nedviga-cli 0.1.0__py3-none-any.whl

nedviga/export_html.py

@@ -0,0 +1,280 @@
+"""v0.3.5 (ТЗ реалтора): экспорт карточек в автономный HTML-каталог.
+
+Команда: nedviga-cli export --input cards.json --format html
+         [--embed-photos] [--output catalog.html]
+
+Карусель: чистый CSS/JS без внешних зависимостей (offline-first).
+Фото: --embed-photos вшивает base64 (сжатие до 800px, JPEG q=70),
+иначе — прямые ссылки (требуют сеть при просмотре).
+"""
+from __future__ import annotations
+
+import base64
+import html as html_mod
+import io
+import json
+from pathlib import Path
+
+from .models import ListingCard
+
+_EMBED_MAX_BYTES = 300_000  # на фото после сжатия
+
+
+def _fetch_photo_bytes(url: str, timeout: float = 15.0) -> bytes | None:
+    try:
+        # v0.3.5: netguard (egress-guard + breaker)
+        from .net_guard import guarded_get
+
+        # CDN Циан требует браузерный UA + Referer, иначе 403 (v0.3.4)
+        resp = guarded_get(
+            url,
+            timeout=timeout,
+            stream=True,
+            headers={
+                "User-Agent": (
+                    "Mozilla/5.0 (X11; Linux x86_64; rv:130.0) "
+                    "Gecko/20100101 Firefox/130.0"
+                ),
+                "Referer": "https://www.cian.ru/",
+                "Accept": "image/avif,image/webp,image/png,image/jpeg,*/*",
+            },
+        )
+        if resp.status_code != 200:
+            return None
+        return resp.content
+    except Exception:
+        return None
+
+
+def _compress(data: bytes, max_side: int = 800, quality: int = 70) -> bytes | None:
+    try:
+        from PIL import Image
+
+        img = Image.open(io.BytesIO(data)).convert("RGB")
+        w, h = img.size
+        if max(w, h) > max_side:
+            scale = max_side / max(w, h)
+            img = img.resize((int(w * scale), int(h * scale)), Image.LANCZOS)
+        buf = io.BytesIO()
+        img.save(buf, format="JPEG", quality=quality, optimize=True)
+        return buf.getvalue()
+    except Exception:
+        return None
+
+
+def _photo_src(card: ListingCard, idx: int, embed: bool) -> str | None:
+    photos = [p.url for p in (card.images or []) if getattr(p, "url", None)]
+    if not photos:
+        return None
+    url = photos[0]  # в карусель идут все, но первый — обложка
+    if not embed:
+        return url
+    data = _fetch_photo_bytes(url)
+    if data is None:
+        return url  # fallback на ссылку
+    compressed = _compress(data)
+    if compressed is None or len(compressed) > _EMBED_MAX_BYTES * 2:
+        return url
+    return "data:image/jpeg;base64," + base64.b64encode(compressed).decode("ascii")
+
+
+def _safe_url(u: object) -> str:
+    """v0.4.4 (независимый аудит sa-2, F-04): allow-list схем + экранирование.
+
+    Внешние данные (images/trading_link из import cards.json) не должны
+    превращаться в XSS в автономном HTML-каталоге. Пустая строка = не выводить.
+    """
+    s = str(u or "").strip()
+    low = s.lower()
+    # v0.4.5 (review R3): data:image/svg+xml может нести <script> — allow-list только
+    # растров; наш собственный embed всегда jpeg, легитимных потребителей svg нет.
+    if not (low.startswith("http://") or low.startswith("https://")
+            or low.startswith("data:image/jpeg") or low.startswith("data:image/png")
+            or low.startswith("data:image/webp") or low.startswith("data:image/gif")):
+        return ""
+    return html_mod.escape(s, quote=True)
+
+
+def _photos_all(card: ListingCard, embed: bool, limit: int = 6) -> list[str]:
+    out: list[str] = []
+    for p in (card.images or [])[:limit]:
+        url = getattr(p, "url", None)
+        if not url:
+            continue
+        if not embed:
+            out.append(url)
+            continue
+        data = _fetch_photo_bytes(url)
+        compressed = _compress(data) if data else None
+        if compressed:
+            out.append("data:image/jpeg;base64," + base64.b64encode(compressed).decode("ascii"))
+        else:
+            out.append(url)
+    return out
+
+
+def _card_html(card: ListingCard, embed: bool, idx: int) -> str:
+    # P3-2: внешний текст мог прийти из cards.json с spotlight-маркерами —
+    # для human-facing каталога снимаем их (HTML и так экранируется).
+    from .net_guard import strip_spotlight
+    title_escaped = html_mod.escape(strip_spotlight(card.title or ""))
+    # v0.4.5 (review R3): card.id — внешние данные (import cards.json), в фолбэке экранируем
+    title = title_escaped or f"Объявление {html_mod.escape(str(card.id), quote=True)}"
+    price = f"{card.pricing.price_total_rub:,.0f}".replace(",", " ") if card.pricing.price_total_rub else "—"
+    area = card.parameters.area_total_m2
+    rooms = card.parameters.rooms
+    floor = card.parameters.floor
+    floors_total = card.parameters.floors_total
+    year = card.building.year_built
+    sale_type = card.pricing.sale_type
+    is_auction = card.pricing.is_auction
+    metro = card.geo.cian_claimed_metro.station if card.geo.cian_claimed_metro else None
+    walk = card.geo.honest_walk_route
+
+    walk_html = ""
+    if walk:
+        overclaim_flag = ""
+        oc = card.geo.honest_overclaim
+        if oc and oc.get("flag"):
+            overclaim_flag = (' <span style="color:#c0392b">⚠ Циан занижает на '
+                              f'{html_mod.escape(str(oc["overclaim_minutes"]), quote=True)} мин</span>')
+        walk_html = (
+            f'<div class="walk">🚇 {html_mod.escape(walk.target_station or "")}: '
+            f'{html_mod.escape(str(walk.walk_distance_meters), quote=True)} м ≈ {walk.walk_duration_minutes:.0f} мин '
+            f'<span class="prov">({html_mod.escape(str(walk.routing_provider), quote=True)})</span>{overclaim_flag}</div>'
+        )
+
+    auction_html = ""
+    if is_auction:
+        link = _safe_url(card.pricing.trading_link)  # v0.4.4 F-04: схемы фильтрует
+        auction_html = (
+            f'<div class="auction">🔨 Аукцион'
+            + (f' — <a href="{link}">торги</a>' if link else "")
+            + "</div>"
+        )
+
+    photos = _photos_all(card, embed)
+    slides = ""
+    dots = ""
+    if photos:
+        slides = "".join(
+            f'<img src="{_safe_url(p)}" alt="фото {idx}-{i}" class="slide{" active" if i == 0 else ""}">'
+            for i, p in enumerate(photos) if _safe_url(p)
+        )
+        dots = "".join(
+            f'<span class="dot{" active" if i == 0 else ""}" onclick="showSlide({idx},{i},this)"></span>'
+            for i in range(len(photos))
+        )
+    carousel = f'<div class="carousel" id="car{idx}">{slides}<div class="dots">{dots}</div></div>' if photos else ""
+
+    # v0.4.6 (the task brief F-13): контакты продавца. Сюда попадают уже маскированные
+    # card.seller.phones — export_html() применяет sanitize_contacts по умолчанию;
+    # сырые номера возможны только при явном with_contacts=True.
+    contacts_html = ""
+    seller_phones = [
+        str(p) for p in (card.seller.phones if card.seller else []) if str(p or "").strip()
+    ]
+    if seller_phones:
+        contacts_html = (
+            '<div class="contacts">📞 '
+            + " · ".join(html_mod.escape(p) for p in seller_phones)
+            + "</div>"
+        )
+
+    return f"""
+<div class="card" id="c{idx}">
+  {carousel}
+  <h3>{title}</h3>
+  <div class="price">{price} ₽</div>
+  <div class="meta">
+    {'комн: ' + str(rooms) if rooms is not None else ''} {'· ' if rooms is not None else ''}{area or '—'} м²
+    {'· этаж ' + str(floor) + '/' + str(floors_total) if floor else ''}
+    {'· ' + str(year) + ' г.' if year else ''}
+    {'· ' + html_mod.escape(sale_type) if sale_type else ''}
+  </div>
+  {walk_html}
+  {auction_html}
+  {contacts_html}
+  <div class="url"><a href="{html_mod.escape(card.url)}">{html_mod.escape(card.url)}</a></div>
+</div>"""
+
+
+_TEMPLATE = """<!DOCTYPE html>
+<html lang="ru"><head><meta charset="utf-8">
+<title>Каталог квартир — nedviga-cli</title>
+<style>
+body{font-family:system-ui,sans-serif;background:#f5f6f7;margin:20px}
+h1{font-size:1.4em}
+.card{background:#fff;border-radius:10px;padding:16px;margin:14px 0;box-shadow:0 1px 4px rgba(0,0,0,.08)}
+.price{font-size:1.3em;font-weight:700;color:#0a7d34}
+.meta{color:#555;margin:6px 0}
+.walk{margin:6px 0;color:#2c3e50}
+.prov{color:#999;font-size:.85em}
+.auction{color:#8e44ad;font-weight:600}
+.contacts{color:#666;margin:4px 0;font-size:.95em}
+.url a{color:#2471a3;font-size:.85em;word-break:break-all}
+.carousel{position:relative;height:240px;overflow:hidden;border-radius:8px;margin-bottom:10px;background:#eee}
+.slide{display:none;width:100%;height:240px;object-fit:cover}
+.slide.active{display:block}
+.dots{position:absolute;bottom:8px;left:0;right:0;text-align:center}
+.dot{display:inline-block;width:10px;height:10px;border-radius:50%;background:#fff;opacity:.6;margin:0 3px;cursor:pointer}
+.dot.active{opacity:1;background:#0a7d34}
+.navbtn{position:absolute;top:50%;transform:translateY(-50%);background:rgba(255,255,255,.8);
+ border:none;border-radius:50%;width:34px;height:34px;font-size:18px;cursor:pointer}
+.navbtn.prev{left:8px}.navbtn.next{right:8px}
+</style></head><body>
+<h1>Каталог квартир — nedviga-cli</h1>
+{cards}
+<script>
+function showSlide(ci, si, dot) {{
+  const car = document.getElementById('car' + ci);
+  car.querySelectorAll('.slide').forEach((s, i) => s.classList.toggle('active', i === si));
+  car.querySelectorAll('.dot').forEach((d, i) => d.classList.toggle('active', i === si));
+}}
+document.querySelectorAll('.carousel').forEach((car, ci) => {{
+  const prev = document.createElement('button'); prev.textContent = '‹'; prev.className = 'navbtn prev';
+  const next = document.createElement('button'); next.textContent = '›'; next.className = 'navbtn next';
+  const go = (d) => {{
+    const slides = [...car.querySelectorAll('.slide')];
+    const dots = [...car.querySelectorAll('.dot')];
+    let cur = dots.findIndex(d => d.classList.contains('active'));
+    cur = (cur + d + slides.length) % slides.length;
+    showSlide(ci, cur, dots[cur]);
+  }};
+  prev.onclick = () => go(-1); next.onclick = () => go(1);
+  car.appendChild(prev); car.appendChild(next);
+}});
+</script>
+</body></html>"""
+
+
+def export_html(
+    cards: list[ListingCard],
+    output: Path,
+    embed_photos: bool = False,
+    with_contacts: bool = False,
+) -> Path:
+    """Собрать автономный HTML-каталог. Возвращает путь к файлу.
+
+    v0.4.6 (the task brief F-13): fail-closed — по умолчанию телефоны во всех
+    строках карточек маскируются (``sanitize_contacts``), и в каталог
+    попадает только маска. Сырые контакты выводятся лишь при явном
+    ``with_contacts=True`` (CLI-флаг ``--with-contacts``).
+    """
+    if with_contacts:
+        prepared = list(cards)
+    else:
+        from .pii import sanitize_contacts
+
+        prepared = [
+            ListingCard(
+                **sanitize_contacts(card.model_dump(mode="json", exclude_none=False))
+            )
+            for card in cards
+        ]
+    parts = [_card_html(c, embed_photos, i) for i, c in enumerate(prepared)]
+    html = _TEMPLATE.replace("{cards}", "\n".join(parts))
+    # Убираем лишние {{}} из f-string-безопасного шаблона
+    html = html.replace("{{", "{").replace("}}", "}")
+    output.write_text(html, encoding="utf-8")
+    return output

nedviga/export_pdf.py

@@ -0,0 +1,96 @@
+"""v0.4.7 (the task brief) F-11: PDF export via WeasyPrint (lazy import)."""
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Any
+
+from .family_card import FamilyCard, build_family_card, render_html
+
+
+def _as_html(card: Any, with_contacts: bool = False) -> str:
+    if isinstance(card, FamilyCard):
+        fc = card
+    else:
+        fc = build_family_card(card, with_contacts=with_contacts)
+    return render_html(fc, with_contacts=with_contacts)
+
+
+def render_pdf(card: Any, with_contacts: bool = False) -> bytes:
+    """Render PDF bytes via WeasyPrint. Raises RuntimeError('pdf-format-unavailable') if missing."""
+    html = _as_html(card, with_contacts=with_contacts)
+    try:
+        from weasyprint import HTML
+    except Exception as exc:  # pragma: no cover - depends on env
+        raise RuntimeError("pdf-format-unavailable") from exc
+    try:
+        return HTML(string=html).write_pdf()
+    except Exception as exc:
+        raise RuntimeError(f"pdf-render-failed: {exc}") from exc
+
+
+def export_pdf(card: Any, output: Any, with_contacts: bool = False) -> Path:
+    """Write PDF to output path. Returns Path."""
+    out = Path(output)
+    data = render_pdf(card, with_contacts=with_contacts)
+    out.write_bytes(data)
+    return out
+
+
+def catalog_pdf(cards, output, title=None, profile=None) -> "Path":
+    cards = list(cards or [])
+    if len(cards) > 30:
+        raise RuntimeError("too_many")
+    # limit photos to 6 per card
+    prepared = []
+    for c in cards:
+        try:
+            import copy
+            cc = copy.copy(c)
+            if hasattr(cc, "images"):
+                cc.images = list(cc.images or [])[:6]
+            prepared.append(cc)
+        except Exception:
+            prepared.append(c)
+    # title page info
+    districts = []
+    for c in prepared:
+        try:
+            d = c.geo.district if getattr(c, "geo", None) else None
+            if d:
+                districts.append(d)
+        except Exception:
+            pass
+    district_str = ", ".join(sorted(set(districts))) if districts else "—"
+    title = title or "Каталог подбора"
+    date_str = __import__("datetime").date.today().isoformat()
+    who = profile or "nedviga-cli"
+    parts = []
+    parts.append(f"<h1>{__import__('html').escape(title)}</h1>")
+    parts.append(f"<p>Дата: {date_str}</p>")
+    parts.append(f"<p>Район: {__import__('html').escape(district_str)}</p>")
+    parts.append(f"<p>Объектов: {len(prepared)}</p>")
+    parts.append(f"<p>Собрал: {__import__('html').escape(str(who))}</p>")
+    parts.append("<hr/>")
+    for c in prepared:
+        try:
+            from .family_card import build_family_card, render_html
+            fc = build_family_card(c)
+            html = render_html(fc, with_contacts=False)
+            # extract body content roughly: keep full html but page break
+            parts.append('<div style="page-break-after: always;"></div>')
+            parts.append(html)
+        except Exception as exc:
+            parts.append(f"<p>render error: {__import__('html').escape(str(exc))}</p>")
+    full_html = "<!DOCTYPE html><html><head><meta charset='utf-8'><style>body{font-family:system-ui}</style></head><body>" + "\n".join(parts) + "</body></html>"
+    try:
+        from weasyprint import HTML
+    except Exception as exc:
+        raise RuntimeError("pdf-format-unavailable") from exc
+    out = Path(output)
+    HTML(string=full_html).write_pdf(str(out))
+    return out
+
+
+# aliases
+pdf_from_card = export_pdf
+to_pdf = export_pdf