mirror of
https://github.com/discourse/discourse-ai.git
synced 2025-03-06 17:30:20 +00:00
d07cf51653
Adds a comprehensive quota management system for LLM models that allows: - Setting per-group (applied per user in the group) token and usage limits with configurable durations - Tracking and enforcing token/usage limits across user groups - Quota reset periods (hourly, daily, weekly, or custom) - Admin UI for managing quotas with real-time updates This system provides granular control over LLM API usage by allowing admins to define limits on both total tokens and number of requests per group. Supports multiple concurrent quotas per model and automatically handles quota resets. Co-authored-by: Keegan George <kgeorge13@gmail.com>
86 lines
2.3 KiB
Ruby
86 lines
2.3 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
class LlmQuota < ActiveRecord::Base
|
|
self.table_name = "llm_quotas"
|
|
|
|
belongs_to :group
|
|
belongs_to :llm_model
|
|
has_many :llm_quota_usages
|
|
|
|
validates :group_id, presence: true
|
|
# we can not validate on create cause it breaks build
|
|
validates :llm_model_id, presence: true, on: :update
|
|
validates :duration_seconds, presence: true, numericality: { greater_than: 0 }
|
|
validates :max_tokens, numericality: { only_integer: true, greater_than: 0, allow_nil: true }
|
|
validates :max_usages, numericality: { greater_than: 0, allow_nil: true }
|
|
|
|
validate :at_least_one_limit
|
|
|
|
def self.check_quotas!(llm, user)
|
|
return true if user.blank?
|
|
quotas = joins(:group).where(llm_model: llm).where(group: user.groups)
|
|
|
|
return true if quotas.empty?
|
|
errors =
|
|
quotas.map do |quota|
|
|
usage = LlmQuotaUsage.find_or_create_for(user: user, llm_quota: quota)
|
|
begin
|
|
usage.check_quota!
|
|
nil
|
|
rescue LlmQuotaUsage::QuotaExceededError => e
|
|
e
|
|
end
|
|
end
|
|
|
|
return if errors.include?(nil)
|
|
|
|
raise errors.first
|
|
end
|
|
|
|
def self.log_usage(llm, user, input_tokens, output_tokens)
|
|
return if user.blank?
|
|
|
|
quotas = joins(:group).where(llm_model: llm).where(group: user.groups)
|
|
|
|
quotas.each do |quota|
|
|
usage = LlmQuotaUsage.find_or_create_for(user: user, llm_quota: quota)
|
|
usage.increment_usage!(input_tokens: input_tokens, output_tokens: output_tokens)
|
|
end
|
|
end
|
|
|
|
def available_tokens
|
|
max_tokens
|
|
end
|
|
|
|
def available_usages
|
|
max_usages
|
|
end
|
|
|
|
private
|
|
|
|
def at_least_one_limit
|
|
if max_tokens.nil? && max_usages.nil?
|
|
errors.add(:base, I18n.t("discourse_ai.errors.quota_required"))
|
|
end
|
|
end
|
|
end
|
|
|
|
# == Schema Information
|
|
#
|
|
# Table name: llm_quotas
|
|
#
|
|
# id :bigint not null, primary key
|
|
# group_id :bigint not null
|
|
# llm_model_id :bigint not null
|
|
# max_tokens :integer
|
|
# max_usages :integer
|
|
# duration_seconds :integer not null
|
|
# created_at :datetime not null
|
|
# updated_at :datetime not null
|
|
#
|
|
# Indexes
|
|
#
|
|
# index_llm_quotas_on_group_id_and_llm_model_id (group_id,llm_model_id) UNIQUE
|
|
# index_llm_quotas_on_llm_model_id (llm_model_id)
|
|
#
|