scripts/reviews/01_scrape.py

#!/usr/bin/env python3
"""
scrape_reviews.py — Fetch the RocketReviews.com review index and scrape each
detail page, saving structured JSON to source/reviews/.

Output
------
    source/reviews/index.jsonl        one record per review (raw index fields)
    source/reviews/detail/{id}.json   full parsed detail per review

Usage
-----
    python scripts/scrape_reviews.py
    python scripts/scrape_reviews.py --delay 2.0 --limit 10
    python scripts/scrape_reviews.py --force   # re-scrape existing files
"""

from __future__ import annotations

import argparse
import json
import logging
import re
import sys
import time
from datetime import datetime, timezone
from pathlib import Path
from typing import Optional

import requests
from bs4 import BeautifulSoup, NavigableString, Tag
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry

# ---------------------------------------------------------------------------
# Config
# ---------------------------------------------------------------------------

BASE_URL = "https://www.rocketreviews.com"
INDEX_URL = f"{BASE_URL}/data/reviews/reviews.php"
USER_AGENT = "RocketReviews-Dataset/1.0"
DEFAULT_DELAY = 1.0  # seconds between requests

ROOT = Path(__file__).parent.parent.parent
SOURCE_DIR = ROOT / "source" / "reviews"
DETAIL_DIR = SOURCE_DIR / "detail"

# ---------------------------------------------------------------------------
# Logging
# ---------------------------------------------------------------------------

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s %(levelname)s %(message)s",
    handlers=[logging.StreamHandler(sys.stdout)],
)
log = logging.getLogger(__name__)

# ---------------------------------------------------------------------------
# HTTP session
# ---------------------------------------------------------------------------


def _build_session() -> requests.Session:
    s = requests.Session()
    s.headers["User-Agent"] = USER_AGENT
    retry = Retry(
        total=3,
        backoff_factor=2.0,
        status_forcelist=[429, 500, 502, 503, 504],
        allowed_methods=["GET"],
    )
    s.mount("https://", HTTPAdapter(max_retries=retry))
    s.mount("http://", HTTPAdapter(max_retries=retry))
    return s


class RateLimiter:
    def __init__(self, delay: float) -> None:
        self.delay = delay
        self._last: float = 0.0

    def wait(self) -> None:
        elapsed = time.monotonic() - self._last
        if elapsed < self.delay:
            time.sleep(self.delay - elapsed)
        self._last = time.monotonic()


# ---------------------------------------------------------------------------
# Parsing helpers
# ---------------------------------------------------------------------------


def _re_first(pattern: str, text: str, group: int = 1) -> Optional[str]:
    m = re.search(pattern, text)
    return m.group(group) if m else None


def _parse_rating_block(soup: BeautifulSoup, label: str) -> Optional[int]:
    """
    Find '<strong>{label} Rating:</strong>' and count filled Material Icons
    stars ('star') that follow it, excluding 'star_border' (empty stars).
    """
    strong = soup.find("strong", string=re.compile(rf"^{label} Rating:?$", re.I))
    if not strong:
        return None

    # Collect raw text from siblings until the next <strong> tag
    text = ""
    for sib in strong.next_siblings:
        if isinstance(sib, Tag) and sib.name == "strong":
            break
        text += sib.get_text() if isinstance(sib, Tag) else str(sib)

    # Count 'star' occurrences that are NOT 'star_border'
    filled = len(re.findall(r"\bstar\b(?!_border)", text))
    return filled if filled else None


def _parse_ratings(soup: BeautifulSoup) -> dict:
    return {
        "construction": _parse_rating_block(soup, "Construction"),
        "flight": _parse_rating_block(soup, "Flight"),
        "overall": _parse_rating_block(soup, "Overall"),
    }


def _parse_product(soup: BeautifulSoup) -> dict:
    html = str(soup)

    # Diameter: /kit-index-diameter-13260.html → 13260 / 10000 = 1.3260 inches
    diameter_raw = _re_first(r"/kit-index-diameter-(\d+)\.html", html)
    diameter = int(diameter_raw) / 10000 if diameter_raw else None

    # Length: /kit-index-length-290000.html → 290000 / 10000 = 29.0000 inches
    length_raw = _re_first(r"/kit-index-length-(\d+)\.html", html)
    length = int(length_raw) / 10000 if length_raw else None

    # Skill level: /kit-index-skilllevel-1.html → 1
    skill_raw = _re_first(r"/kit-index-skilllevel-(\d+)\.html", html)
    skill = int(skill_raw) if skill_raw else None

    # Style: /kit-index-style-{name}-{id}.html — one or more values
    styles = [
        m.replace("-", " ").title()
        for m in re.findall(r"/kit-index-style-([a-z][a-z-]*)-\d+\.html", html)
    ]

    # Price: first dollar amount mentioned in the page text
    price_raw = _re_first(r"\$\s*(\d+(?:\.\d{1,2})?)", soup.get_text())
    price = float(price_raw) if price_raw else None

    return {
        "diameter_in": diameter,
        "length_in": length,
        "skill_level": skill,
        "style": styles,
        "price_usd": price,
    }


# Section headings that contain grids/tables rather than review text
_SKIP_SECTION_RE = re.compile(
    r"^\s*(flights?|.*reviews?|what you can do|sign in|create account)\s*$",
    re.I,
)


def _parse_sections(soup: BeautifulSoup) -> dict[str, str]:
    """
    Walk all h2/h4 headings and collect the plain text that follows each one
    until the next heading of the same or higher level. Skips headings that
    correspond to data grids (flight logs, related reviews).
    """
    sections: dict[str, str] = {}

    for heading in soup.find_all(["h2", "h4"]):
        title = heading.get_text(strip=True)
        if not title or _SKIP_SECTION_RE.match(title):
            continue

        parts: list[str] = []
        for sib in heading.next_siblings:
            if isinstance(sib, Tag) and sib.name in ("h2", "h4"):
                break
            if isinstance(sib, Tag):
                text = sib.get_text(separator=" ", strip=True)
            elif isinstance(sib, NavigableString):
                text = str(sib).strip()
            else:
                continue
            if text:
                parts.append(text)

        content = " ".join(parts).strip()
        if content:
            sections[title] = content

    return sections


def _parse_detail(html: str, index_rec: dict) -> dict:
    soup = BeautifulSoup(html, "lxml")

    # Canonical URL from any /review-{N}.html href in the page
    canonical_path = _re_first(r'href=["\']?(/review-\d+\.html)', html)
    canonical_url = f"{BASE_URL}{canonical_path}" if canonical_path else None

    # Alias URL — slug URL from the index
    alias_path = index_rec.get("url", "")
    url_alias = f"{BASE_URL}{alias_path}" if alias_path else None

    # Kit URL constructed from the Kit_ID in the JS data source
    # e.g. data/flightlog/flights.php?column=Kit_ID&value=7163
    kit_id_raw = _re_first(r"Kit_ID&value=(-?\d+)", html)
    kit_id_int = int(kit_id_raw) if kit_id_raw else None
    kit_url = f"{BASE_URL}/product-{kit_id_int}.html" if kit_id_int and kit_id_int > 0 else None

    # Manufacturer canonical URL: e.g. /estes-1093.html
    mfr_canonical = _re_first(r'href="(/[a-z][a-z0-9-]+-\d+\.html)"', html)
    manufacturer_url = f"{BASE_URL}{mfr_canonical}" if mfr_canonical else None

    # Manufacturer alias URL from the kit-index filter link
    # e.g. /kit-index-manufacturer-estes.html
    mfr_alias = _re_first(r"(/kit-index-manufacturer-[^\"']+\.html)", html)
    manufacturer_url_alias = f"{BASE_URL}{mfr_alias}" if mfr_alias else None

    # Manufacturer name from the kit-index link (more reliable than the index field)
    manufacturer = index_rec.get("manufacturer") or None
    mfr_link = soup.find("a", href=re.compile(r"/kit-index-manufacturer-"))
    if mfr_link:
        manufacturer = mfr_link.get_text(strip=True) or manufacturer

    # Contributor URL from profile link with 4+ digit numeric suffix
    # e.g. /darrell-ritchies-darrell-8979.html
    contributor_path = _re_first(r"(/[a-z][a-z0-9-]+-\d{4,}\.html)", html)
    contributor_url = f"{BASE_URL}{contributor_path}" if contributor_path else None

    return {
        "id": int(index_rec["id"]),
        "url": canonical_url,
        "url_alias": url_alias,
        "date": index_rec.get("date"),
        "contributor": index_rec.get("contributor"),
        "contributor_url": contributor_url,
        "kit": index_rec.get("kit"),
        "kit_url": kit_url,
        "manufacturer": manufacturer,
        "manufacturer_url": manufacturer_url,
        "manufacturer_url_alias": manufacturer_url_alias,
        "type": index_rec.get("type"),
        "scraped_at": datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
        "ratings": _parse_ratings(soup),
        "product": _parse_product(soup),
        "sections": _parse_sections(soup),
    }


# ---------------------------------------------------------------------------
# Fetch helpers
# ---------------------------------------------------------------------------


def fetch_index(session: requests.Session) -> list[dict]:
    log.info("Fetching review index from %s", INDEX_URL)
    resp = session.get(INDEX_URL, timeout=30)
    resp.raise_for_status()
    records = resp.json().get("records", [])
    log.info("Index returned %d records.", len(records))
    return records


def scrape_detail(
    session: requests.Session,
    rate: RateLimiter,
    index_rec: dict,
    force: bool = False,
) -> Optional[dict]:
    review_id = int(index_rec["id"])
    shard = f"{review_id // 1000:03d}"
    shard_dir = DETAIL_DIR / shard
    dest = shard_dir / f"{review_id:06d}.json"

    if dest.exists() and not force:
        log.debug("Already scraped %s, skipping.", review_id)
        return None

    alias_path = index_rec.get("url", "")
    if not alias_path:
        log.warning("No URL for review %s, skipping.", review_id)
        return None

    url = f"{BASE_URL}{alias_path}"
    rate.wait()

    try:
        resp = session.get(url, timeout=30)
        resp.raise_for_status()
    except requests.RequestException as exc:
        log.warning("Failed to fetch review %s: %s", review_id, exc)
        return None

    return _parse_detail(resp.text, index_rec)


# ---------------------------------------------------------------------------
# Main
# ---------------------------------------------------------------------------


def main() -> None:
    parser = argparse.ArgumentParser(description="Scrape RocketReviews.com reviews.")
    parser.add_argument(
        "--delay",
        type=float,
        default=DEFAULT_DELAY,
        help=f"Seconds between requests (default: {DEFAULT_DELAY})",
    )
    parser.add_argument(
        "--limit",
        type=int,
        default=None,
        help="Stop after scraping this many detail pages (useful for testing)",
    )
    parser.add_argument(
        "--force",
        action="store_true",
        help="Re-scrape reviews that already have a saved detail file",
    )
    args = parser.parse_args()

    SOURCE_DIR.mkdir(parents=True, exist_ok=True)
    DETAIL_DIR.mkdir(parents=True, exist_ok=True)

    session = _build_session()
    rate = RateLimiter(args.delay)

    # ------------------------------------------------------------------
    # Step 1: fetch and write the full index
    # ------------------------------------------------------------------
    records = fetch_index(session)

    scraped_at = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
    index_path = SOURCE_DIR / "index.jsonl"
    with index_path.open("w", encoding="utf-8") as f:
        for rec in records:
            f.write(json.dumps({**rec, "scraped_at": scraped_at}) + "\n")
    log.info("Wrote %d index records to %s", len(records), index_path)

    # ------------------------------------------------------------------
    # Step 2: scrape each detail page
    # ------------------------------------------------------------------
    if args.limit:
        records = records[: args.limit]

    ok = skipped = failed = 0
    total = len(records)

    for i, rec in enumerate(records, 1):
        result = scrape_detail(session, rate, rec, force=args.force)

        if result is None:
            # Already exists and --force not set, or missing URL
            skipped += 1
            continue

        review_id = int(rec["id"])
        shard = f"{review_id // 1000:03d}"
        shard_dir = DETAIL_DIR / shard
        shard_dir.mkdir(parents=True, exist_ok=True)
        dest = shard_dir / f"{review_id:06d}.json"
        
        try:
            dest.write_text(
                json.dumps(result, indent=2, ensure_ascii=False), encoding="utf-8"
            )
            ok += 1
            log.debug("Saved %s", dest.name)
        except OSError as exc:
            log.warning("Could not write %s: %s", dest, exc)
            failed += 1

        if i % 25 == 0 or i == total:
            log.info(
                "Progress: %d/%d — ok=%d  skipped=%d  failed=%d",
                i,
                total,
                ok,
                skipped,
                failed,
            )

    log.info("Done — ok=%d  skipped=%d  failed=%d", ok, skipped, failed)


if __name__ == "__main__":
    main()