all specs working at last

Author: SamSaffron

lib/completions/llm.rb

@@ -307,7 +307,7 @@ def initialize(dialect_klass, gateway_klass, llm_model, gateway: nil)
       # @param response_format { Hash - Optional } - JSON schema passed to the API as the desired structured output.
       # @param [Experimental] extra_model_params { Hash - Optional } - Other params that are not available accross models. e.g. response_format JSON schema.
       #
-      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response alongside a cancel function.
+      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response.
       #
       # @returns String | ToolCall - Completion result.
       # if multiple tools or a tool and a message come back, the result will be an array of ToolCall / String objects.

lib/discord/bot/persona_replier.rb

@@ -24,7 +24,7 @@ def handle_interaction!
         full_reply =
           @bot.reply(
             { conversation_context: [{ type: :user, content: @query }], skip_tool_details: true },
-          ) do |partial, _cancel, _something|
+          ) do |partial, _something|
             reply << partial
             next if reply.blank?
 

lib/summarization/fold_content.rb

@@ -18,7 +18,7 @@ def initialize(bot, strategy, persist_summaries: true)
       attr_reader :bot, :strategy
 
       # @param user { User } - User object used for auditing usage.
-      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response alongside a cancel function.
+      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response.
       # Note: The block is only called with results of the final summary, not intermediate summaries.
       #
       # This method doesn't care if we already have an up to date summary. It always regenerate.
@@ -77,7 +77,7 @@ def latest_sha
 
       # @param items { Array<Hash> } - Content to summarize. Structure will be: { poster: who wrote the content, id: a way to order content, text: content }
       # @param user { User } - User object used for auditing usage.
-      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response alongside a cancel function.
+      # @param &on_partial_blk { Block - Optional } - The passed block will get called with the LLM partial response.
       # Note: The block is only called with results of the final summary, not intermediate summaries.
       #
       # The summarization algorithm.
@@ -112,20 +112,20 @@ def fold(items, user, &on_partial_blk)
         summary = +""
 
         buffer_blk =
-          Proc.new do |partial, cancel, _, type|
+          Proc.new do |partial, _, type|
             if type == :structured_output
               json_summary_schema_key = bot.persona.response_format&.first.to_h
               partial_summary =
                 partial.read_latest_buffered_chunk[json_summary_schema_key[:key].to_sym]
 
               if partial_summary.present?
                 summary << partial_summary
-                on_partial_blk.call(partial_summary, cancel) if on_partial_blk
+                on_partial_blk.call(partial_summary) if on_partial_blk
               end
             elsif type.blank?
               # Assume response is a regular completion.
               summary << partial
-              on_partial_blk.call(partial, cancel) if on_partial_blk
+              on_partial_blk.call(partial) if on_partial_blk
             end
           end