Module: Html2rss::Doctor::Botasaurus

Defined in:
lib/html2rss/doctor/botasaurus.rb

Overview

Preflight checks for Botasaurus scrape-api connectivity.

Defined Under Namespace

Classes: Check, Result

Class Method Summary collapse

Class Method Details

.call(sample_url: nil) ⇒ Result

Parameters:

  • (defaults to: nil)

    optional URL for a sample scrape smoke test

Returns:



41
42
43
44
45
46
47
48
49
50
51
52
53
54
 
# File 'lib/html2rss/doctor/botasaurus.rb', line 41

def call(sample_url: nil) # rubocop:disable Metrics/AbcSize
  env = env_check
  checks = [env_check_record(env)]
  return missing_env_failure(checks) unless checks.first.ok

  checks << health_check_record(env[:base_url])
  return failure(checks, 'Botasaurus health check failed.') unless checks.last.ok

  checks << sample_scrape_record(sample_url) if sample_url
  return failure(checks, 'Sample scrape failed.') if sample_url && !checks.last.ok

  Log.info('doctor botasaurus: preflight passed')
  Result.new(ok: true, checks:, message: 'Botasaurus preflight passed.')
end

.env_check ⇒ Hash{Symbol => Object}

Returns:



58
59
60
61
62
63
64
65
 
# File 'lib/html2rss/doctor/botasaurus.rb', line 58

def env_check
  base = ENV.fetch(ENV_VAR, '').strip
  {
    configured: MCP::Runtime.botasaurus_configured?,
    var: ENV_VAR,
    base_url: base.empty? ? nil : base
  }
end

.health_check(base_url) ⇒ Hash{Symbol => Object}

Parameters:

Returns:



70
71
72
73
74
75
76
77
78
79
 
# File 'lib/html2rss/doctor/botasaurus.rb', line 70

def health_check(base_url)
  endpoint = "#{Url.for_channel(base_url).to_s.chomp('/')}/health"
  response = health_session.get(endpoint)
  raise response.error if response.is_a?(HTTPX::ErrorResponse)

  body = JSON.parse(response.body.to_s)
  { ok: response.status == 200, status: response.status, version: body['version'] }
rescue StandardError => error
  { ok: false, error: "#{error.class}: #{error.message}" }
end

.sample_scrape(url) ⇒ Hash{Symbol => Object}

Parameters:

Returns:



91
92
93
94
95
96
97
98
99
100
 
# File 'lib/html2rss/doctor/botasaurus.rb', line 91

def sample_scrape(url)
  wire = PageRecon::Diagnostics.call(url:, strategy: :botasaurus).to_wire_h
  {
    ok: wire[:status].to_i.between?(200, 399),
    status: wire[:status],
    articles_count: wire[:articles_count]
  }
rescue StandardError => error
  { ok: false, error: "#{error.class}: #{error.message}" }
end