refactor: vwts_scraper — proper Python package structure

- Split single file into 8 focused modules (399 lines total):
  config.py — constants
  fetch.py — HTTP GET with retries
  paginate.py — page_count + HEAD verification
  parse.py — HTML parsing (topics, posts, tags)
  state.py — per-section state.json
  output.py — JSONL with file rotation
  run.py — main orchestration loop
  __main__.py — CLI entry point
- Run: python -m vwts_scraper <URL>
- Page count method fully tested (13 sections, 100+ topics, 0 errors)
- Resume support (Ctrl+C → state saved)
- Per-section state files (no conflicts between sections)

vwts_scraper/state.py

@@ -0,0 +1,24 @@
+"""Управление состоянием (state.json в папке раздела)."""
+
+import json
+from pathlib import Path
+
+OUTPUT_DIR = Path("vwts_data")
+
+
+def _state_file(section_slug: str) -> Path:
+    d = OUTPUT_DIR / section_slug
+    d.mkdir(parents=True, exist_ok=True)
+    return d / "state.json"
+
+
+def load(section_slug: str) -> dict:
+    f = _state_file(section_slug)
+    if f.exists():
+        return json.loads(f.read_text(encoding="utf-8"))
+    return {"topics_done": {}, "pages_done": [], "file_index": 0, "topic_count": 0}
+
+
+def save(section_slug: str, st: dict):
+    _state_file(section_slug).write_text(
+        json.dumps(st, ensure_ascii=False, indent=2), encoding="utf-8")