    settings = read_json("settings.json", {})
+    candidates: list[Any] = []
+    for key in ("facebook_pages", "facebook_page_accounts", "meta_pages"):
+        value = settings.get(key, []) if isinstance(settings, Mapping) else []
+        candidates.extend(value if isinstance(value, list) else [value] if isinstance(value, Mapping) else [])
+    for channel in read_json("channels.json", []):
+        if isinstance(channel, dict) and str(channel.get("platform") or "").casefold() in {"facebook", "facebook_pages", "facebook page"}:
+            candidates.append(channel)
+    unique: dict[str, dict[str, Any]] = {}
+    for item in candidates:
+        if not isinstance(item, Mapping):
+            continue
+        page = dict(item)
+        key = str(page.get("id") or page.get("page_id") or page.get("name") or page.get("url") or "").strip()
+        if key:
+            unique[key] = page
+    return [item for item in unique.values() if item.get("active", True)]
+
+
+def _report_markdown(record: Mapping[str, Any]) -> str:
+    lines = [f"# FACEBOOK GROWTH AUDIT REPORT: {record.get('page_name', 'Página')}", "", f"- Código: `{record.get('code')}`", "- Plataforma: Facebook Pages", f"- Data: {record.get('created_at')}", f"- Nota geral: **{record.get('overall_score', 0)}/100**", "", "## Projecção financeira e contexto", "", "A auditoria pública não expõe CPM, minutos de visualização, CTR ou dados de monetização. O relatório não inventa estes valores. A elegibilidade deve confirmar 10.000 seguidores, 600.000 minutos de watch time em 60 dias, conteúdo original e conta em país elegível.", "", "## Diagnóstico", "", "| Pilar | Nota | Métrica observada | Origem | Diagnóstico |", "|---|---:|---|---|---|"]
+    for metric in record.get("metrics", []):
+        lines.append(f"| **{metric['label']}** | {metric['score']}/100 | {metric['value']} | {metric.get('source', 'unknown')} | {metric['diagnosis']} |")
+    lines += ["", "## Últimas publicações públicas", "", "| Título | Visualizações | Reacções | Comentários | Partilhas |", "|---|---:|---:|---:|---:|"]
+    for post in record.get("posts", []):
+        lines.append(f"| {str(post.get('title') or post.get('message') or 'Sem título').replace('|', '/')} | {_safe_int(post.get('view_count') or post.get('views')):,} | {_safe_int(post.get('like_count') or post.get('reactions')):,} | {_safe_int(post.get('comment_count') or post.get('comments')):,} | {_safe_int(post.get('share_count') or post.get('shares')):,} |")
+    lines += ["", "## Roadmap accionável", "", "1. **REMAKE THUMBNAILS**: aumentar contraste, usar sujeito claro, emoção forte e no máximo três palavras legíveis no telemóvel.", "2. **REWRITE TITLES**: começar pela keyword principal e combinar curiosidade, clareza e urgência sem prometer algo que o vídeo não entrega.", "3. **EDIT HOOK**: abrir com o resultado ou estatística forte; adicionar pattern interrupt a cada 3–5 segundos e CTA visual a 40% do vídeo.", "", "## Limitações", "", "CTR, retenção de 3/15 segundos, tráfego Recommended, alcance, minutos de watch time, seguidores ganhos e monetização exigem dados autorizados do Facebook Page Insights/Creator Studio."]
+    return "\n".join(lines) + "\n"
+
+
+def run_audit(page: Mapping[str, Any], settings: Mapping[str, Any] | None = None, *, progress: Callable[[str], None] | None = None) -> dict[str, Any]:
+    del settings
+    code = analysis_code()
+    root = ANALYSES_DIR / code
+    root.mkdir(parents=True, exist_ok=True)
+    notify = progress or (lambda _message: None)
+    notify("A preparar a auditoria pública da Facebook Page…")
+    posts = _public_posts(page, 5)
+    views = sum(_safe_int(item.get("view_count") or item.get("views")) for item in posts)
+    reactions = sum(_safe_int(item.get("like_count") or item.get("reactions")) for item in posts)
+    comments = sum(_safe_int(item.get("comment_count") or item.get("comments")) for item in posts)
+    shares = sum(_safe_int(item.get("share_count") or item.get("shares")) for item in posts)
+    engagement = ((reactions + comments + shares) / views * 100) if views else None
+    followers = _safe_int(page.get("followers") or page.get("follower_count") or page.get("fan_count")) or None
+    conversion = (followers / views * 100) if followers and views else None
+    cadence = 70 if len(posts) >= 3 else 40 if posts else 0
+    public_value = f"{len(posts)} publicações públicas" if posts else "Sem publicações públicas estruturadas"
+    metrics = [
+        {"label": "Demanda validada", "score": 50 if posts else 0, "value": public_value, "source": "public", "diagnosis": "Concorrência, pesquisa e Recommended exigem Page Insights."},
+        {"label": "Qualidade da thumbnail", "score": 50, "value": "CTR indisponível", "source": "unavailable", "diagnosis": "CTR privado; meta do agente: acima de 6%."},
+        {"label": "Qualidade do título", "score": 50, "value": "Retenção cruzada indisponível", "source": "unavailable", "diagnosis": "Avaliar os 3 C's: Curiosity, Clarity e Urgency."},
+        {"label": "Hook — retenção inicial", "score": 50, "value": "Retenção de 3/15s indisponível", "source": "unavailable", "diagnosis": "Creator Studio é necessário para a retenção absoluta."},
+        {"label": "Ritmo e edição", "score": 50, "value": "Quedas de retenção indisponíveis", "source": "unavailable", "diagnosis": "Aplicar pattern interrupt a cada 3–5 segundos."},
+        {"label": "Origem Recommended", "score": 50, "value": "Fontes de tráfego indisponíveis", "source": "unavailable", "diagnosis": "Meta: 40–70% Recommended/Suggestions."},
+        {"label": "Conversão em seguidores", "score": 50 if conversion is None else (85 if conversion > 1.5 else 60 if conversion >= .5 else 30), "value": f"{conversion:.2f}% proxy" if conversion is not None else "Seguidores ganhos por publicação indisponíveis", "source": "estimated" if conversion is not None else "unavailable", "diagnosis": "Meta: acima de 1,5%; proxy não substitui seguidores ganhos."},
+        {"label": "Cadência", "score": cadence, "value": f"{len(posts)} publicações na amostra", "source": "public", "diagnosis": "Meta do agente: 3–5 publicações/Reels por semana."},
+    ]
+    overall = round(sum(item["score"] for item in metrics) / len(metrics))
+    record = {"code": code, "page_id": str(page.get("page_id") or page.get("id") or ""), "page_name": str(page.get("name") or page.get("page_name") or "Facebook Page"), "platform": "Facebook Pages", "created_at": datetime.now(timezone.utc).isoformat(), "overall_score": overall, "score_color": _score_color(overall), "metrics": metrics, "posts": posts, "sample_views": views, "engagement_rate": engagement, "monetization_status": "A verificar — confirmar requisitos no Creator Studio", "status": "completed"}
+    report_path = root / f"{code}.md"
+    report_path.write_text(_report_markdown(record), encoding="utf-8")
+    record["report_path"] = str(report_path)
+    analyses = read_json(ANALYSES_FILENAME, [])
+    if not isinstance(analyses, list): analyses = []
+    analyses.append(record)
+    write_json(ANALYSES_FILENAME, analyses[-100:])
+    return record
+
+
+def _card(title: str, score: Any, featured: bool, rows: list[tuple[str, str, str, str, str]]) -> str:
+    numeric = _safe_int(score)
+    colour = "#ef4444" if numeric <= 30 else "#f59e0b" if numeric < 70 else "#22c55e"
+    row_html = "".join(f'<tr><td>{escape(label)}</td><td><b>{escape(value)}</b></td><td>{escape(target)}</td><td style="color:{status_colour};font-weight:700;">● {escape(status)}</td></tr>' for label, value, target, status, status_colour in rows)
+    return f'<div style="min-height:{"350px" if featured else "280px"};border:1px solid #2b3b52;border-radius:14px;padding:16px;background:linear-gradient(145deg,#111c2b,#0b121c);box-shadow:0 8px 22px rgba(0,0,0,.16);"><div style="font-size:{"19px" if featured else "16px"};font-weight:750;color:#f8fafc;margin-top:8px;">{escape(title)}</div><div style="display:flex;align-items:baseline;gap:5px;margin:8px 0 10px;"><span style="font-size:{"38px" if featured else "30px"};font-weight:850;color:{colour};">{numeric}</span><span style="font-size:13px;color:#94a3b8;">/100</span></div><div style="overflow-x:auto;"><table style="width:100%;border-collapse:collapse;font-size:11px;color:#cbd5e1;"><thead><tr><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">KPI</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Valor</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Meta</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Status</th></tr></thead><tbody>{row_html}</tbody></table></div></div>'
+
+
+def render_growth_facebook_pages() -> None:
+    st.title("Analista Facebook Pages")
+    st.caption("Auditoria de Facebook Pages baseada nos 3 pilares críticos e nas métricas secundárias do agente Growth.")
+    pages = _pages()
+    if not pages:
+        st.info("Configure pelo menos uma Facebook Page em Configuração API ou no registo de páginas para iniciar a análise.")
+        return
+    selector_col, action_col = st.columns([1.05, 1.15], gap="small")
+    with selector_col:
+        selected = st.selectbox("Página a analisar", pages, format_func=lambda item: str(item.get("name") or item.get("page_name") or "Página sem nome"), key="growth_facebook_page")
+    with action_col:
+        st.write("")
+        analyse_clicked = st.button("ANALISAR PÁGINA", type="primary", use_container_width=True, key="growth_facebook_analyse")
+    st.markdown("**Estado das APIs de Growth**")
+    api_cols = st.columns(2, gap="small")
+    with api_cols[0]:
+        st.caption("Dados públicos e registos locais")
+        st.markdown('<span style="color:#22c55e;font-weight:700;">● Disponível quando houver publicações guardadas</span>', unsafe_allow_html=True)
+    with api_cols[1]:
+        st.caption("Facebook Page Insights / Creator Studio")
+        st.markdown('<span style="color:#94a3b8;font-weight:700;">● Requer dados privados autorizados</span>', unsafe_allow_html=True)
+    if analyse_clicked:
+        with st.status("A preparar a análise…", expanded=True) as status:
+            try:
+                record = run_audit(selected, read_json("settings.json", {}), progress=st.write)
+                status.update(label="Análise concluída", state="complete")
+            except Exception as exc:
+                status.update(label="Análise falhou", state="error")
+                st.error(str(exc)[:500])
+    records = list_analyses()
+    selected_record = next((item for item in reversed(records) if str(item.get("page_id")) == str(selected.get("page_id") or selected.get("id"))), None)
+    metric_by_label = {str(item.get("label")): item for item in (selected_record.get("metrics", []) if selected_record else [])}
+    metrics = selected_record.get("metrics", []) if selected_record else [{"label": label, "score": 0, "value": "Aguardando análise", "source": "unknown"} for label in AGENT_METRICS]
+    score = selected_record.get("overall_score") if selected_record else None
+    colour = "#64748b" if score is None else "#ef4444" if score <= 30 else "#f59e0b" if score < 70 else "#22c55e"
+    summary_col, download_col = st.columns([3.25, 1], gap="large")
+    with summary_col:
+        st.markdown(f'<div style="border:1px solid #263447;border-radius:12px;padding:18px;margin:12px 0 20px;background:#101722;display:flex;align-items:center;justify-content:space-between;min-height:76px;"><div><div style="font-size:13px;color:#94a3b8;text-transform:uppercase;letter-spacing:.08em;">Nota geral da página</div><div style="font-size:18px;font-weight:700;margin-top:5px;">{escape(str(selected.get("name") or selected.get("page_name") or "Facebook Page"))}</div></div><div style="font-size:42px;font-weight:800;color:{colour};line-height:1;">{score if score is not None else "—"}<span style="font-size:16px;color:#94a3b8;">{"/100" if score is not None else ""}</span></div></div>', unsafe_allow_html=True)
+    with download_col:
+        st.write("")
+        path = Path(str(selected_record.get("report_path") or "")) if selected_record else Path()
+        if path.is_file():
+            st.download_button("BAIXAR ANALISE COMPLETA", path.read_bytes(), file_name=path.name, mime="text/markdown", use_container_width=True, key=f"facebook_download_{selected_record['code']}")
+        else:
+            st.button("BAIXAR ANALISE COMPLETA", disabled=True, use_container_width=True, key="facebook_download_disabled")
+    st.subheader("Dashboard de Growth")
+    st.caption("3 pilares críticos em destaque e camadas operacionais abaixo. Vermelho: 0–30 · Amarelo: 31–69 · Verde: 70–100.")
+    def item(label: str, target: str, available: bool = False) -> tuple[str, str, str, str, str]:
+        metric = metric_by_label.get(label, {})
+        return (label, str(metric.get("value", "Aguardando análise")), target, "Meta atingida" if available else "A verificar", "#22c55e" if available else "#94a3b8")
+    cards = [("Demanda validada", True, [item("Demanda validada", "> 500k concorrência", True)]), ("Qualidade da thumbnail", True, [item("Qualidade da thumbnail", "> 6% CTR")]), ("Qualidade do título", True, [item("Qualidade do título", "3 C's")]), ("Hook — retenção inicial", False, [item("Hook — retenção inicial", "> 50% aos 3s")]), ("Ritmo e edição", False, [item("Ritmo e edição", "Interrupt 3–5s")]), ("Origem Recommended", False, [item("Origem Recommended", "40–70%")]), ("Conversão em seguidores", False, [item("Conversão em seguidores", "> 1.5%")]), ("Cadência", False, [item("Cadência", "3–5/semana")])]
+    st.markdown('<div style="display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:14px;margin:14px 0 18px;">' + "".join(_card(title, metric_by_label.get(title, {}).get("score", 0), featured, rows) for title, featured, rows in cards[:3]) + '</div>', unsafe_allow_html=True)
+    st.markdown('<div style="display:grid;grid-template-columns:repeat(5,minmax(0,1fr));gap:12px;margin:0 0 18px;">' + "".join(_card(title, metric_by_label.get(title, {}).get("score", 0), featured, rows) for title, featured, rows in cards[3:]) + '</div>', unsafe_allow_html=True)
+    st.caption("Legenda: ● Verde = meta atingida · ● Amarelo = atenção · ● Vermelho = crítico · Cinzento = dado não disponível na fonte autorizada.")
+    st.divider()
+    history_tab, dashboard_tab = st.tabs(["Últimas análises", "Leitura estratégica"])
+    with dashboard_tab:
+        st.write("Priorize Thumbnail, Título e Hook. CTR, retenção, tráfego Recommended, minutos de watch time e monetização ficam pendentes do Page Insights/Creator Studio.")
+    with history_tab:
+        if not records:
+            st.info("Ainda não existem análises guardadas.")
+        for record in reversed(records[-20:]):
+            row = st.columns([2, 1, 2.2, 1.2, 1, 1.4])
+            row[0].write(f"**{record.get('page_name', 'Facebook Page')}**")
+            row[1].write("Facebook Pages")
+            row[2].caption(str(record.get("code", "")))
+            row[3].caption(str(record.get("created_at", "")).replace("T", " ")[:19])
+            row[4].write(f"{record.get('overall_score', '—')}/100")
+            path = Path(str(record.get("report_path") or ""))
+            if path.is_file():
+                row[5].download_button("Análise", path.read_bytes(), file_name=path.name, mime="text/markdown", key=f"facebook_history_{record.get('code')}")
+
+
+def list_analyses() -> list[dict[str, Any]]:
+    value = read_json(ANALYSES_FILENAME, [])
+    return [dict(item) for item in value if isinstance(item, Mapping)] if isinstance(value, list) else []
+
+
+__all__ = ["AGENT_METRICS", "ANALYSES_DIR", "analysis_code", "list_analyses", "run_audit", "render_growth_facebook_pages"]

package/hermes_ui/growth_instagram.py

@@ -0,0 +1,264 @@
+from __future__ import annotations
+
+import json
+import re
+import uuid
+from datetime import datetime, timezone
+from html import escape
+from pathlib import Path
+from typing import Any, Callable, Mapping
+
+import streamlit as st
+
+from integrations.instagram_public import fetch_public_instagram_profile
+from .storage import STORAGE, read_json, write_json
+
+ANALYSES_FILENAME = "instagram_growth_analyses.json"
+ANALYSES_DIR = STORAGE / "growth" / "instagram"
+AGENT_METRICS = (
+    "Demanda e potencial de viralização", "Qualidade visual — stop the scroll", "Hook e storytelling",
+    "Ritmo e edição", "Tráfego Explore/Reels", "Engajamento e saves", "Conversão em seguidores", "Mix de formatos", "Cadência",
+)
+
+
+def analysis_code(now: datetime | None = None) -> str:
+    moment = now or datetime.now(timezone.utc)
+    return f"GIA-{moment.strftime('%Y%m%d-%H%M%S')}-{uuid.uuid4().hex[:6].upper()}"
+
+
+def _score_color(score: int | float) -> str:
+    return "red" if score <= 30 else "yellow" if score < 70 else "green"
+
+
+def _safe_int(value: Any, default: int = 0) -> int:
+    try:
+        return int(value or default)
+    except (TypeError, ValueError):
+        return default
+
+
+def _number(value: Any) -> float | None:
+    try:
+        return float(value) if value not in (None, "", "—") else None
+    except (TypeError, ValueError):
+        return None
+
+
+def _public_videos(source: str, limit: int = 5) -> list[dict[str, Any]]:
+    try:
+        import yt_dlp
+    except ImportError as exc:
+        raise RuntimeError("yt-dlp não está instalado. Instale as dependências do Thunderbolt.") from exc
+    options = {"quiet": True, "skip_download": True, "extract_flat": True, "playlistend": limit, "noplaylist": False}
+    with yt_dlp.YoutubeDL(options) as downloader:
+        info = downloader.extract_info(source, download=False)
+    entries = info.get("entries") if isinstance(info, Mapping) else []
+    videos: list[dict[str, Any]] = []
+    for item in entries or []:
+        if not isinstance(item, Mapping):
+            continue
+        video_id = str(item.get("id") or item.get("webpage_url") or "").strip()
+        if not video_id:
+            continue
+        videos.append({
+            "id": video_id,
+            "title": str(item.get("title") or "Sem título").strip(),
+            "url": str(item.get("webpage_url") or f"https://www.instagram.com/@{source.lstrip('@').split('/')[-1]}/video/{video_id}"),
+            "view_count": _safe_int(item.get("view_count")),
+            "like_count": _safe_int(item.get("like_count")),
+            "comment_count": _safe_int(item.get("comment_count")),
+            "repost_count": _safe_int(item.get("repost_count") or item.get("share_count")),
+            "save_count": _safe_int(item.get("save_count")),
+            "duration": _safe_int(item.get("duration")),
+            "upload_date": str(item.get("upload_date") or ""),
+        })
+    return videos[:limit]
+
+
+def _score_from_rate(rate: float | None, thresholds: tuple[float, float]) -> int:
+    if rate is None:
+        return 50
+    low, high = thresholds
+    if rate < low:
+        return 30
+    if rate < high:
+        return 60
+    return 85
+
+
+def _report_markdown(record: Mapping[str, Any]) -> str:
+    profile = record.get("profile") or {}
+    lines = [
+        f"# INSTAGRAM CHANNEL AUDIT REPORT: {record.get('channel_name', 'Canal')}", "",
+        f"- Código: `{record.get('code')}`", "- Plataforma: Instagram",
+        f"- Data: {record.get('created_at')}", f"- Nota geral: **{record.get('overall_score', 0)}/100**", "",
+        "## Resumo executivo", "",
+        f"- Elegível para ferramentas de monetização: **{record.get('monetization_status', 'A verificar')}**",
+        f"- Seguidores públicos: **{profile.get('subscriber_count', '—')}**",
+        f"- Visualizações públicas na amostra: **{record.get('sample_views', '—')}**", "",
+        "## Diagnóstico", "", "| Pilar | Nota | Métrica observada | Origem | Diagnóstico |", "|---|---:|---|---|---|",
+    ]
+    for metric in record.get("metrics", []):
+        lines.append(f"| **{metric['label']}** | {metric['score']}/100 | {metric['value']} | {metric.get('source', 'unknown')} | {metric['diagnosis']} |")
+    lines += ["", "## Últimos vídeos públicos", "", "| Título | Visualizações | Gostos | Comentários | Duração |", "|---|---:|---:|---:|---:|"]
+    for video in record.get("videos", []):
+        lines.append(f"| {str(video.get('title', '')).replace('|', '/')} | {video.get('view_count', 0):,} | {video.get('like_count', 0):,} | {video.get('comment_count', 0):,} | {video.get('duration', 0)}s |")
+    lines += ["", "## Roadmap accionável", "", "1. **REBUILD THE HOOK**: nos primeiros 3 segundos, começar com resultado, estatística, pergunta directa ou pattern interrupt; sem logo ou introdução lenta.", "2. **OPTIMIZE RETENTION**: renovar visual/texto a cada 3–5 segundos, inserir pattern interrupt a cada 5–8 segundos e evitar frames estáticos por mais de 3 segundos.", "3. **PUBLISH CONSISTENTLY**: manter 3–5 Reels por semana, combinando Reels para alcance, Carrosséis para saves e Stories para relacionamento.", "", "## Limitações", "", "Retenção de 3 segundos, skip rate, conclusão, alcance, impressões, fontes Explore/Reels, saves, RPM e seguidores ganhos exigem dados autorizados do Instagram Insights. Estes valores não são inventados a partir de uma página pública."]
+    return "\n".join(lines) + "\n"
+
+
+def run_audit(channel: Mapping[str, Any], settings: Mapping[str, Any] | None = None, *, progress: Callable[[str], None] | None = None) -> dict[str, Any]:
+    del settings
+    code = analysis_code()
+    root = ANALYSES_DIR / code
+    root.mkdir(parents=True, exist_ok=True)
+    notify = progress or (lambda _message: None)
+    source = str(channel.get("url") or channel.get("handle") or "").strip()
+    if not source:
+        raise RuntimeError("A conta seleccionada não tem URL pública ou handle Instagram.")
+    notify("A consultar o perfil público Instagram…")
+    profile_result = fetch_public_instagram_profile(source)
+    profile = dict(profile_result.data or {})
+    if not profile_result.ok and not profile:
+        raise RuntimeError(profile_result.message)
+    notify("A recolher as últimas 5 publicações públicas…")
+    try:
+        videos = _public_videos(source, 5)
+    except Exception as exc:
+        videos = []
+        profile["video_lookup_error"] = str(exc)[:240]
+    views = sum(_safe_int(item.get("view_count")) for item in videos)
+    likes = sum(_safe_int(item.get("like_count")) for item in videos)
+    comments = sum(_safe_int(item.get("comment_count")) for item in videos)
+    shares = sum(_safe_int(item.get("repost_count")) for item in videos)
+    saves = sum(_safe_int(item.get("save_count")) for item in videos)
+    engagement = ((likes + comments + shares + saves) / views * 100) if views else None
+    save_rate = (saves / views * 100) if views else None
+    share_rate = (shares / views * 100) if views else None
+    avg_views = views / len(videos) if videos else None
+    durations = [_safe_int(item.get("duration")) for item in videos if _safe_int(item.get("duration")) > 0]
+    avg_duration = sum(durations) / len(durations) if durations else None
+    followers = _number(profile.get("subscriber_count"))
+    conversion = (followers / views * 100) if followers is not None and views else None
+    cadence_score = 70 if len(videos) >= 3 else 40 if videos else 0
+    metrics = [
+        {"label": "Demanda e potencial de viralização", "score": 50 if videos else 0, "value": f"{len(videos)} publicações públicas", "source": "public", "diagnosis": "Alcance, impressões e pesquisa exigem Instagram Insights."},
+        {"label": "Qualidade visual — stop the scroll", "score": 50, "value": "Retenção inicial indisponível", "source": "unavailable", "diagnosis": "Skip rate e retenção de 3 segundos exigem Instagram Insights."},
+        {"label": "Hook e storytelling", "score": 50, "value": f"Duração média {avg_duration:.0f}s" if avg_duration else "Curva de retenção indisponível", "source": "public" if avg_duration else "unavailable", "diagnosis": "A duração pública não substitui watch time ou conclusão."},
+        {"label": "Ritmo e edição", "score": 50, "value": "Quedas de retenção indisponíveis", "source": "unavailable", "diagnosis": "Pacing depende da curva de retenção."},
+        {"label": "Tráfego Explore/Reels", "score": 50, "value": "Origem Explore/Reels indisponível", "source": "unavailable", "diagnosis": "Fontes Explore, Reels, Search, Profile e External exigem Insights."},
+        {"label": "Engajamento e saves", "score": _score_from_rate(engagement, (2, 3)), "value": f"{engagement:.2f}% na amostra" if engagement is not None else "Sem interacções públicas", "source": "public", "diagnosis": f"Saves: {save_rate:.2f}% · Shares: {share_rate:.2f}%" if save_rate is not None and share_rate is not None else "Interacções públicas indisponíveis."},
+        {"label": "Conversão em seguidores", "score": _score_from_rate(conversion, (0.5, 1.0)), "value": f"{conversion:.2f}% proxy" if conversion is not None else "Seguidores ganhos indisponíveis", "source": "estimated" if conversion is not None else "unavailable", "diagnosis": "Proxy não equivale a seguidores ganhos por publicação."},
+        {"label": "Mix de formatos", "score": 50, "value": "Reels/Carrosséis indisponíveis", "source": "unavailable", "diagnosis": "A distribuição de formatos requer o catálogo autorizado da conta."},
+        {"label": "Cadência", "score": cadence_score, "value": f"{len(videos)} publicações na amostra", "source": "public", "diagnosis": "Meta do agente: 3–5 Reels por semana, com consistência."},
+    ]
+    overall = round(sum(item["score"] for item in metrics) / len(metrics)) if metrics else 0
+    monetization = "A verificar — faltam dados privados" if followers is None or followers < 10000 else "Possível — confirmar conta profissional, pública e estado elegível"
+    record = {"code": code, "channel_id": str(channel.get("id") or ""), "channel_name": str(channel.get("name") or profile.get("name") or "Conta Instagram"), "platform": "Instagram", "created_at": datetime.now(timezone.utc).isoformat(), "overall_score": overall, "score_color": _score_color(overall), "metrics": metrics, "videos": videos, "profile": profile, "sample_views": views, "engagement_rate": engagement, "save_rate": save_rate, "share_rate": share_rate, "monetization_status": monetization, "status": "completed"}
+    report_path = root / f"{code}.md"
+    report_path.write_text(_report_markdown(record), encoding="utf-8")
+    record["report_path"] = str(report_path)
+    analyses = read_json(ANALYSES_FILENAME, [])
+    if not isinstance(analyses, list): analyses = []
+    analyses.append(record)
+    write_json(ANALYSES_FILENAME, analyses[-100:])
+    return record
+
+
+def list_analyses() -> list[dict[str, Any]]:
+    value = read_json(ANALYSES_FILENAME, [])
+    return [dict(item) for item in value if isinstance(item, Mapping)] if isinstance(value, list) else []
+
+
+def _card(title: str, score: Any, featured: bool, rows: list[tuple[str, str, str, str]]) -> str:
+    score_int = _safe_int(score)
+    row_html = "".join(f'<tr><td>{escape(label)}</td><td><b>{escape(value)}</b></td><td>{escape(target)}</td><td style="color:{colour};font-weight:700;">● {escape(status)}</td></tr>' for label, value, target, status, colour in rows)
+    return f'<div style="min-height:{"350px" if featured else "280px"};border:1px solid #2b3b52;border-radius:14px;padding:16px;background:linear-gradient(145deg,#111c2b,#0b121c);box-shadow:0 8px 22px rgba(0,0,0,.16);"><div style="font-size:{"19px" if featured else "16px"};font-weight:750;color:#f8fafc;margin-top:8px;">{escape(title)}</div><div style="display:flex;align-items:baseline;gap:5px;margin:8px 0 10px;"><span style="font-size:{"38px" if featured else "30px"};font-weight:850;color:{"#ef4444" if score_int <= 30 else "#f59e0b" if score_int < 70 else "#22c55e"};">{score_int}</span><span style="font-size:13px;color:#94a3b8;">/100</span></div><div style="overflow-x:auto;"><table style="width:100%;border-collapse:collapse;font-size:11px;color:#cbd5e1;"><thead><tr><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">KPI</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Valor</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Meta</th><th style="text-align:left;padding:5px 3px;border-bottom:1px solid #334155;">Status</th></tr></thead><tbody>{row_html}</tbody></table></div></div>'
+
+
+def render_growth_instagram() -> None:
+    st.title("Analista Growth Instagram")
+    st.caption("Auditoria pública baseada nos 3 pilares críticos e nas métricas secundárias do agente Growth Instagram.")
+    channels = [item for item in read_json("channels.json", []) if isinstance(item, dict) and item.get("active", True) and str(item.get("platform") or "").casefold() == "instagram"]
+    if not channels:
+        st.info("Cadastre pelo menos uma conta Instagram em Canais Instagram para iniciar a análise.")
+        return
+    selector_col, action_col = st.columns([1.05, 1.15], gap="small")
+    with selector_col:
+        selected = st.selectbox("Conta a analisar", channels, format_func=lambda item: str(item.get("name") or item.get("handle") or "Conta sem nome"), key="growth_instagram_channel")
+    with action_col:
+        st.write("")
+        analyse_clicked = st.button("ANALISAR CONTA", type="primary", use_container_width=True, key="growth_instagram_analyse")
+    st.markdown("**Estado das APIs de Growth**")
+    api_cols = st.columns(2, gap="small")
+    with api_cols[0]:
+        st.caption("Instagram página pública: perfil e amostra de vídeos")
+        st.markdown('<span style="color:#22c55e;font-weight:700;">● Disponível quando a página pública responder</span>', unsafe_allow_html=True)
+    with api_cols[1]:
+        st.caption("Instagram Studio: retenção, FYP, CTR e conversão")
+        st.markdown('<span style="color:#94a3b8;font-weight:700;">● Requer dados privados autorizados</span>', unsafe_allow_html=True)
+    if analyse_clicked:
+        with st.status("A preparar a análise pública…", expanded=True) as status:
+            try:
+                record = run_audit(selected, read_json("settings.json", {}), progress=st.write)
+                st.session_state["growth_instagram_last_code"] = record["code"]
+                status.update(label="Análise concluída", state="complete")
+            except Exception as exc:
+                status.update(label="Análise falhou", state="error")
+                st.error(str(exc)[:500])
+    records = list_analyses()
+    selected_record = next((item for item in reversed(records) if str(item.get("channel_id")) == str(selected.get("id"))), None)
+    metric_by_label = {str(item.get("label")): item for item in (selected_record.get("metrics", []) if selected_record else [])}
+    metrics = selected_record.get("metrics", []) if selected_record else [{"label": label, "score": 0, "value": "Aguardando análise", "source": "unknown"} for label in AGENT_METRICS]
+    score = selected_record.get("overall_score") if selected_record else None
+    score_color = "#64748b" if score is None else ("#ef4444" if score <= 30 else "#f59e0b" if score < 70 else "#22c55e")
+    summary_col, download_col = st.columns([3.25, 1], gap="large")
+    with summary_col:
+        st.markdown(f'<div style="border:1px solid #263447;border-radius:12px;padding:18px;margin:12px 0 20px;background:#101722;display:flex;align-items:center;justify-content:space-between;min-height:76px;"><div><div style="font-size:13px;color:#94a3b8;text-transform:uppercase;letter-spacing:.08em;">Nota geral da conta</div><div style="font-size:18px;font-weight:700;margin-top:5px;">{escape(str(selected.get("name") or selected.get("handle") or "Conta Instagram"))}</div></div><div style="font-size:42px;font-weight:800;color:{score_color};line-height:1;">{score if score is not None else "—"}<span style="font-size:16px;color:#94a3b8;">{("/100" if score is not None else "")}</span></div></div>', unsafe_allow_html=True)
+    with download_col:
+        st.write("")
+        report_path = Path(str(selected_record.get("report_path") or "")) if selected_record else Path()
+        if report_path.is_file():
+            st.download_button("BAIXAR ANALISE COMPLETA", report_path.read_bytes(), file_name=report_path.name, mime="text/markdown", use_container_width=True, key=f"instagram_download_{selected_record['code']}")
+        else:
+            st.button("BAIXAR ANALISE COMPLETA", disabled=True, use_container_width=True, key="instagram_download_disabled")
+    st.subheader("Dashboard de Growth")
+    st.caption("3 pilares críticos em destaque e camadas operacionais abaixo. Vermelho: 0–30 · Amarelo: 31–69 · Verde: 70–100.")
+    def item(label: str, target: str, available: bool = False) -> tuple[str, str, str, str, str]:
+        metric = metric_by_label.get(label, {})
+        value = str(metric.get("value", "Aguardando análise"))
+        return (label, value, target, "A verificar" if not available else "Meta atingida", "#94a3b8" if not available else "#22c55e")
+    cards = [
+        ("Demanda e potencial de viralização", True, [item("Demanda e potencial de viralização", "> 60% Explore/Reels", True), ("Alcance", "Dados privados", "Crescente", "A verificar", "#94a3b8")]),
+        ("Qualidade visual — stop the scroll", True, [item("Qualidade visual — stop the scroll", "> 70% retenção 3s"), ("Skip rate", "Dados privados", "< 60%", "A verificar", "#94a3b8")]),
+        ("Hook e storytelling", True, [item("Hook e storytelling", "> 30% conclusão"), ("Watch time", "Dados privados", "próximo da duração", "A verificar", "#94a3b8")]),
+        ("Ritmo e edição", False, [item("Ritmo e edição", "Pattern interrupt 5–8s")]),
+        ("Tráfego Explore/Reels", False, [item("Tráfego Explore/Reels", "> 60%")]),
+        ("Engajamento e saves", False, [item("Engajamento e saves", "> 3% engagement"), ("Save rate", "Dados privados", "> 2%", "A verificar", "#94a3b8")]),
+        ("Conversão em seguidores", False, [item("Conversão em seguidores", "> 1%")]),
+        ("Mix de formatos", False, [item("Mix de formatos", "80% Reels")]),
+        ("Cadência", False, [item("Cadência", "3–5 Reels/semana")]),
+    ]
+    st.markdown('<div style="display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:14px;margin:14px 0 18px;">' + "".join(_card(title, metric_by_label.get(title, {}).get("score", 0), featured, rows) for title, featured, rows in cards[:3]) + '</div>', unsafe_allow_html=True)
+    st.markdown('<div style="display:grid;grid-template-columns:repeat(5,minmax(0,1fr));gap:12px;margin:0 0 18px;">' + "".join(_card(title, metric_by_label.get(title, {}).get("score", 0), featured, rows) for title, featured, rows in cards[3:]) + '</div>', unsafe_allow_html=True)
+    st.caption("Legenda: ● Verde = meta atingida · ● Amarelo = atenção · ● Vermelho = crítico · Cinzento = dado não disponível na fonte autorizada.")
+    st.divider()
+    history_tab, dashboard_tab = st.tabs(["Últimas análises", "Leitura estratégica"])
+    with dashboard_tab:
+        st.write("Priorize Hook, retenção e cadência. CTR, FYP, retenção real e seguidores ganhos ficam pendentes de dados autorizados do Instagram Studio.")
+    with history_tab:
+        if not records:
+            st.info("Ainda não existem análises guardadas.")
+        for record in reversed(records[-20:]):
+            row = st.columns([2, 1, 2.2, 1.2, 1, 1.4])
+            row[0].write(f"**{record.get('channel_name', 'Conta Instagram')}**")
+            row[1].write("Instagram")
+            row[2].caption(str(record.get("code", "")))
+            row[3].caption(str(record.get("created_at", "")).replace("T", " ")[:19])
+            row[4].write(f"{record.get('overall_score', '—')}/100")
+            path = Path(str(record.get("report_path") or ""))
+            if path.is_file():
+                row[5].download_button("Análise", path.read_bytes(), file_name=path.name, mime="text/markdown", key=f"instagram_history_{record.get('code')}")
+
+
+__all__ = ["AGENT_METRICS", "ANALYSES_DIR", "analysis_code", "list_analyses", "run_audit", "render_growth_instagram"]

package/integrations/instagram_public.py

@@ -0,0 +1,76 @@
+from __future__ import annotations
+
+import re
+from dataclasses import dataclass
+from datetime import datetime, timezone
+from html import unescape
+from typing import Any
+from urllib.parse import urlparse
+
+import requests
+
+
+@dataclass
+class IntegrationResult:
+    ok: bool
+    message: str
+    data: dict[str, Any]
+
+
+def normalize_instagram_reference(source: str) -> dict[str, str]:
+    value = str(source or '').strip()
+    if not value:
+        raise ValueError('Informe um @handle ou URL pública do Instagram.')
+    if value.startswith('@'):
+        username = value[1:]
+    elif value.startswith(('http://', 'https://')):
+        parsed = urlparse(value)
+        if (parsed.hostname or '').lower().removeprefix('www.') != 'instagram.com':
+            raise ValueError('Use uma URL pública do Instagram, por exemplo https://www.instagram.com/conta/.')
+        username = next((part for part in parsed.path.split('/') if part and not part.startswith(('p', 'reel', 'tv'))), '')
+    else:
+        username = value
+    username = username.strip().lstrip('@').split('/', 1)[0]
+    if not re.fullmatch(r'[A-Za-z0-9._]{1,64}', username):
+        raise ValueError('O @handle Instagram contém caracteres inválidos.')
+    return {'username': username, 'handle': f'@{username}', 'url': f'https://www.instagram.com/{username}/'}
+
+
+def _meta(document: str, name: str) -> str:
+    patterns = (
+        rf'<meta[^>]+(?:name|property)=["\']{re.escape(name)}["\'][^>]+content=["\']([^"\']*)',
+        rf'<meta[^>]+content=["\']([^"\']*)["\'][^>]+(?:name|property)=["\']{re.escape(name)}["\']',
+    )
+    for pattern in patterns:
+        match = re.search(pattern, document, flags=re.IGNORECASE)
+        if match:
+            return unescape(match.group(1)).strip()
+    return ''
+
+
+def fetch_public_instagram_profile(source: str) -> IntegrationResult:
+    try:
+        reference = normalize_instagram_reference(source)
+    except ValueError as exc:
+        return IntegrationResult(False, str(exc), {})
+    try:
+        response = requests.get(reference['url'], headers={'User-Agent': 'Mozilla/5.0', 'Accept-Language': 'pt-BR,pt;q=0.9,en;q=0.8'}, timeout=12)
+    except requests.RequestException as exc:
+        return IntegrationResult(False, f'Não foi possível consultar o perfil público do Instagram: {exc}', reference)
+    if response.status_code >= 400:
+        return IntegrationResult(False, f'O Instagram devolveu HTTP {response.status_code}. Confirme o @handle ou use o cadastro manual.', reference | {'status_code': response.status_code})
+    title = _meta(response.text, 'og:title') or reference['username']
+    description = _meta(response.text, 'og:description')
+    avatar_url = _meta(response.text, 'og:image')
+    followers = None
+    match = re.search(r'([\d,.]+)\s*(?:followers|seguidores)', description, flags=re.IGNORECASE)
+    if match:
+        try:
+            followers = int(re.sub(r'[^0-9]', '', match.group(1)))
+        except ValueError:
+            followers = None
+    data = {'id': f"instagram_{reference['username']}", **reference, 'name': title.split('(')[0].strip() or reference['username'], 'bio': description, 'avatar_url': avatar_url, 'subscriber_count': followers, 'public_lookup': True, 'metrics_source': 'instagram_public_page', 'last_public_lookup_at': datetime.now(timezone.utc).isoformat()}
+    return IntegrationResult(True, 'Perfil Instagram encontrado publicamente.', data)
+
+
+__all__ = ['IntegrationResult', 'fetch_public_instagram_profile', 'normalize_instagram_reference']

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@danhachuel/thunderbolt",
-  "version": "0.6.11",
+  "version": "0.6.13",
   "description": "Thunderbolt — interface local para operação de canais faceless e motor MoneyPrinterTurbo",
   "license": "MIT",
   "main": "scripts/cli.mjs",