we-promise/sure · warning

Invalid institution URL for <%= class_name %> account #{prov

Error message

Invalid institution URL for <%= class_name %> account #{provider_account.id}: #{url}

What it means

In the generated provider adapter, institution_domain reads institution_metadata and, when the "domain" key is blank but "url" is present, tries URI.parse(url).host to derive the domain (stripping a leading www.). URI::InvalidURIError is rescued with this warning, leaving domain nil, which only degrades institution logo/website lookups. This is a template file — the warning text literally contains <%= class_name %> until the generator renders it.

Source

Thrown at lib/generators/provider/family/templates/adapter.rb.tt:100

<% if investment_provider? -%>
  def can_delete_holdings?
    false
  end
<% end -%>

  def institution_domain
    metadata = provider_account.institution_metadata
    return nil unless metadata.present?

    domain = metadata["domain"]
    url = metadata["url"]

    # Derive domain from URL if missing
    if domain.blank? && url.present?
      begin
        domain = URI.parse(url).host&.gsub(/^www\./, "")
      rescue URI::InvalidURIError
        Rails.logger.warn("Invalid institution URL for <%= class_name %> account #{provider_account.id}: #{url}")
      end
    end

    domain
  end

  def institution_name
    metadata = provider_account.institution_metadata
    return nil unless metadata.present?

    metadata["name"] || item&.institution_name
  end

  def institution_url
    metadata = provider_account.institution_metadata
    return nil unless metadata.present?

    metadata["url"] || item&.institution_url

View on GitHub (pinned to e69894adb9)

Solutions

1. Inspect provider_account.institution_metadata to see the exact url value that failed to parse
2. Prefer populating the "domain" key in institution_metadata directly (in the generated extract_institution_metadata) so URL parsing is never needed
3. Harden the generated adapter: rescue and retry with a cleaned string, or fall back to a regex host extraction
4. Accept nil domain when the institution has no usable website

Example fix

# generated adapter - before
begin
  domain = URI.parse(url).host&.gsub(/^www\./, "")
rescue URI::InvalidURIError
  Rails.logger.warn("Invalid institution URL for #{provider_account.id}: #{url}")
end

# after
begin
  domain = URI.parse(url.to_s.strip).host&.gsub(/^www\./, "")
  domain ||= url.to_s[/\b(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z]{2,}\b/i]
rescue URI::InvalidURIError
  domain = url.to_s[/\b(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z]{2,}\b/i]
  Rails.logger.warn("Invalid institution URL for #{provider_account.id}: #{url}") if domain.nil?
end

# Cheap pre-check before URI.parse in the generated adapter
def parseable_url?(url)
  url.is_a?(String) && url.strip.match?(%r{\Ahttps?://\S+\z})
end

domain = parseable_url?(url) ? URI.parse(url.strip).host&.gsub(/^www\./, "") : nil

def plausible_url?(value)
  value.is_a?(String) && value.strip.match?(%r{\Ahttps?://[\w.-]+\.[a-z]{2,}(?:/\S*)?\z}i)
end

begin
  domain = URI.parse(url.to_s.strip).host&.gsub(/^www\./, "")
rescue URI::InvalidURIError => e
  Rails.logger.warn("Invalid institution URL for #{provider_account.id}: #{url}")
  domain = nil # explicit degradation, logo lookup simply skipped
end

When it happens

Trigger: institution_metadata["url"] is a malformed URI that raises URI::InvalidURIError: strings with spaces ("http://my bank.com"), missing scheme with bad characters ("mybank com"), placeholder text ("Not Available"), or control characters. Note: URI.parse("mybank.com") does not raise (host is nil), so the warning fires only for genuinely unparseable strings.

Common situations: Aggregator institution metadata containing display text instead of URLs; manually curated/free-text institution fields; providers that fill the url slot with a null-like sentinel ("N/A", "-"); encoding issues in copied URLs.

Related errors

• Invalid currency code '#{currency_value}' for <%= class_name
• <%= class_name %>Account::ActivitiesProcessor - Unmapped act
• <%= class_name %>ConnectionCleanupJob - Failed: #{e.class} -
• Invalid currency code '#{currency_value}' for <%= class_name
• Unexpected response format from search endpoint

AI-assisted analysis of we-promise/sure@e69894adb9 (2026-08-21). Data as JSON: /api/errors/f16399187249ffe3. Report an issue: GitHub.