tas_councils/scrapers/waratah_wynyard.rb

# Waratah–Wynyard Council — Advertised / Planning Applications

require "nokogiri"
require "uri"
require "cgi"

require_relative "../lib/enrich"
require_relative "../lib/log"
require_relative "../lib/util"
TABLE = ENV.fetch("TABLE_NAME")  # da_waratah_wynyard
URL   = "https://www.warwyn.tas.gov.au/planning-and-development/advertised-permits/"

DB.ensure_table!(TABLE)

def abs_url(base, href)
  return "" if href.to_s.strip.empty?
  URI.join(base, href).to_s
rescue URI::InvalidURIError
  href.to_s
end

# DA 2025/0123, DA2025-0123, DA 114-2025 etc.
REF_RX1 = %r{\bDA\s*(20\d{2})\s*/\s*([A-Za-z0-9\-\._]+)}i
REF_RX2 = %r{\bDA(20\d{2})\s*[-\/]?\s*([0-9]{3,})\b}i
REF_RX3 = %r{\bDA\s*([0-9]{1,4})\s*-\s*(20\d{2})\b}i

def extract_ref(str)
  s = CGI.unescape(str.to_s)
  if (m = s.match(REF_RX1)) then return "DA #{m[1]} / #{m[2]}" end
  if (m = s.match(REF_RX2)) then return "DA #{m[1]} / #{m[2]}" end
  if (m = s.match(REF_RX3)) then return "DA #{m[2]} / #{m[1]}" end
  nil
end

def extract_date_like(str)
  s = str.to_s
  return $1 if s =~ /(\b\d{1,2}\/\d{1,2}\/\d{2,4}\b)/
  return $1 if s =~ /(\b\d{1,2}\s+[A-Za-z]{3,}\s+\d{4}\b)/
  return $1 if s =~ /(\b[A-Za-z]{3,}\s+\d{1,2},?\s+\d{4}\b)/
  ""
end

def extract_on_notice_raw(text)
  s = text.to_s.gsub(/\s+/, " ")
  if s =~ /\bon\s*notice\s*(until|to)\s*[:\-]?\s*([A-Za-z0-9\/ ,]+)/i
    d = extract_date_like($2); return d unless d.empty?
  end
  if s =~ /clos(?:e|ing|es)\s*(on)?\s*[:\-]?\s*([A-Za-z0-9\/ ,]+)/i
    d = extract_date_like($2); return d unless d.empty?
  end
  if s =~ /submissions?\s*close\s*[:\-]?\s*([A-Za-z0-9\/ ,]+)/i
    d = extract_date_like($1); return d unless d.empty?
  end
  extract_date_like(s)
end

def nearest_context_text(a)
  host = a.ancestors("li, p, div, tr, article").first || a.parent
  host ? host.text.to_s.strip.gsub(/\s+/, " ") : ""
end

def parse_list_items(doc, base_url)
  rows = []
  anchors = doc.css("a").select { |a|
    href = a["href"].to_s
    a.text.to_s.strip.match?(/application|permit|planning|advertis/i) || href.downcase.end_with?(".pdf")
  }

  anchors.each do |a|
    href = a["href"].to_s
    link_text = a.text.to_s.strip
    document_url = abs_url(base_url, href)
    ctx = nearest_context_text(a)

    title_reference = link_text.empty? ? ctx[0,200] : link_text
    text_for_parse  = [link_text, ctx].reject(&:empty?).uniq.join(" — ")

    # Address guess
    address = link_text.length >= 6 ? link_text : ctx[0, 140]

    ref = extract_ref(text_for_parse) || extract_ref(File.basename(document_url))

    on_raw = extract_on_notice_raw(text_for_parse)
    on_dt  = Util.parse_aus_date(on_raw)

    description = if text_for_parse =~ /proposal\s*[:\-]\s*([^—\-]+)\b/i
      $1.strip
    else
      "Development Application"
    end

    next if ref.nil? || address.to_s.strip.empty?

    rows << {
      council_reference: ref,
      address: address.to_s.strip,
      description: description,
      date_received: on_dt,
      date_received_raw: on_raw,
      document_url: document_url,
      title_reference: title_reference
    }
  end

  rows.uniq { |r| [r[:council_reference], r[:address]] }
end

def parse_detail_page(url)
  html = Http.get(url)
  doc  = Nokogiri::HTML(html)

  # Try simple two-column tables first
  kv = {}
  doc.css("table tr").each do |tr|
    cells = tr.css("th, td")
    next unless cells.length >= 2
    key = cells[0].text.strip
    val = cells[1].text.strip
    kv[key] = val unless key.empty?
  end

  if kv.any?
    find = ->(rx) { kv.find { |k, _| k =~ rx }&.last.to_s.strip }

    council_reference = find.call(/(Application\s*(No|Number|ID)|Reference)/i)
    address           = find.call(/(Address|Location|Property)/i)
    description       = find.call(/(Proposal|Description)/i)
    on_notice_raw     = find.call(/(On\s*Notice\s*(until|to)|Closing\s*Date|Closes|Submissions)/i)

    pdf = doc.at_css("a[href$='.pdf'], a[href*='.pdf?']")&.[]("href")
    document_url = pdf ? abs_url(url, pdf) : ""

    unless council_reference.empty? || address.empty?
      return {
        council_reference: council_reference,
        address: address,
        description: description.empty? ? "Development Application" : description,
        date_received_raw: on_notice_raw,
        date_received: Util.parse_aus_date(on_notice_raw),
        document_url: document_url,
        title_reference: doc.at_css("h1, .entry-title")&.text&.strip.to_s
      }
    end
  end

  # Fallback: parse from page text
  page_text = doc.text.to_s.strip.gsub(/\s+/, " ")
  ref = extract_ref(page_text)
  on_raw = extract_on_notice_raw(page_text)
  on_dt  = Util.parse_aus_date(on_raw)
  h1 = doc.at_css("h1, .entry-title")&.text&.strip.to_s
  address = h1.empty? ? page_text[0, 140] : h1
  pdf = doc.at_css("a[href$='.pdf'], a[href*='.pdf?']")&.[]("href")
  document_url = pdf ? abs_url(url, pdf) : ""

  return nil if ref.nil? || address.empty?

  {
    council_reference: ref,
    address: address,
    description: "Development Application",
    date_received_raw: on_raw,
    date_received: on_dt,
    document_url: document_url,
    title_reference: h1
  }
end

begin
  html = URL.include?("/eservice/") ? Http.dorset_session_get(URL) : Http.get(URL)
rescue StandardError => e
  Log.warn "scraper", "Failed to fetch #{URL}: #{e.class} #{e.message}"
  exit 1
end

doc = Nokogiri::HTML(html)

host = begin
  URI.parse(URL).host
rescue URI::InvalidURIError
  nil
end

anchors = doc.css("a").map { |a| abs_url(URL, a["href"].to_s) }.select { |u|
  next false if u.empty? || u.start_with?("#")
  u.downcase.end_with?(".pdf") || begin
    uh = URI.parse(u).host rescue nil
    host && uh == host
  end
}.uniq

rows = []

anchors.each do |u|
  if u.downcase.end_with?(".pdf")
    if (a = doc.at_css(%Q{a[href="#{u}"]}))
      ctx_text = nearest_context_text(a)
      title    = a.text.to_s.strip
      ref      = extract_ref([title, ctx_text].join(" — "))
      addr     = title.length >= 6 ? title : ctx_text[0, 140]
      on_raw   = extract_on_notice_raw([title, ctx_text].join(" — "))
      on_dt    = Util.parse_aus_date(on_raw)
      next if ref.nil? || addr.to_s.strip.empty?
      rows << {
        council_reference: ref,
        address: addr,
        description: "Development Application",
        date_received: on_dt,
        date_received_raw: on_raw,
        document_url: u,
        title_reference: title.empty? ? ctx_text[0,200] : title
      }
    end
  else
    begin
      item = parse_detail_page(u)
      rows << item if item
    rescue StandardError => e
      Log.warn "scraper", "Skip detail #{u}: #{e.class} #{e.message}"
    end
  end
end

# Safety net: scrape items from the main page content too
rows += parse_list_items(doc, URL)
rows.uniq! { |r| [r[:council_reference], r[:address]] }

puts "Found #{rows.length} item(s) for #{TABLE}"

rows.each do |r|
  cr   = r[:council_reference].to_s
  addr = r[:address].to_s

  next if addr.strip.empty?
  next if addr =~ /\A(?:download|advertised planning applications)\z/i
  next if cr.strip.empty?
  next if addr == cr

  DB.upsert(TABLE, {
    description: r[:description],
    date_received: r[:date_received],
    date_received_raw: r[:date_received_raw],
    address: addr,
    council_reference: cr,
    applicant: "",
    owner: ""
  })
  
  enrich_after_upsert!(
    table: TABLE,
    council_reference: cr,
    address: addr
  )

  begin
    upd = DB.client.prepare(
      "UPDATE `#{DB.client.escape(TABLE)}` " \
      "SET document_url = ?, on_notice_to = ?, on_notice_to_raw = ?, title_reference = ? " \
      "WHERE council_reference = ? AND address = ?"
    )
    upd.execute(r[:document_url], r[:date_received], r[:date_received_raw], r[:title_reference], cr, addr)
  rescue StandardError => e
    Log.warn "scraper", "Extras update skipped for #{cr}: #{e.class} #{e.message}"
  end

  puts "Upserted #{cr} -> #{addr}"
end

puts "Done #{TABLE}."