#!/usr/bin/env python3 """Извлечение таблиц совместимости Yealink через pdfplumber (финальный подход). Находит страницы с заголовком "Accessories Compatibility" и берёт таблицы с колонками Device | Accessories | Max Qty | Note. """ import os import re import sys OUT = os.environ.get("COMPAT_OUT", "/opt/work-hub/compat") def role_of(m): for pat, role in [ (r"^CPW\d|^VCM\d|^CP9\d\d|^CP7\d\d|^MSpeaker|^CM\d|^CS\d|^BT\d|Dect", "audio"), (r"^CTP\d", "touch"), (r"^UVC\d", "camera"), (r"^VCH\d|^WPP\d|^WF\d|^DD1\d", "hub_or_sharing"), (r"^VCR\d", "remote"), ]: if re.match(pat, m, re.I): return role return "other" def clean(x): if x is None: return "" return re.sub(r"\s+", " ", str(x)).strip() JUNK_RE = re.compile( r"^(Recommended solution|Name|Custom Key|Mute Key|Device|Accessories|" r"\u2022|Note|Model|Description)", re.I) def is_junk(model): """Отсекаем не-аксессуары: длинные фразы, заголовки других таблиц, маркеры.""" if not model or JUNK_RE.match(model): return True if model.startswith("\u2022") or model.startswith("-"): return True if len(model) > 26: # длинная фраза, а не модель return True if len(model.split()) > 3: # «VCS device: VC800 Pickup device» return True if ":" in model: return True return False def parse(pdf, plumber): devices = {} cur = None with plumber.open(pdf) as doc: for page in doc.pages: txt = page.extract_text() or "" # определяем устройство на странице m = re.search(r"Compatibility of (.+?) with accessories", txt, re.I) if m: cur = m.group(1).strip() devices.setdefault(cur, []) if not cur: continue if "Accessories Compatibility" not in txt and "Max Qty" not in txt: continue for table in page.extract_tables() or []: # MAXQTY_GATE head = " ".join(clean(c) for c in (table[0] if table else [])).lower() if "max" not in head and "qty" not in head: continue for row in table: cells = [clean(c) for c in (row or [])] if len(cells) < 3: continue model, qty, note = cells[1] if len(cells) > 3 else cells[1], None, "" # типовой порядок: [None/cat, model, qty, note] if len(cells) >= 4: model, qty, note = cells[1], cells[2], " ".join(cells[3:]) else: model, qty = cells[1], cells[2] note = " ".join(cells[3:]) if len(cells) > 3 else "" if is_junk(model): continue try: q = int(qty) except (TypeError, ValueError): q = 0 if str(qty).lower().startswith("nonsupport") else None devices[cur].append({"model": model, "role": role_of(model), "max_qty": q, "note": note[:200] or None}) return devices def dedupe(acc): """DEDUPE: убираем повторы модели внутри одного устройства (multi-page таблицы).""" seen, out = {}, [] for a in acc: key = a["model"].lower().replace(" ", "") if key in seen: prev = seen[key] # оставляем максимальный лимит и самую длинную заметку if a["max_qty"] and (prev["max_qty"] is None or a["max_qty"] > prev["max_qty"]): prev["max_qty"] = a["max_qty"] if a["note"] and len(a["note"]) > len(prev["note"] or ""): prev["note"] = a["note"] else: seen[key] = dict(a) out.append(seen[key]) return out def main(): import pdfplumber import yaml if len(sys.argv) < 2: print("использование: parse_compat3.py ...") return 1 merged = {} for pdf in sys.argv[1:]: devs = parse(pdf, pdfplumber) for name, acc in devs.items(): acc = dedupe(acc) if not acc: continue key = name.lower() if key in merged: have = {a["model"] for a in merged[key]["accessories"]} merged[key]["accessories"] += [a for a in acc if a["model"] not in have] else: merged[key] = {"device": name, "accessories": acc, "source": os.path.basename(pdf)} total = 0 for key, d in sorted(merged.items()): safe = re.sub(r"[^A-Za-z0-9._-]+", "_", d["device"])[:60] os.makedirs(os.path.join(OUT, "yealink"), exist_ok=True) with open(os.path.join(OUT, "yealink", safe + ".yaml"), "w", encoding="utf-8") as f: yaml.safe_dump({"compatibility": { "vendor": "yealink", "device": d["device"], "source": {"file": d["source"], "kind": "official_guide", "verified": True}, "accessories": d["accessories"]}}, f, allow_unicode=True, sort_keys=False, width=160) total += len(d["accessories"]) print(" %-24s аксессуаров: %2d" % (d["device"][:24], len(d["accessories"]))) print("\nустройств: %d | рёбер: %d" % (len(merged), total)) if __name__ == "__main__": sys.exit(main())