Facility: 001967

B & B Mini Storage

Stale Data Warning: This facility has not been successfully scraped in 26 days (threshold: 3 days). Data may be outdated.
Facility Information active
Facility ID
001967
Name
B & B Mini Storage
URL
http://bnbministorage.com/
Address
1051 38th St, Peru, IL 61354, USA, Peru, Illinois 61354
Platform
custom_facility_001967
Parser File
src/parsers/custom/facility_001967_parser.py
Last Scraped
2026-03-27 13:49:17.449809
Created
2026-03-14 16:21:53.706708
Updated
2026-03-27 13:49:17.478847
Parser & Healing Diagnosis working
Parser Status
✓ Working
Status Reason
N/A
Last Healing Attempt
Not attempted
Parser Source (src/parsers/custom/facility_001967_parser.py) 
"""Parser for B & B Mini Storage."""

from __future__ import annotations

import re

from bs4 import BeautifulSoup

from src.parsers.base import BaseParser, ParseResult, UnitResult


class Facility001967Parser(BaseParser):
    """Extract storage units from B & B Mini Storage."""

    platform = "custom_facility_001967"

    _UNIT_RE = re.compile(
        r"(\d+\s*[\'\'\u2032]?\s*[xX\u00d7]\s*\d+\s*[\'\'\u2032]?)"
        r"[^\$]{0,120}"
        r"\$(\d[\d,.]*)",
        re.DOTALL,
    )

    _PRICE_SIZE_RE = re.compile(
        r"\$(\d[\d,.]*)"
        r".{0,120}"
        r"(\d+\s*[\'\'\u2032]?\s*[xX\u00d7]\s*\d+\s*[\'\'\u2032]?)",
        re.DOTALL,
    )

    _SIZE_ONLY_RE = re.compile(
        r"(\d+\s*[\'\'\u2032]?\s*[xX\u00d7]\s*\d+\s*[\'\'\u2032]?)"
    )

    def parse(self, html: str, url: str = "") -> ParseResult:
        soup = BeautifulSoup(html, "lxml")
        result = ParseResult(platform=self.platform, parser_name=self.__class__.__name__)

        for tag in soup.find_all(["script", "style"]):
            tag.decompose()

        body_text = soup.get_text(separator="\n")

        seen: set[tuple[str, str]] = set()

        # Try size-then-price pattern
        for m in self._UNIT_RE.finditer(body_text):
            size_text = m.group(1).strip()
            price_text = m.group(2).strip()
            key = (size_text, price_text)
            if key in seen:
                continue
            seen.add(key)

            unit = UnitResult()
            unit.size = size_text
            w, ln, sq = self.normalize_size(size_text)
            if w is not None:
                unit.metadata = {"width": w, "length": ln, "sqft": sq}
            unit.price = self.normalize_price(price_text)
            unit.description = m.group(0).strip()[:200]
            if unit.size or unit.price:
                result.units.append(unit)

        # Try price-then-size pattern if no results
        if not result.units:
            for m in self._PRICE_SIZE_RE.finditer(body_text):
                price_text = m.group(1).strip()
                size_text = m.group(2).strip()
                key = (size_text, price_text)
                if key in seen:
                    continue
                seen.add(key)

                unit = UnitResult()
                unit.size = size_text
                w, ln, sq = self.normalize_size(size_text)
                if w is not None:
                    unit.metadata = {"width": w, "length": ln, "sqft": sq}
                unit.price = self.normalize_price(price_text)
                unit.description = m.group(0).strip()[:200]
                if unit.size or unit.price:
                    result.units.append(unit)

        # Fallback: extract sizes without prices
        if not result.units:
            seen_sizes: set[str] = set()
            for m in self._SIZE_ONLY_RE.finditer(body_text):
                size_text = m.group(1).strip()
                if size_text in seen_sizes:
                    continue
                w, ln, sq = self.normalize_size(size_text)
                if w is None or w < 3 or ln < 3:
                    continue
                seen_sizes.add(size_text)
                unit = UnitResult()
                unit.size = size_text
                unit.metadata = {"width": w, "length": ln, "sqft": sq}
                result.units.append(unit)

        if not result.units:
            result.warnings.append("No units found via regex")

        return result

Scrape Runs (7)

Run #1750 Details

Status
exported
Parser Used
Facility001967Parser
Platform Detected
table_layout
Units Found
7
Stage Reached
exported
Timestamp
2026-03-27 13:48:56.675595
Timing
Stage Duration
Fetch20606ms
Detect25ms
Parse11ms
Export17ms

Snapshot: 001967_20260327T134917Z.html · Show Snapshot · Open in New Tab

Parsed Units (7)

5x10

No price

10x10

No price

10x15

No price

10×15

No price

10x20

No price

10x25

No price

10x30

No price

All Failures for this Facility (2)
fetch TimeoutException website timeout transient Run #644 | 2026-03-21 18:41:36.748676

Message: timeout: Timed out receiving message from renderer: 1.528 (Session info: chrome=146.0.7680.153) Stacktrace: #0 0x55fba75ff8ce <unknown> #1 0x55fba6fbd1d2 <unknown> #2 0x55fba6fa81fc <unknown> #3 0x55fba6fa7fe9 <unknown> #4 0x55fba6fa65e6 <unknown> #5 0x55fba6fa6a96 <unknown> #6 0x55fba6fb50f7 <unknown> #7 0x55fba6fca77d <unknown> #8 0x55fba6fd002b <unknown> #9 0x55fba6fa70a1 <unknown> #10 0x55fba6fca5bf <unknown> #11 0x55fba704cb79 <unknown> #12 0x55fba702ca03 <unknown> #13 0x55fba6ffd5d5 <unknown> #14 0x55fba6ffe1c1 <unknown> #15 0x55fba75c3f10 <unknown> #16 0x55fba75c71d8 <unknown> #17 0x55fba75c6c8a <unknown> #18 0x55fba75c7645 <unknown> #19 0x55fba75b38fb <unknown> #20 0x55fba75c79a7 <unknown> #21 0x55fba759b836 <unknown> #22 0x55fba75ec8a5 <unknown> #23 0x55fba75eca9c <unknown> #24 0x55fba75fe38a <unknown> #25 0x7f34a7eb71f5 <unknown>

Stack trace 
Traceback (most recent call last):
  File "/app/src/pipeline.py", line 361, in _process_facility
    fetch_result = fetch_page(driver, url, snapshot_mgr, facility_id, **fetch_kwargs)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/src/fetcher/fetcher.py", line 125, in fetch_page
    driver.get(url)
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/webdriver.py", line 466, in get
    self.execute(Command.GET, {"url": url})
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/webdriver.py", line 446, in execute
    self.error_handler.check_response(response)
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/errorhandler.py", line 232, in check_response
    raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.TimeoutException: Message: timeout: Timed out receiving message from renderer: 1.528
  (Session info: chrome=146.0.7680.153)
Stacktrace:
#0 0x55fba75ff8ce <unknown>
#1 0x55fba6fbd1d2 <unknown>
#2 0x55fba6fa81fc <unknown>
#3 0x55fba6fa7fe9 <unknown>
#4 0x55fba6fa65e6 <unknown>
#5 0x55fba6fa6a96 <unknown>
#6 0x55fba6fb50f7 <unknown>
#7 0x55fba6fca77d <unknown>
#8 0x55fba6fd002b <unknown>
#9 0x55fba6fa70a1 <unknown>
#10 0x55fba6fca5bf <unknown>
#11 0x55fba704cb79 <unknown>
#12 0x55fba702ca03 <unknown>
#13 0x55fba6ffd5d5 <unknown>
#14 0x55fba6ffe1c1 <unknown>
#15 0x55fba75c3f10 <unknown>
#16 0x55fba75c71d8 <unknown>
#17 0x55fba75c6c8a <unknown>
#18 0x55fba75c7645 <unknown>
#19 0x55fba75b38fb <unknown>
#20 0x55fba75c79a7 <unknown>
#21 0x55fba759b836 <unknown>
#22 0x55fba75ec8a5 <unknown>
#23 0x55fba75eca9c <unknown>
#24 0x55fba75fe38a <unknown>
#25 0x7f34a7eb71f5 <unknown>
fetch TimeoutException website timeout transient Run #643 | 2026-03-21 18:41:01.698651

Message: timeout: Timed out receiving message from renderer: -0.003 (Session info: chrome=146.0.7680.153) Stacktrace: #0 0x55fba75ff8ce <unknown> #1 0x55fba6fbd1d2 <unknown> #2 0x55fba6fa81fc <unknown> #3 0x55fba6fa7fe9 <unknown> #4 0x55fba6fa65e6 <unknown> #5 0x55fba6fa6a96 <unknown> #6 0x55fba6fb50f7 <unknown> #7 0x55fba6fca77d <unknown> #8 0x55fba6fd002b <unknown> #9 0x55fba6fa70a1 <unknown> #10 0x55fba6fca5bf <unknown> #11 0x55fba704c86a <unknown> #12 0x55fba702ca03 <unknown> #13 0x55fba6ffd5d5 <unknown> #14 0x55fba6ffe1c1 <unknown> #15 0x55fba75c3f10 <unknown> #16 0x55fba75c71d8 <unknown> #17 0x55fba75c6c8a <unknown> #18 0x55fba75c7645 <unknown> #19 0x55fba75b38fb <unknown> #20 0x55fba75c79a7 <unknown> #21 0x55fba759b836 <unknown> #22 0x55fba75ec8a5 <unknown> #23 0x55fba75eca9c <unknown> #24 0x55fba75fe38a <unknown> #25 0x7f34a7eb71f5 <unknown>

Stack trace 
Traceback (most recent call last):
  File "/app/src/pipeline.py", line 361, in _process_facility
    fetch_result = fetch_page(driver, url, snapshot_mgr, facility_id, **fetch_kwargs)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/src/fetcher/fetcher.py", line 125, in fetch_page
    driver.get(url)
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/webdriver.py", line 466, in get
    self.execute(Command.GET, {"url": url})
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/webdriver.py", line 446, in execute
    self.error_handler.check_response(response)
  File "/app/.venv/lib/python3.11/site-packages/selenium/webdriver/remote/errorhandler.py", line 232, in check_response
    raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.TimeoutException: Message: timeout: Timed out receiving message from renderer: -0.003
  (Session info: chrome=146.0.7680.153)
Stacktrace:
#0 0x55fba75ff8ce <unknown>
#1 0x55fba6fbd1d2 <unknown>
#2 0x55fba6fa81fc <unknown>
#3 0x55fba6fa7fe9 <unknown>
#4 0x55fba6fa65e6 <unknown>
#5 0x55fba6fa6a96 <unknown>
#6 0x55fba6fb50f7 <unknown>
#7 0x55fba6fca77d <unknown>
#8 0x55fba6fd002b <unknown>
#9 0x55fba6fa70a1 <unknown>
#10 0x55fba6fca5bf <unknown>
#11 0x55fba704c86a <unknown>
#12 0x55fba702ca03 <unknown>
#13 0x55fba6ffd5d5 <unknown>
#14 0x55fba6ffe1c1 <unknown>
#15 0x55fba75c3f10 <unknown>
#16 0x55fba75c71d8 <unknown>
#17 0x55fba75c6c8a <unknown>
#18 0x55fba75c7645 <unknown>
#19 0x55fba75b38fb <unknown>
#20 0x55fba75c79a7 <unknown>
#21 0x55fba759b836 <unknown>
#22 0x55fba75ec8a5 <unknown>
#23 0x55fba75eca9c <unknown>
#24 0x55fba75fe38a <unknown>
#25 0x7f34a7eb71f5 <unknown>

← Back to dashboard