Module: Relaton::W3c::RateLimitHandler

Included in:
DataFetcher, DataParser
Defined in:
lib/relaton/w3c/rate_limit_handler.rb

Constant Summary collapse

MAX_RETRIES = 
5
RETRYABLE_ERRORS = 
[
  NameError, Lutaml::Hal::ConnectionError, Lutaml::Hal::TimeoutError,
  Lutaml::Hal::ServerError, Faraday::ConnectionFailed, Net::OpenTimeout,
].freeze

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.fetched_objects ⇒ Object

Concurrent::Map so multiple fetcher threads can hit the cache without a global lock. Duplicate concurrent fetches of the same URL are possible but harmless; the second write just replaces the first.



15
16
17
 
# File 'lib/relaton/w3c/rate_limit_handler.rb', line 15

def self.fetched_objects
  @fetched_objects ||= Concurrent::Map.new
end

Instance Method Details

#realize(obj) ⇒ Object 



19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
 
# File 'lib/relaton/w3c/rate_limit_handler.rb', line 19

def realize(obj)
  href = resolve_href(obj)
  return RateLimitHandler.fetched_objects[href] if RateLimitHandler.fetched_objects.key?(href)

  attempt = 1
  begin
    RateLimitHandler.fetched_objects[href] = obj.realize
  rescue *RETRYABLE_ERRORS => e
    if attempt < MAX_RETRIES
      sleep_time = attempt * attempt
      attempt += 1
      Util.warn "Rate limit exceeded for #{href}, retrying in #{sleep_time} seconds..."
      sleep sleep_time
      retry
    elsif e.is_a?(Lutaml::Hal::ServerError)
      # Persistent 5xx — cache nil so a permanently broken upstream
      # resource is skipped on the next lookup instead of re-tried.
      Util.warn "Server error for #{href}, skipping: #{e.message}"
      RateLimitHandler.fetched_objects[href] = nil
    else
      # Do not cache on retries exhausted — transient failures should not
      # permanently poison the cache; subsequent calls will retry fresh.
      Util.warn "Failed to realize object: #{href}, error: #{e.message}"
    end
  rescue Lutaml::Hal::NotFoundError
    Util.warn "Object not found: #{href}"
    RateLimitHandler.fetched_objects[href] = nil
  rescue Lutaml::Hal::Error => e
    # Non-retryable client-side errors (403/401/400 and any other
    # Lutaml::Hal::Error not matched above) — skip the resource and
    # continue rather than aborting the whole crawl.
    Util.warn "Client error for #{href}, skipping: #{e.message}"
    RateLimitHandler.fetched_objects[href] = nil
  end
end