Module: Legion::LLM::Inference::Executor::ContextWindow

Included in:
Legion::LLM::Inference::Executor
Defined in:
lib/legion/llm/inference/executor/context_window.rb

Overview

ContextWindow methods extracted from Executor verbatim (P4b §1.5, refactor-under-green). Functional message-list transformations: empty/thinking/tool-result trimming and context-window-aware compaction. Operates on the messages argument; reads

Instance Method Summary collapse

Instance Method Details

#compact_to_fit(messages, target_tokens) ⇒ Object 



74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 74

def compact_to_fit(messages, target_tokens)
  return messages if estimate_message_tokens(messages) <= target_tokens

  filtered = messages.reject do |msg|
    cw_role(msg) == 'tool' && cw_text(msg).length >
      Legion::Settings[:llm][:tools][:context_compaction][:threshold_chars]
  end
  messages = filtered.map do |msg|
    next msg unless cw_role(msg) == 'tool'

    content = cw_text(msg)
    result_chars = Legion::Settings[:llm][:tools][:context_compaction][:result_chars]
    content.length > result_chars ? cw_with_content(msg, "#{content[0, result_chars]}\n[compacted]") : msg
  end

  return messages if estimate_message_tokens(messages) <= target_tokens

  messages = messages.last(messages.size / 2) while messages.size > 2 && estimate_message_tokens(messages) > target_tokens
  messages
end

#cw_content(msg) ⇒ Object 



241
242
243
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 241

def cw_content(msg)
  msg.content
end

#cw_role(msg) ⇒ Object

G3: the pipeline message world is Canonical::Message — these are plain member reads (the former Canonical-OR-Hash dual shape was the split-world seam).



237
238
239
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 237

def cw_role(msg)
  msg.role.to_s
end

#cw_text(msg) ⇒ Object 



245
246
247
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 245

def cw_text(msg)
  msg.text.to_s
end

#cw_tool_call_id(msg) ⇒ Object 



249
250
251
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 249

def cw_tool_call_id(msg)
  msg.tool_call_id
end

#cw_tool_calls(msg) ⇒ Object 



253
254
255
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 253

def cw_tool_calls(msg)
  msg.tool_calls
end

#cw_with_content(msg, content) ⇒ Object 



257
258
259
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 257

def cw_with_content(msg, content)
  msg.with(content: content)
end

#empty_assistant_message?(msg) ⇒ Boolean

Returns:

  • (Boolean) 


221
222
223
224
225
226
227
228
229
230
231
232
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 221

def empty_assistant_message?(msg)
  return false unless cw_role(msg) == 'assistant'

  content = cw_content(msg)
  has_content = content.is_a?(String) ? !content.strip.empty? : !content.nil?
  return false if has_content

  tool_calls = cw_tool_calls(msg)
  return false if tool_calls.is_a?(Array) && tool_calls.any?

  true
end

#enforce_context_window(messages) ⇒ Object 



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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 29

def enforce_context_window(messages)
  context_window = resolved_context_window
  @context_accounting[:component_status][:context_window] = :observed
  return messages unless context_window&.positive?

  threshold = (context_window * Legion::Settings[:llm][:context_curation][:context_window_threshold]).to_i
  tool_budget = estimate_tool_token_budget
  available_for_messages = threshold - tool_budget
  estimated = estimate_message_tokens(messages)
  return messages if estimated <= available_for_messages

  log.warn "[llm][executor] action=context_compaction request_id=#{@request.id} " \
           "estimated_tokens=#{estimated} context_window=#{context_window} " \
           "threshold=#{threshold} tool_budget=#{tool_budget} available=#{available_for_messages}"

  preserve_after = last_user_message_index(messages)
  recent = messages[preserve_after..]
  older = messages[0...preserve_after]

  target_tokens = available_for_messages - estimate_message_tokens(recent)
  compacted = compact_to_fit(older, target_tokens)

  result = compacted + recent
  after_tokens = estimate_message_tokens(result)
  saved = [estimated - after_tokens, 0].max

  @context_accounting[:tokens][:context_window_saved_estimated_tokens] += saved
  @context_accounting[:counts][:context_window_message_count_before] = messages.size
  @context_accounting[:counts][:context_window_message_count_after] = result.size
  @context_accounting[:events] << ContextAccounting.event(
    event_type:    :context_window_enforcement,
    component:     :context_window,
    before_tokens: estimated,
    after_tokens:  after_tokens,
    before_count:  messages.size,
    after_count:   result.size,
    metadata:      { context_window: context_window, threshold: threshold }
  )

  log.info "[llm][executor] action=context_compaction_complete request_id=#{@request.id} " \
           "before=#{messages.size} after=#{result.size} " \
           "tokens_before=#{estimated} tokens_after=#{after_tokens}"
  result
end

#estimate_message_tokens(messages) ⇒ Object 



101
102
103
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 101

def estimate_message_tokens(messages)
  messages.sum { |m| (cw_text(m).length / 4.0).ceil }
end

#estimate_tool_token_budget ⇒ Object 



105
106
107
108
109
110
111
112
113
114
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 105

def estimate_tool_token_budget
  tools = @request.tools
  return 0 if tools.nil? || tools.empty?

  tool_list = tools.is_a?(Hash) ? tools.values : Array(tools)
  tool_list.sum do |tool|
    json_repr = tool.respond_to?(:to_h) ? Legion::JSON.dump(tool.to_h) : tool.to_s
    (json_repr.length / 3.5).ceil
  end
end

#last_user_message_index(messages) ⇒ Object 



213
214
215
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 213

def last_user_message_index(messages)
  messages.rindex { |m| cw_role(m) == 'user' } || messages.size
end

#native_dispatch_messages ⇒ Object 



12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 12

def native_dispatch_messages
  messages = apply_conversation_breakpoint(@request.messages)
  rejected = messages.count { |m| empty_assistant_message?(m) }
  if rejected.positive?
    log.warn "[llm][executor] action=strip_empty_assistants request_id=#{@request.id} removed=#{rejected}"
    messages = messages.reject { |m| empty_assistant_message?(m) }
  end
  messages = strip_thinking_from_history(messages)
  post_thinking_rejected = messages.count { |m| empty_assistant_message?(m) }
  if post_thinking_rejected.positive?
    log.warn "[llm][executor] action=strip_empty_assistants_post_thinking request_id=#{@request.id} removed=#{post_thinking_rejected}"
    messages = messages.reject { |m| empty_assistant_message?(m) }
  end
  messages = trim_oversized_tool_results(messages)
  enforce_context_window(messages)
end

#reduce_messages_for_dispatch(messages) ⇒ Object

Lane-INDEPENDENT reduction applied before dispatch: empty-assistant prune + leading-thinking strip + oversized-tool-result trim. PURE — no call it to measure exactly what native_dispatch_messages will send (minus the lane-dependent enforce_context_window compaction, which is correctly excluded from routing since it depends on the chosen lane).



122
123
124
125
126
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 122

def reduce_messages_for_dispatch(messages)
  msgs = Array(messages).reject { |m| empty_assistant_message?(m) }
  msgs = strip_thinking_pure(msgs)
  trim_oversized_tool_results_pure(msgs)
end

#resolved_context_window ⇒ Object 



95
96
97
98
99
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 95

def resolved_context_window
  @resolved_offering_metadata&.dig(:limits, :context_window) ||
    @resolved_offering_metadata&.dig(:context_window) ||
    @resolved_offering_metadata&.dig('limits', 'context_window')
end

#strip_leading_thinking_block(text) ⇒ Object 



172
173
174
175
176
177
178
179
180
181
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 172

def strip_leading_thinking_block(text)
  result = text.lstrip
  THINKING_TAG_PAIRS.each do |open_tag, close_tag|
    next unless result.start_with?(open_tag)

    close_idx = result.index(close_tag, open_tag.length)
    return close_idx ? result[(close_idx + close_tag.length)..].lstrip : ''
  end
  text
end

#strip_thinking_from_history(messages) ⇒ Object 



146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 146

def strip_thinking_from_history(messages)
  before_tokens = ContextAccounting.estimate_message_tokens(messages)
  result = strip_thinking_pure(messages)
  stripped_count = messages.zip(result).count { |before, after| before != after }

  after_tokens = ContextAccounting.estimate_message_tokens(result)
  saved = [before_tokens - after_tokens, 0].max
  @context_accounting[:component_status][:thinking_strip] = :observed
  if saved.positive?
    @context_accounting[:tokens][:stripped_thinking_estimated_tokens] += saved
    @context_accounting[:counts][:stripped_thinking_message_count] += stripped_count
    @context_accounting[:events] << ContextAccounting.event(
      event_type:    :thinking_stripped,
      component:     :stripped_thinking,
      before_tokens: before_tokens,
      after_tokens:  after_tokens,
      before_count:  messages.size,
      after_count:   result.size,
      metadata:      { stripped_count: stripped_count }
    )
  end

  log.info "[llm][executor] action=strip_thinking_history request_id=#{@request.id} stripped=#{stripped_count}" if stripped_count.positive?
  result
end

#strip_thinking_pure(messages) ⇒ Object

Pure leading-thinking strip. Shared by strip_thinking_from_history (which adds accounting) and reduce_messages_for_dispatch.



130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 130

def strip_thinking_pure(messages)
  preserve_after = last_user_message_index(messages)
  messages.each_with_index.map do |msg, idx|
    next msg if idx >= preserve_after
    next msg unless cw_role(msg) == 'assistant'

    content = cw_content(msg)
    next msg unless content.is_a?(String)

    cleaned = strip_leading_thinking_block(content)
    next msg if cleaned == content

    cw_with_content(msg, cleaned)
  end
end

#tool_result_message?(msg) ⇒ Boolean

Returns:

  • (Boolean) 


217
218
219
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 217

def tool_result_message?(msg)
  cw_role(msg) == 'tool' || !cw_tool_call_id(msg).nil?
end

#trim_oversized_tool_results(messages) ⇒ Object 



203
204
205
206
207
208
209
210
211
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 203

def trim_oversized_tool_results(messages)
  result = trim_oversized_tool_results_pure(messages)
  trimmed_count = messages.zip(result).count { |before, after| before != after }
  if trimmed_count.positive?
    log.info "[llm][executor] action=trim_tool_results request_id=#{@request.id} trimmed=#{trimmed_count} " \
             "max_chars=#{Legion::Settings[:llm][:tools][:result_max_dispatch_chars]}"
  end
  result
end

#trim_oversized_tool_results_pure(messages) ⇒ Object

Pure oversized-tool-result trim. Shared by trim_oversized_tool_results (which adds logging) and reduce_messages_for_dispatch.



185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
 
# File 'lib/legion/llm/inference/executor/context_window.rb', line 185

def trim_oversized_tool_results_pure(messages)
  max_chars = Legion::Settings[:llm][:tools][:result_max_dispatch_chars]
  return messages unless max_chars.positive?

  preserve_after = last_user_message_index(messages)
  messages.each_with_index.map do |msg, idx|
    next msg if idx >= preserve_after
    next msg unless tool_result_message?(msg)

    content = cw_content(msg)
    next msg unless content.is_a?(String) && content.length > max_chars

    cw_with_content(msg, "#{content[0, max_chars]}\n\n[TRUNCATED: showing first #{max_chars} of #{content.length} chars. " \
                         'If you need more content, make multiple smaller targeted requests ' \
                         '(e.g. read specific line ranges, grep for specific patterns, or request smaller sections).]')
  end
end