fix: Added authentication to FireCrawl API, remove unused RobinAI references (#10737)

- Fixed Firecrawl webhook payloads to ensure proper data handling and
delivery.
- Removed unused Robin AI code to improve codebase cleanliness and
maintainability.
- Implement authentication for the Firecrawl endpoint to improve
security. A key is generated to secure the webhook URLs from FireCrawl.

---------

Co-authored-by: Pranav <pranavrajs@gmail.com>
This commit is contained in:
Sojan Jose
2025-01-23 07:44:25 +05:30
committed by GitHub
parent 3b366f43e6
commit be8205657e
28 changed files with 345 additions and 470 deletions

View File

@@ -1,17 +0,0 @@
class ResponseSourcePolicy < ApplicationPolicy
def parse?
@account_user.administrator?
end
def create?
@account_user.administrator?
end
def add_document?
@account_user.administrator?
end
def remove_document?
@account_user.administrator?
end
end

View File

@@ -1,3 +0,0 @@
json.array! @response_sources do |response_source|
json.partial! 'api/v1/models/response_source', formats: [:json], resource: response_source
end

View File

@@ -1 +0,0 @@
json.partial! 'api/v1/models/response_source', formats: [:json], resource: @response_source

View File

@@ -1 +0,0 @@
json.partial! 'api/v1/models/response_source', formats: [:json], resource: @response_source

View File

@@ -1 +0,0 @@
json.partial! 'api/v1/models/response_source', formats: [:json], resource: @response_source

View File

@@ -1,15 +0,0 @@
json.id resource.id
json.name resource.name
json.source_link resource.source_link
json.source_type resource.source_type
json.account_id resource.account_id
json.created_at resource.created_at.to_i
json.updated_at resource.updated_at.to_i
json.response_documents do
json.array! resource.response_documents do |response_document|
json.id response_document.id
json.document_link response_document.document_link
json.created_at response_document.created_at.to_i
json.updated_at response_document.updated_at.to_i
end
end

View File

@@ -32,7 +32,7 @@ as defined by the routes in the `admin/` namespace
<ul class="my-4">
<%= render partial: "nav_item", locals: { icon: 'icon-grid-line', url: super_admin_root_url, label: 'Dashboard' } %>
<% Administrate::Namespace.new(namespace).resources.each do |resource| %>
<% next if ["account_users", "access_tokens", "installation_configs", "dashboard", "devise/sessions", "app_configs", "instance_statuses", "responses", "response_sources", "response_documents" , "settings"].include? resource.resource %>
<% next if ["account_users", "access_tokens", "installation_configs", "dashboard", "devise/sessions", "app_configs", "instance_statuses", "settings"].include? resource.resource %>
<%= render partial: "nav_item", locals: {
icon: sidebar_icons[resource.resource.to_sym],
url: resource_index_route(resource),

View File

@@ -142,6 +142,10 @@
display_title: 'OpenAI Model'
description: 'The OpenAI model configured for use in Captain AI. Default: gpt-4o-mini'
locked: false
- name: CAPTAIN_FIRECRAWL_API_KEY
display_title: 'FireCrawl API Key (optional)'
description: 'The FireCrawl API key for the Captain AI service'
locked: false
- name: CAPTAIN_CLOUD_PLAN_LIMITS
display_title: 'Captain Cloud Plan Limits'
description: 'The limits for the Captain AI service for different plans'

View File

@@ -481,8 +481,6 @@ Rails.application.routes.draw do
end
resources :access_tokens, only: [:index, :show]
resources :response_documents, only: [:index, :show, :new, :create, :edit, :update, :destroy]
resources :responses, only: [:index, :show, :new, :create, :edit, :update, :destroy]
resources :installation_configs, only: [:index, :new, :create, :show, :edit, :update]
resources :agent_bots, only: [:index, :new, :create, :show, :edit, :update] do
delete :avatar, on: :member, action: :destroy_avatar

View File

@@ -10,7 +10,7 @@ module Enterprise::SuperAdmin::AppConfigsController
when 'internal'
@allowed_configs = internal_config_options
when 'captain'
@allowed_configs = %w[CAPTAIN_OPEN_AI_API_KEY CAPTAIN_OPEN_AI_MODEL]
@allowed_configs = %w[CAPTAIN_OPEN_AI_API_KEY CAPTAIN_OPEN_AI_MODEL CAPTAIN_FIRECRAWL_API_KEY]
else
super
end

View File

@@ -1,16 +1,32 @@
class Enterprise::Webhooks::FirecrawlController < ActionController::API
before_action :validate_token
def process_payload
if crawl_page_event?
Captain::Tools::FirecrawlParserJob.perform_later(
assistant_id: permitted_params[:assistant_id],
payload: permitted_params[:data]
)
end
Captain::Tools::FirecrawlParserJob.perform_later(assistant_id: assistant.id, payload: payload) if crawl_page_event?
head :ok
end
private
include Captain::FirecrawlHelper
def payload
permitted_params[:data]&.first&.to_h
end
def validate_token
render json: { error: 'Invalid access_token' }, status: :unauthorized if assistant_token != permitted_params[:token]
end
def assistant
@assistant ||= Captain::Assistant.find(permitted_params[:assistant_id])
end
def assistant_token
generate_firecrawl_token(assistant.id, assistant.account_id)
end
def crawl_page_event?
permitted_params[:type] == 'crawl.page'
end
@@ -19,12 +35,13 @@ class Enterprise::Webhooks::FirecrawlController < ActionController::API
params.permit(
:type,
:assistant_id,
:token,
:success,
:id,
:metadata,
:format,
:firecrawl,
{ data: {} }
data: [:markdown, { metadata: {} }]
)
end
end

View File

@@ -1,89 +0,0 @@
require 'administrate/base_dashboard'
class ResponseDashboard < Administrate::BaseDashboard
# ATTRIBUTE_TYPES
# a hash that describes the type of each of the model's fields.
#
# Each different type represents an Administrate::Field object,
# which determines how the attribute is displayed
# on pages throughout the dashboard.
ATTRIBUTE_TYPES = {
id: Field::Number.with_options(searchable: true),
account: Field::BelongsToSearch.with_options(class_name: 'Account', searchable_field: [:name, :id], order: 'id DESC'),
response_source: Field::BelongsToSearch.with_options(class_name: 'ResponseSource', searchable_field: [:name, :id, :source_link],
order: 'id DESC'),
answer: Field::Text.with_options(searchable: true),
question: Field::String.with_options(searchable: true),
status: Field::Select.with_options(searchable: false, collection: lambda { |field|
field.resource.class.send(field.attribute.to_s.pluralize).keys
}),
response_document: Field::BelongsToSearch.with_options(class_name: 'ResponseDocument', searchable_field: [:document_link, :content, :id],
order: 'id DESC'),
created_at: Field::DateTime,
updated_at: Field::DateTime
}.freeze
# COLLECTION_ATTRIBUTES
# an array of attributes that will be displayed on the model's index page.
#
# By default, it's limited to four items to reduce clutter on index pages.
# Feel free to add, remove, or rearrange items.
COLLECTION_ATTRIBUTES = %i[
id
question
answer
status
response_document
response_source
account
].freeze
# SHOW_PAGE_ATTRIBUTES
# an array of attributes that will be displayed on the model's show page.
SHOW_PAGE_ATTRIBUTES = %i[
id
status
question
answer
response_document
response_source
account
created_at
updated_at
].freeze
# FORM_ATTRIBUTES
# an array of attributes that will be displayed
# on the model's form (`new` and `edit`) pages.
FORM_ATTRIBUTES = %i[
response_source
response_document
question
answer
status
].freeze
# COLLECTION_FILTERS
# a hash that defines filters that can be used while searching via the search
# field of the dashboard.
#
# For example to add an option to search for open resources by typing "open:"
# in the search field:
#
# COLLECTION_FILTERS = {
# open: ->(resources) { resources.where(open: true) }
# }.freeze
COLLECTION_FILTERS = {
account: ->(resources, attr) { resources.where(account_id: attr) },
response_source: ->(resources, attr) { resources.where(response_source_id: attr) },
response_document: ->(resources, attr) { resources.where(response_document_id: attr) },
status: ->(resources, attr) { resources.where(status: attr) }
}.freeze
# Overwrite this method to customize how responses are displayed
# across all pages of the admin dashboard.
#
def display_resource(response)
"Response: ##{response.id} - #{response.question}"
end
end

View File

@@ -1,84 +0,0 @@
require 'administrate/base_dashboard'
class ResponseDocumentDashboard < Administrate::BaseDashboard
# ATTRIBUTE_TYPES
# a hash that describes the type of each of the model's fields.
#
# Each different type represents an Administrate::Field object,
# which determines how the attribute is displayed
# on pages throughout the dashboard.
ATTRIBUTE_TYPES = {
id: Field::Number.with_options(searchable: true),
account: Field::BelongsToSearch.with_options(class_name: 'Account', searchable_field: [:name, :id], order: 'id DESC'),
content: Field::Text.with_options(searchable: true),
document_id: Field::Number,
document_link: Field::String.with_options(searchable: true),
document_type: Field::String,
response_source: Field::BelongsToSearch.with_options(class_name: 'ResponseSource', searchable_field: [:name, :id, :source_link],
order: 'id DESC'),
responses: Field::HasMany,
created_at: Field::DateTime,
updated_at: Field::DateTime
}.freeze
# COLLECTION_ATTRIBUTES
# an array of attributes that will be displayed on the model's index page.
#
# By default, it's limited to four items to reduce clutter on index pages.
# Feel free to add, remove, or rearrange items.
COLLECTION_ATTRIBUTES = %i[
id
account
response_source
document_link
].freeze
# SHOW_PAGE_ATTRIBUTES
# an array of attributes that will be displayed on the model's show page.
SHOW_PAGE_ATTRIBUTES = %i[
id
account
response_source
document_link
document_id
document_type
content
created_at
updated_at
responses
].freeze
# FORM_ATTRIBUTES
# an array of attributes that will be displayed
# on the model's form (`new` and `edit`) pages.
FORM_ATTRIBUTES = %i[
account
response_source
document_link
document_id
document_type
content
].freeze
# COLLECTION_FILTERS
# a hash that defines filters that can be used while searching via the search
# field of the dashboard.
#
# For example to add an option to search for open resources by typing "open:"
# in the search field:
#
# COLLECTION_FILTERS = {
# open: ->(resources) { resources.where(open: true) }
# }.freeze
COLLECTION_FILTERS = {
account: ->(resources, attr) { resources.where(account_id: attr) },
response_source: ->(resources, attr) { resources.where(response_source_id: attr) }
}.freeze
# Overwrite this method to customize how response documents are displayed
# across all pages of the admin dashboard.
#
def display_resource(response_document)
"Document: ##{response_document.id} - #{response_document.document_link}"
end
end

View File

@@ -1,86 +0,0 @@
require 'administrate/base_dashboard'
class ResponseSourceDashboard < Administrate::BaseDashboard
# ATTRIBUTE_TYPES
# a hash that describes the type of each of the model's fields.
#
# Each different type represents an Administrate::Field object,
# which determines how the attribute is displayed
# on pages throughout the dashboard.
ATTRIBUTE_TYPES = {
id: Field::Number.with_options(searchable: true),
account: Field::BelongsToSearch.with_options(class_name: 'Account', searchable_field: [:name, :id], order: 'id DESC'),
name: Field::String.with_options(searchable: true),
response_documents: Field::HasMany,
responses: Field::HasMany,
source_link: Field::String.with_options(searchable: true),
source_model_id: Field::Number,
source_model_type: Field::String,
source_type: Field::Select.with_options(searchable: false, collection: lambda { |field|
field.resource.class.send(field.attribute.to_s.pluralize).keys
}),
created_at: Field::DateTime,
updated_at: Field::DateTime
}.freeze
# COLLECTION_ATTRIBUTES
# an array of attributes that will be displayed on the model's index page.
#
# By default, it's limited to four items to reduce clutter on index pages.
# Feel free to add, remove, or rearrange items.
COLLECTION_ATTRIBUTES = %i[
id
name
account
source_link
].freeze
# SHOW_PAGE_ATTRIBUTES
# an array of attributes that will be displayed on the model's show page.
SHOW_PAGE_ATTRIBUTES = %i[
id
name
account
source_link
source_model_id
source_model_type
source_type
created_at
updated_at
response_documents
responses
].freeze
# FORM_ATTRIBUTES
# an array of attributes that will be displayed
# on the model's form (`new` and `edit`) pages.
FORM_ATTRIBUTES = %i[
account
name
source_link
source_model_id
source_model_type
source_type
].freeze
# COLLECTION_FILTERS
# a hash that defines filters that can be used while searching via the search
# field of the dashboard.
#
# For example to add an option to search for open resources by typing "open:"
# in the search field:
#
# COLLECTION_FILTERS = {
# open: ->(resources) { resources.where(open: true) }
# }.freeze
COLLECTION_FILTERS = {
account: ->(resources, attr) { resources.where(account_id: attr) }
}.freeze
# Overwrite this method to customize how response sources are displayed
# across all pages of the admin dashboard.
#
def display_resource(response_source)
"Source: ##{response_source.id} - #{response_source.name}"
end
end

View File

@@ -0,0 +1,9 @@
module Captain::FirecrawlHelper
def generate_firecrawl_token(assistant_id, account_id)
api_key = InstallationConfig.find_by(name: 'CAPTAIN_FIRECRAWL_API_KEY')&.value
return nil unless api_key
token_base = "#{api_key[-4..]}#{assistant_id}#{account_id}"
Digest::SHA256.hexdigest(token_base)
end
end

View File

@@ -11,6 +11,8 @@ class Captain::Documents::CrawlJob < ApplicationJob
private
include Captain::FirecrawlHelper
def perform_simple_crawl(document)
page_links = Captain::Tools::SimplePageCrawlService.new(document.external_link).page_links
@@ -28,13 +30,22 @@ class Captain::Documents::CrawlJob < ApplicationJob
end
def perform_firecrawl_crawl(document)
webhook_url = Rails.application.routes.url_helpers.enterprise_webhooks_firecrawl_url
captain_usage_limits = document.account.usage_limits[:captain] || {}
document_limit = captain_usage_limits[:documents] || {}
crawl_limit = [document_limit[:available] || 10, 500].min
Captain::Tools::FirecrawlService
.new
.perform(
document.external_link,
"#{webhook_url}?assistant_id=#{document.assistant_id}"
firecrawl_webhook_url(document),
crawl_limit
)
end
def firecrawl_webhook_url(document)
webhook_url = Rails.application.routes.url_helpers.enterprise_webhooks_firecrawl_url
"#{webhook_url}?assistant_id=#{document.assistant_id}&token=#{generate_firecrawl_token(document.assistant_id, document.account_id)}"
end
end

View File

@@ -6,12 +6,12 @@ class Captain::Tools::FirecrawlParserJob < ApplicationJob
metadata = payload[:metadata]
document = assistant.documents.find_or_initialize_by(
external_link: metadata[:ogUrl]
external_link: metadata['url']
)
document.update!(
content: payload[:markdown],
name: metadata[:ogTitle],
name: metadata['title'],
status: :available
)
rescue StandardError => e

View File

@@ -29,6 +29,7 @@ class Captain::Document < ApplicationRecord
validates :external_link, presence: true
validates :external_link, uniqueness: { scope: :assistant_id }
validates :content, length: { maximum: 200_000 }
before_validation :ensure_account_id
enum status: {

View File

@@ -4,10 +4,10 @@ class Captain::Tools::FirecrawlService
raise 'Missing API key' if @api_key.nil?
end
def perform(url, webhook_url = '')
def perform(url, webhook_url, crawl_limit = 10)
HTTParty.post(
'https://api.firecrawl.dev/v1/crawl',
body: crawl_payload(url, webhook_url),
body: crawl_payload(url, webhook_url, crawl_limit),
headers: headers
)
rescue StandardError => e
@@ -16,12 +16,12 @@ class Captain::Tools::FirecrawlService
private
def crawl_payload(url, webhook_url)
def crawl_payload(url, webhook_url, crawl_limit)
{
url: url,
maxDepth: 50,
ignoreSitemap: false,
limit: 10,
limit: crawl_limit,
webhook: webhook_url,
scrapeOptions: {
onlyMainContent: false,

View File

@@ -1,5 +0,0 @@
<% content_for :title, "Robin AI playground: #{@response_source.name}" %>
<%= render_vue_component('PlaygroundIndex', {
responseSourceName: @response_source.name,
responseSourcePath: super_admin_response_source_path(@response_source)
}) %>

View File

@@ -1,71 +0,0 @@
<%#
# Show
This view is the template for the show page.
It renders the attributes of a resource,
as well as a link to its edit page.
## Local variables:
- `page`:
An instance of [Administrate::Page::Show][1].
Contains methods for accessing the resource to be displayed on the page,
as well as helpers for describing how each attribute of the resource
should be displayed.
[1]: http://www.rubydoc.info/gems/administrate/Administrate/Page/Show
%>
<% content_for(:title) { t("administrate.actions.show_resource", name: page.page_title) } %>
<header class="main-content__header">
<h1 class="main-content__page-title">
<%= content_for(:title) %>
</h1>
<div>
<%= link_to(
"Chat",
[:chat, namespace, page.resource],
class: "button"
) %>
<%= link_to(
"Edit",
[:edit, namespace, page.resource],
class: "button",
) if accessible_action?(page.resource, :edit) %>
<%= link_to(
t("administrate.actions.destroy"),
[namespace, page.resource],
class: "button button--danger",
method: :delete,
data: { confirm: t("administrate.actions.confirm") }
) if accessible_action?(page.resource, :destroy) %>
</div>
</header>
<section class="main-content__body">
<dl>
<% page.attributes.each do |title, attributes| %>
<fieldset class="<%= "field-unit--nested" if title.present? %>">
<% if title.present? %>
<legend><%= t "helpers.label.#{page.resource_name}.#{title}", default: title %></legend>
<% end %>
<% attributes.each do |attribute| %>
<dt class="attribute-label" id="<%= attribute.name %>">
<%= t(
"helpers.label.#{resource_name}.#{attribute.name}",
default: page.resource.class.human_attribute_name(attribute.name),
) %>
</dt>
<dd class="attribute-data attribute-data--<%=attribute.html_class%>"
><%= render_field attribute, page: page %></dd>
<% end %>
</fieldset>
<% end %>
</dl>
</section>

View File

@@ -1,57 +1,120 @@
require 'rails_helper'
RSpec.describe 'Firecrawl Webhooks', type: :request do
describe 'POST /enterprise/webhooks/firecrawl?assistant_id=:assistant_id' do
let(:assistant_id) { 'asst_123' }
describe 'POST /enterprise/webhooks/firecrawl?assistant_id=:assistant_id&token=:token' do
let!(:api_key) { create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test_api_key_123') }
let!(:account) { create(:account) }
let!(:assistant) { create(:captain_assistant, account: account) }
let(:payload_data) do
{
'markdown' => 'hello world',
'metadata' => {
'ogUrl' => 'https://example.com'
}
markdown: 'hello world',
metadata: { ogUrl: 'https://example.com' }
}
end
context 'with crawl.page event type' do
let(:valid_params) do
{
data: payload_data,
type: 'crawl.page'
}
# Generate actual token using the helper
let(:valid_token) do
token_base = "#{api_key.value[-4..]}#{assistant.id}#{assistant.account_id}"
Digest::SHA256.hexdigest(token_base)
end
context 'with valid token' do
context 'with crawl.page event type' do
let(:valid_params) do
{
type: 'crawl.page',
data: [payload_data]
}
end
it 'processes the webhook and returns success' do
expect(Captain::Tools::FirecrawlParserJob).to receive(:perform_later)
.with(
assistant_id: assistant.id,
payload: payload_data
)
post(
"/enterprise/webhooks/firecrawl?assistant_id=#{assistant.id}&token=#{valid_token}",
params: valid_params,
as: :json
)
expect(response).to have_http_status(:ok)
expect(response.body).to be_empty
end
end
it 'processes the webhook and returns success' do
expect(Captain::Tools::FirecrawlParserJob).to(
receive(:perform_later)
.with(
assistant_id: assistant_id,
payload: payload_data
)
)
context 'with crawl.completed event type' do
let(:valid_params) do
{
type: 'crawl.completed'
}
end
post("/enterprise/webhooks/firecrawl?assistant_id=#{assistant_id}",
params: valid_params,
as: :json)
it 'returns success without enqueuing job' do
expect(Captain::Tools::FirecrawlParserJob).not_to receive(:perform_later)
expect(response).to have_http_status(:ok)
expect(response.body).to be_empty
post("/enterprise/webhooks/firecrawl?assistant_id=#{assistant.id}&token=#{valid_token}",
params: valid_params,
as: :json)
expect(response).to have_http_status(:ok)
expect(response.body).to be_empty
end
end
end
context 'with crawl.completed event type' do
let(:valid_params) do
{ type: 'crawl.completed' }
context 'with invalid token' do
let(:invalid_params) do
{
type: 'crawl.page',
data: [payload_data]
}
end
it 'returns success without enqueuing job' do
expect(Captain::Tools::FirecrawlParserJob).not_to receive(:perform_later)
post("/enterprise/webhooks/firecrawl?assistant_id=#{assistant_id}",
params: valid_params,
it 'returns unauthorized status' do
post("/enterprise/webhooks/firecrawl?assistant_id=#{assistant.id}&token=invalid_token",
params: invalid_params,
as: :json)
expect(response).to have_http_status(:ok)
expect(response.body).to be_empty
expect(response).to have_http_status(:unauthorized)
end
end
context 'with invalid assistant_id' do
context 'with non-existent assistant_id' do
it 'returns not found status' do
post("/enterprise/webhooks/firecrawl?assistant_id=invalid_id&token=#{valid_token}",
params: { type: 'crawl.page', data: [payload_data] },
as: :json)
expect(response).to have_http_status(:not_found)
end
end
context 'with nil assistant_id' do
it 'returns not found status' do
post("/enterprise/webhooks/firecrawl?token=#{valid_token}",
params: { type: 'crawl.page', data: [payload_data] },
as: :json)
expect(response).to have_http_status(:not_found)
end
end
end
context 'when CAPTAIN_FIRECRAWL_API_KEY is not configured' do
before do
api_key.destroy
end
it 'returns unauthorized status' do
post("/enterprise/webhooks/firecrawl?assistant_id=#{assistant.id}&token=#{valid_token}",
params: { type: 'crawl.page', data: [payload_data] },
as: :json)
expect(response).to have_http_status(:unauthorized)
end
end
end

View File

@@ -8,21 +8,61 @@ RSpec.describe Captain::Documents::CrawlJob, type: :job do
describe '#perform' do
context 'when CAPTAIN_FIRECRAWL_API_KEY is configured' do
let(:firecrawl_service) { instance_double(Captain::Tools::FirecrawlService) }
let(:account) { document.account }
let(:token) { Digest::SHA256.hexdigest("-key#{document.assistant_id}#{document.account_id}") }
before do
allow(Captain::Tools::FirecrawlService).to receive(:new).and_return(firecrawl_service)
allow(firecrawl_service).to receive(:perform)
create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test-key')
end
it 'uses FirecrawlService to crawl the page' do
create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test-key')
context 'with account usage limits' do
before do
allow(account).to receive(:usage_limits).and_return({ captain: { documents: { available: 20 } } })
end
expect(firecrawl_service).to receive(:perform).with(
document.external_link,
"#{webhook_url}?assistant_id=#{assistant_id}"
)
it 'uses FirecrawlService with the correct crawl limit' do
expect(firecrawl_service).to receive(:perform).with(
document.external_link,
"#{webhook_url}?assistant_id=#{assistant_id}&token=#{token}",
20
)
described_class.perform_now(document)
described_class.perform_now(document)
end
end
context 'when crawl limit exceeds maximum' do
before do
allow(account).to receive(:usage_limits).and_return({ captain: { documents: { available: 1000 } } })
end
it 'caps the crawl limit at 500' do
expect(firecrawl_service).to receive(:perform).with(
document.external_link,
"#{webhook_url}?assistant_id=#{assistant_id}&token=#{token}",
500
)
described_class.perform_now(document)
end
end
context 'with no usage limits configured' do
before do
allow(account).to receive(:usage_limits).and_return({})
end
it 'uses default crawl limit of 10' do
expect(firecrawl_service).to receive(:perform).with(
document.external_link,
"#{webhook_url}?assistant_id=#{assistant_id}&token=#{token}",
10
)
described_class.perform_now(document)
end
end
end

View File

@@ -7,9 +7,9 @@ RSpec.describe Captain::Tools::FirecrawlParserJob, type: :job do
{
markdown: 'Launch Week I is here! 🚀',
metadata: {
title: 'Home - Firecrawl',
ogTitle: 'Firecrawl',
ogUrl: 'https://www.firecrawl.dev/'
'title' => 'Home - Firecrawl',
'ogTitle' => 'Firecrawl',
'url' => 'https://www.firecrawl.dev/'
}
}
end
@@ -22,8 +22,8 @@ RSpec.describe Captain::Tools::FirecrawlParserJob, type: :job do
document = assistant.documents.last
expect(document).to have_attributes(
content: payload[:markdown],
name: payload[:metadata][:ogTitle],
external_link: payload[:metadata][:ogUrl],
name: payload[:metadata]['title'],
external_link: payload[:metadata]['url'],
status: 'available'
)
end
@@ -32,7 +32,7 @@ RSpec.describe Captain::Tools::FirecrawlParserJob, type: :job do
existing_document = create(:captain_document,
assistant: assistant,
account: assistant.account,
external_link: payload[:metadata][:ogUrl],
external_link: payload[:metadata]['url'],
content: 'old content',
name: 'old title',
status: :in_progress)
@@ -44,7 +44,7 @@ RSpec.describe Captain::Tools::FirecrawlParserJob, type: :job do
existing_document.reload
expect(existing_document).to have_attributes(
content: payload[:markdown],
name: payload[:metadata][:ogTitle],
name: payload[:metadata]['title'],
status: 'available'
)
end

View File

@@ -0,0 +1,134 @@
require 'rails_helper'
RSpec.describe Captain::Tools::FirecrawlService do
let(:api_key) { 'test-api-key' }
let(:url) { 'https://example.com' }
let(:webhook_url) { 'https://webhook.example.com/callback' }
let(:crawl_limit) { 15 }
before do
create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: api_key)
end
describe '#initialize' do
context 'when API key is configured' do
it 'initializes successfully' do
expect { described_class.new }.not_to raise_error
end
end
context 'when API key is missing' do
before do
InstallationConfig.find_by(name: 'CAPTAIN_FIRECRAWL_API_KEY').destroy
end
it 'raises an error' do
expect { described_class.new }.to raise_error(ActiveRecord::RecordNotFound)
end
end
context 'when API key is nil' do
before do
InstallationConfig.find_by(name: 'CAPTAIN_FIRECRAWL_API_KEY').update(value: nil)
end
it 'raises an error' do
expect { described_class.new }.to raise_error('Missing API key')
end
end
end
describe '#perform' do
let(:service) { described_class.new }
let(:expected_payload) do
{
url: url,
maxDepth: 50,
ignoreSitemap: false,
limit: crawl_limit,
webhook: webhook_url,
scrapeOptions: {
onlyMainContent: false,
formats: ['markdown'],
excludeTags: ['iframe']
}
}.to_json
end
let(:expected_headers) do
{
'Authorization' => "Bearer #{api_key}",
'Content-Type' => 'application/json'
}
end
context 'when the API call is successful' do
before do
stub_request(:post, 'https://api.firecrawl.dev/v1/crawl')
.with(
body: expected_payload,
headers: expected_headers
)
.to_return(status: 200, body: '{"status": "success"}')
end
it 'makes a POST request with correct parameters' do
service.perform(url, webhook_url, crawl_limit)
expect(WebMock).to have_requested(:post, 'https://api.firecrawl.dev/v1/crawl')
.with(
body: expected_payload,
headers: expected_headers
)
end
it 'uses default crawl limit when not specified' do
default_payload = expected_payload.gsub(crawl_limit.to_s, '10')
stub_request(:post, 'https://api.firecrawl.dev/v1/crawl')
.with(
body: default_payload,
headers: expected_headers
)
.to_return(status: 200, body: '{"status": "success"}')
service.perform(url, webhook_url)
expect(WebMock).to have_requested(:post, 'https://api.firecrawl.dev/v1/crawl')
.with(
body: default_payload,
headers: expected_headers
)
end
end
context 'when the API call fails' do
before do
stub_request(:post, 'https://api.firecrawl.dev/v1/crawl')
.to_raise(StandardError.new('Connection failed'))
end
it 'raises an error with the failure message' do
expect { service.perform(url, webhook_url, crawl_limit) }
.to raise_error('Failed to crawl URL: Connection failed')
end
end
context 'when the API returns an error response' do
before do
stub_request(:post, 'https://api.firecrawl.dev/v1/crawl')
.to_return(status: 422, body: '{"error": "Invalid URL"}')
end
it 'makes the request but does not raise an error' do
expect { service.perform(url, webhook_url, crawl_limit) }.not_to raise_error
expect(WebMock).to have_requested(:post, 'https://api.firecrawl.dev/v1/crawl')
.with(
body: expected_payload,
headers: expected_headers
)
end
end
end
end

View File

@@ -1,8 +0,0 @@
# frozen_string_literal: true
FactoryBot.define do
factory :inbox_response_source do
inbox
response_source
end
end

View File

@@ -1,11 +0,0 @@
# frozen_string_literal: true
FactoryBot.define do
factory :response do
response_document
response_source
question { Faker::Lorem.sentence }
answer { Faker::Lorem.paragraph }
account
end
end

View File

@@ -1,10 +0,0 @@
# frozen_string_literal: true
FactoryBot.define do
factory :response_document do
response_source
content { Faker::Lorem.paragraph }
document_link { Faker::Internet.url }
account
end
end