openclacky 1.3.6 → 1.3.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/CHANGELOG.md +54 -0
- data/lib/clacky/agent/hook_manager.rb +10 -1
- data/lib/clacky/agent/system_prompt_builder.rb +1 -1
- data/lib/clacky/agent.rb +17 -3
- data/lib/clacky/agent_profile.rb +124 -53
- data/lib/clacky/billing/billing_store.rb +3 -2
- data/lib/clacky/brand_config.rb +309 -6
- data/lib/clacky/cli.rb +4 -57
- data/lib/clacky/default_extensions/coding/agents/coding/avatar.png +0 -0
- data/lib/clacky/default_extensions/coding/ext.yml +19 -0
- data/lib/clacky/default_extensions/ext-studio/agents/ext-developer/avatar.png +0 -0
- data/lib/clacky/default_extensions/ext-studio/agents/ext-developer/system_prompt.md +95 -0
- data/lib/clacky/default_extensions/ext-studio/api/handler.rb +254 -0
- data/lib/clacky/default_extensions/ext-studio/ext.yml +33 -0
- data/lib/clacky/default_extensions/ext-studio/panels/studio/view.js +1288 -0
- data/lib/clacky/default_extensions/ext-studio/skills/ext-debug/SKILL.md +71 -0
- data/lib/clacky/default_extensions/ext-studio/skills/ext-publish/SKILL.md +74 -0
- data/lib/clacky/default_extensions/ext-studio/skills/ext-scaffold/SKILL.md +65 -0
- data/lib/clacky/default_extensions/general/agents/general/avatar.png +0 -0
- data/lib/clacky/default_extensions/general/ext.yml +19 -0
- data/lib/clacky/default_extensions/git/ext.yml +12 -0
- data/lib/clacky/{default_agents/_panels/git/panel.js → default_extensions/git/panels/git/view.js} +4 -5
- data/lib/clacky/default_extensions/meeting/{handler.rb → api/handler.rb} +78 -26
- data/lib/clacky/default_extensions/meeting/ext.yml +15 -0
- data/lib/clacky/default_extensions/meeting/{meeting.js → panels/meeting/view.js} +266 -90
- data/lib/clacky/default_extensions/meeting/skills/meeting-summarizer/SKILL.md +2 -1
- data/lib/clacky/default_extensions/time_machine/ext.yml +12 -0
- data/lib/clacky/{default_agents/_panels/time_machine/panel.js → default_extensions/time_machine/panels/time_machine/view.js} +58 -47
- data/lib/clacky/default_skills/channel-manager/discord_setup.rb +1 -1
- data/lib/clacky/default_skills/extend-openclacky/SKILL.md +83 -18
- data/lib/clacky/{api_extension.rb → extension/api_extension.rb} +27 -11
- data/lib/clacky/extension/api_loader.rb +136 -0
- data/lib/clacky/extension/cli_commands.rb +226 -0
- data/lib/clacky/{server/api_extension_dispatcher.rb → extension/dispatcher.rb} +22 -9
- data/lib/clacky/extension/hook_loader.rb +77 -0
- data/lib/clacky/extension/loader.rb +489 -0
- data/lib/clacky/extension/packager.rb +226 -0
- data/lib/clacky/{patch_loader.rb → extension/patch_loader.rb} +45 -0
- data/lib/clacky/extension/scaffold/template_renderer.rb +64 -0
- data/lib/clacky/extension/scaffold/templates/full/README.md.erb +48 -0
- data/lib/clacky/extension/scaffold/templates/full/agents/designer.md +5 -0
- data/lib/clacky/extension/scaffold/templates/full/api/handler.rb.erb +16 -0
- data/lib/clacky/extension/scaffold/templates/full/channels/noop.rb.erb +40 -0
- data/lib/clacky/extension/scaffold/templates/full/ext.yml.erb +41 -0
- data/lib/clacky/extension/scaffold/templates/full/hooks/audit.rb +9 -0
- data/lib/clacky/extension/scaffold/templates/full/panels/dashboard/view.js.erb +194 -0
- data/lib/clacky/extension/scaffold/templates/full/patches/audit.rb +15 -0
- data/lib/clacky/extension/scaffold/templates/full/skills/__slug__-skill/SKILL.md.erb +8 -0
- data/lib/clacky/extension/scaffold/templates/hello/api/handler.rb.erb +9 -0
- data/lib/clacky/extension/scaffold/templates/hello/ext.yml.erb +12 -0
- data/lib/clacky/extension/scaffold/templates/hello/panels/hello/view.js.erb +31 -0
- data/lib/clacky/extension/scaffold.rb +55 -0
- data/lib/clacky/extension/verifier.rb +196 -0
- data/lib/clacky/identity.rb +70 -0
- data/lib/clacky/locales/en.rb +19 -1
- data/lib/clacky/locales/zh.rb +19 -1
- data/lib/clacky/platform_http_client.rb +18 -15
- data/lib/clacky/server/channel/extension_adapter_loader.rb +32 -0
- data/lib/clacky/server/channel.rb +6 -0
- data/lib/clacky/server/http_server.rb +439 -146
- data/lib/clacky/server/session_registry.rb +9 -2
- data/lib/clacky/shell_hook_loader.rb +266 -22
- data/lib/clacky/skill_loader.rb +28 -1
- data/lib/clacky/ui2/components/welcome_banner.rb +1 -1
- data/lib/clacky/ui2/ui_controller.rb +9 -6
- data/lib/clacky/utils/workspace_rules.rb +2 -2
- data/lib/clacky/version.rb +1 -1
- data/lib/clacky/web/app.css +1034 -675
- data/lib/clacky/web/app.js +63 -15
- data/lib/clacky/web/auth.js +2 -0
- data/lib/clacky/web/components/notify.js +2 -0
- data/lib/clacky/web/components/sidebar.js +1 -5
- data/lib/clacky/web/core/ext.js +432 -89
- data/lib/clacky/web/features/backup/store.js +1 -0
- data/lib/clacky/web/features/backup/view.js +0 -1
- data/lib/clacky/web/features/brand/view.js +13 -5
- data/lib/clacky/web/features/extensions/store.js +184 -0
- data/lib/clacky/web/features/extensions/view.js +390 -0
- data/lib/clacky/web/features/new-session/store.js +199 -0
- data/lib/clacky/web/features/new-session/view.js +619 -0
- data/lib/clacky/web/features/share/view.js +63 -14
- data/lib/clacky/web/features/skills/store.js +3 -1
- data/lib/clacky/web/features/skills/view.js +1 -1
- data/lib/clacky/web/features/tasks/view.js +12 -0
- data/lib/clacky/web/features/trash/store.js +24 -6
- data/lib/clacky/web/features/trash/view.js +48 -12
- data/lib/clacky/web/features/workspace/store.js +2 -0
- data/lib/clacky/web/features/workspace/view.js +0 -1
- data/lib/clacky/web/i18n.js +152 -106
- data/lib/clacky/web/index.html +172 -190
- data/lib/clacky/web/sessions.js +67 -272
- data/lib/clacky/web/settings.js +1 -1
- data/lib/clacky/web/skills.js +88 -48
- data/lib/clacky/web/theme.js +5 -0
- data/lib/clacky/web/ws-dispatcher.js +12 -2
- data/lib/clacky/web/ws.js +2 -0
- data/lib/clacky.rb +21 -10
- metadata +52 -21
- data/benchmark/runner.rb +0 -441
- data/lib/clacky/api_extension_loader.rb +0 -168
- data/lib/clacky/default_agents/SOUL.md +0 -3
- data/lib/clacky/default_agents/USER.md +0 -1
- data/lib/clacky/default_agents/coding/profile.yml +0 -5
- data/lib/clacky/default_agents/coding/webui/.gitkeep +0 -0
- data/lib/clacky/default_agents/general/profile.yml +0 -2
- data/lib/clacky/default_extensions/meeting/meta.yml +0 -3
- data/lib/clacky/web/features/creator/store.js +0 -81
- data/lib/clacky/web/features/creator/view.js +0 -380
- /data/lib/clacky/{default_agents → default_extensions/coding/agents}/coding/system_prompt.md +0 -0
- /data/lib/clacky/{default_agents → default_extensions/general/agents}/general/system_prompt.md +0 -0
- /data/lib/clacky/{default_agents/base_prompt.md → prompts/base.md} +0 -0
data/benchmark/runner.rb
DELETED
|
@@ -1,441 +0,0 @@
|
|
|
1
|
-
#!/usr/bin/env ruby
|
|
2
|
-
# frozen_string_literal: true
|
|
3
|
-
|
|
4
|
-
# Benchmark runner for system prompt A/B testing
|
|
5
|
-
# Usage: ruby benchmark/runner.rb
|
|
6
|
-
|
|
7
|
-
require "fileutils"
|
|
8
|
-
require "json"
|
|
9
|
-
require "tmpdir"
|
|
10
|
-
|
|
11
|
-
project_root = File.expand_path("..", __dir__)
|
|
12
|
-
$LOAD_PATH.unshift File.join(project_root, "lib")
|
|
13
|
-
require "clacky"
|
|
14
|
-
|
|
15
|
-
class BenchmarkRunner
|
|
16
|
-
PROMPT_FILES = {
|
|
17
|
-
base: "lib/clacky/default_agents/base_prompt.md",
|
|
18
|
-
coding: "lib/clacky/default_agents/coding/system_prompt.md",
|
|
19
|
-
general: "lib/clacky/default_agents/general/system_prompt.md"
|
|
20
|
-
}.freeze
|
|
21
|
-
|
|
22
|
-
FIXTURE_DIR = File.expand_path("fixtures/sample_project", __dir__)
|
|
23
|
-
RESULTS_DIR = File.expand_path("results", __dir__)
|
|
24
|
-
|
|
25
|
-
TASKS = [
|
|
26
|
-
{
|
|
27
|
-
name: "simple_edit",
|
|
28
|
-
description: "Rename methods to snake_case across files",
|
|
29
|
-
prompt: "Rename the method `calculateTotal` to `calculate_total` in all files. Also rename `calculateTotalWithTax` to `calculate_total_with_tax` and `applyDiscount` to `apply_discount`. Update all references and the test file too.",
|
|
30
|
-
agent_profile: "coding"
|
|
31
|
-
},
|
|
32
|
-
{
|
|
33
|
-
name: "feature_addition",
|
|
34
|
-
description: "Add a new /products API endpoint with tests",
|
|
35
|
-
prompt: "Add a new `/products` endpoint to the ApiHandler that returns products from the store with optional pagination via `page` and `per_page` params. Also create a test file `spec/api_handler_spec.rb` with basic tests for this endpoint.",
|
|
36
|
-
agent_profile: "coding"
|
|
37
|
-
},
|
|
38
|
-
{
|
|
39
|
-
name: "refactoring",
|
|
40
|
-
description: "Extract helper method from duplicated pattern",
|
|
41
|
-
prompt: "In order_calculator.rb, both `calculate_total_with_tax` and `apply_discount` call `calculate_total` as their first step. Refactor to eliminate this duplication in the cleanest way possible. Do not over-engineer.",
|
|
42
|
-
agent_profile: "coding"
|
|
43
|
-
},
|
|
44
|
-
{
|
|
45
|
-
name: "bug_fix",
|
|
46
|
-
description: "Fix XSS vulnerability in HTML rendering",
|
|
47
|
-
prompt: "Fix the XSS vulnerability in user_renderer.rb. The methods directly interpolate user input into HTML without escaping. Make the rendering safe against XSS attacks.",
|
|
48
|
-
agent_profile: "coding"
|
|
49
|
-
},
|
|
50
|
-
{
|
|
51
|
-
name: "git_workflow",
|
|
52
|
-
description: "Fix bug and stage changes safely with git",
|
|
53
|
-
prompt: "Fix the XSS vulnerability in user_renderer.rb, then use git to stage only the changed file for commit. Do NOT stage all files.",
|
|
54
|
-
agent_profile: "coding"
|
|
55
|
-
}
|
|
56
|
-
].freeze
|
|
57
|
-
|
|
58
|
-
def initialize
|
|
59
|
-
@project_root = File.expand_path("..", __dir__)
|
|
60
|
-
@original_prompts = read_current_prompts
|
|
61
|
-
FileUtils.mkdir_p(RESULTS_DIR)
|
|
62
|
-
end
|
|
63
|
-
|
|
64
|
-
def run
|
|
65
|
-
run_baseline
|
|
66
|
-
run_treatment
|
|
67
|
-
run_report
|
|
68
|
-
end
|
|
69
|
-
|
|
70
|
-
def run_baseline
|
|
71
|
-
puts "=" * 70
|
|
72
|
-
puts "OpenClacky System Prompt Benchmark - BASELINE"
|
|
73
|
-
puts "=" * 70
|
|
74
|
-
puts "Project: #{@project_root}"
|
|
75
|
-
puts "Model: #{agent_config.model_name}"
|
|
76
|
-
puts "Tasks: #{TASKS.length}"
|
|
77
|
-
puts
|
|
78
|
-
|
|
79
|
-
unless git_clean?
|
|
80
|
-
puts "WARNING: Prompt files have uncommitted changes. Baseline may not reflect main."
|
|
81
|
-
puts
|
|
82
|
-
end
|
|
83
|
-
|
|
84
|
-
baseline_prompts = read_baseline_prompts
|
|
85
|
-
write_prompts(baseline_prompts)
|
|
86
|
-
results = run_all_tasks(:baseline)
|
|
87
|
-
write_results("baseline", results)
|
|
88
|
-
|
|
89
|
-
# Restore treatment prompts
|
|
90
|
-
write_prompts(@original_prompts)
|
|
91
|
-
puts "\nBaseline complete. Results saved."
|
|
92
|
-
results
|
|
93
|
-
rescue => e
|
|
94
|
-
write_prompts(@original_prompts)
|
|
95
|
-
raise
|
|
96
|
-
end
|
|
97
|
-
|
|
98
|
-
def run_treatment
|
|
99
|
-
puts "=" * 70
|
|
100
|
-
puts "OpenClacky System Prompt Benchmark - TREATMENT"
|
|
101
|
-
puts "=" * 70
|
|
102
|
-
puts "Project: #{@project_root}"
|
|
103
|
-
puts "Model: #{agent_config.model_name}"
|
|
104
|
-
puts "Tasks: #{TASKS.length}"
|
|
105
|
-
puts
|
|
106
|
-
|
|
107
|
-
# Ensure treatment prompts are active
|
|
108
|
-
write_prompts(@original_prompts)
|
|
109
|
-
results = run_all_tasks(:treatment)
|
|
110
|
-
write_results("treatment", results)
|
|
111
|
-
|
|
112
|
-
puts "\nTreatment complete. Results saved."
|
|
113
|
-
results
|
|
114
|
-
end
|
|
115
|
-
|
|
116
|
-
def run_report
|
|
117
|
-
baseline_file = Dir.glob(File.join(RESULTS_DIR, "baseline_*.json")).max
|
|
118
|
-
treatment_file = Dir.glob(File.join(RESULTS_DIR, "treatment_*.json")).max
|
|
119
|
-
|
|
120
|
-
unless baseline_file
|
|
121
|
-
puts "ERROR: No baseline results found in #{RESULTS_DIR}"
|
|
122
|
-
exit 1
|
|
123
|
-
end
|
|
124
|
-
unless treatment_file
|
|
125
|
-
puts "ERROR: No treatment results found in #{RESULTS_DIR}"
|
|
126
|
-
exit 1
|
|
127
|
-
end
|
|
128
|
-
|
|
129
|
-
baseline = JSON.parse(File.read(baseline_file), symbolize_names: true)
|
|
130
|
-
treatment = JSON.parse(File.read(treatment_file), symbolize_names: true)
|
|
131
|
-
|
|
132
|
-
puts "=" * 70
|
|
133
|
-
puts "COMPARISON REPORT"
|
|
134
|
-
puts "=" * 70
|
|
135
|
-
puts "Baseline: #{File.basename(baseline_file)}"
|
|
136
|
-
puts "Treatment: #{File.basename(treatment_file)}"
|
|
137
|
-
puts
|
|
138
|
-
compare_and_print(baseline, treatment)
|
|
139
|
-
|
|
140
|
-
# Save combined report
|
|
141
|
-
report_path = File.join(RESULTS_DIR, "report_#{timestamp}.json")
|
|
142
|
-
File.write(report_path, JSON.pretty_generate({
|
|
143
|
-
baseline: baseline,
|
|
144
|
-
treatment: treatment,
|
|
145
|
-
meta: {
|
|
146
|
-
model: agent_config.model_name,
|
|
147
|
-
timestamp: Time.now.iso8601,
|
|
148
|
-
tasks: TASKS.map { |t| t[:name] }
|
|
149
|
-
}
|
|
150
|
-
}))
|
|
151
|
-
puts
|
|
152
|
-
puts "Full report saved to: #{report_path}"
|
|
153
|
-
end
|
|
154
|
-
|
|
155
|
-
private
|
|
156
|
-
|
|
157
|
-
def agent_config
|
|
158
|
-
@agent_config ||= Clacky::AgentConfig.load
|
|
159
|
-
end
|
|
160
|
-
|
|
161
|
-
def read_current_prompts
|
|
162
|
-
prompts = {}
|
|
163
|
-
PROMPT_FILES.each do |key, rel_path|
|
|
164
|
-
full_path = File.join(@project_root, rel_path)
|
|
165
|
-
prompts[key] = File.read(full_path)
|
|
166
|
-
end
|
|
167
|
-
prompts
|
|
168
|
-
end
|
|
169
|
-
|
|
170
|
-
def read_baseline_prompts
|
|
171
|
-
prompts = {}
|
|
172
|
-
PROMPT_FILES.each do |key, rel_path|
|
|
173
|
-
content = `git -C "#{@project_root}" show main:"#{rel_path}" 2>/dev/null`
|
|
174
|
-
if $?.success? && !content.empty?
|
|
175
|
-
prompts[key] = content
|
|
176
|
-
else
|
|
177
|
-
puts " Warning: Could not read #{rel_path} from main, using current"
|
|
178
|
-
prompts[key] = @original_prompts[key]
|
|
179
|
-
end
|
|
180
|
-
end
|
|
181
|
-
prompts
|
|
182
|
-
end
|
|
183
|
-
|
|
184
|
-
def write_prompts(prompts)
|
|
185
|
-
prompts.each do |key, content|
|
|
186
|
-
rel_path = PROMPT_FILES[key]
|
|
187
|
-
full_path = File.join(@project_root, rel_path)
|
|
188
|
-
File.write(full_path, content)
|
|
189
|
-
end
|
|
190
|
-
end
|
|
191
|
-
|
|
192
|
-
def git_clean?
|
|
193
|
-
PROMPT_FILES.values.all? do |rel_path|
|
|
194
|
-
status = `git -C "#{@project_root}" status --porcelain "#{rel_path}" 2>/dev/null`
|
|
195
|
-
status.strip.empty?
|
|
196
|
-
end
|
|
197
|
-
end
|
|
198
|
-
|
|
199
|
-
def run_all_tasks(variant)
|
|
200
|
-
results = {}
|
|
201
|
-
TASKS.each_with_index do |task, idx|
|
|
202
|
-
puts
|
|
203
|
-
puts "[#{idx + 1}/#{TASKS.length}] #{task[:name]}: #{task[:description]}"
|
|
204
|
-
results[task[:name]] = run_task(task, variant)
|
|
205
|
-
end
|
|
206
|
-
results
|
|
207
|
-
end
|
|
208
|
-
|
|
209
|
-
def run_task(task, variant)
|
|
210
|
-
tmp_dir = File.join(Dir.tmpdir, "clacky_benchmark_#{variant}_#{task[:name]}_#{Process.pid}_#{Time.now.to_i}")
|
|
211
|
-
FileUtils.cp_r(FIXTURE_DIR, tmp_dir)
|
|
212
|
-
|
|
213
|
-
# Ensure tmp_dir is a git repo (cp_r preserves .git)
|
|
214
|
-
Dir.chdir(tmp_dir) do
|
|
215
|
-
system("git config user.email 'benchmark@test.com' >/dev/null 2>&1")
|
|
216
|
-
system("git config user.name 'Benchmark' >/dev/null 2>&1")
|
|
217
|
-
end
|
|
218
|
-
|
|
219
|
-
config = agent_config.dup
|
|
220
|
-
config.permission_mode = :auto_approve
|
|
221
|
-
|
|
222
|
-
client = Clacky::Client.new(
|
|
223
|
-
config.api_key,
|
|
224
|
-
base_url: config.base_url,
|
|
225
|
-
model: config.model_name,
|
|
226
|
-
anthropic_format: config.anthropic_format?
|
|
227
|
-
)
|
|
228
|
-
|
|
229
|
-
agent = Clacky::Agent.new(
|
|
230
|
-
client, config,
|
|
231
|
-
working_dir: tmp_dir,
|
|
232
|
-
ui: BenchmarkUI.new,
|
|
233
|
-
profile: task[:agent_profile],
|
|
234
|
-
session_id: Clacky::SessionManager.generate_id,
|
|
235
|
-
source: :manual
|
|
236
|
-
)
|
|
237
|
-
|
|
238
|
-
start_time = Time.now
|
|
239
|
-
agent.run(task[:prompt])
|
|
240
|
-
duration = Time.now - start_time
|
|
241
|
-
|
|
242
|
-
# Collect metrics
|
|
243
|
-
metrics = {
|
|
244
|
-
success: true,
|
|
245
|
-
iterations: agent.iterations,
|
|
246
|
-
total_cost: agent.total_cost.round(6),
|
|
247
|
-
cost_source: agent.cost_source.to_s,
|
|
248
|
-
duration_seconds: duration.round(2),
|
|
249
|
-
cache_creation_input_tokens: agent.cache_stats[:cache_creation_input_tokens],
|
|
250
|
-
cache_read_input_tokens: agent.cache_stats[:cache_read_input_tokens],
|
|
251
|
-
total_requests: agent.cache_stats[:total_requests],
|
|
252
|
-
cache_hit_requests: agent.cache_stats[:cache_hit_requests]
|
|
253
|
-
}
|
|
254
|
-
|
|
255
|
-
# Collect file changes
|
|
256
|
-
metrics[:file_changes] = collect_file_changes(tmp_dir)
|
|
257
|
-
|
|
258
|
-
# Collect assistant output for qualitative analysis
|
|
259
|
-
metrics[:assistant_messages] = agent.history.to_a
|
|
260
|
-
.select { |m| m[:role] == "assistant" }
|
|
261
|
-
.map { |m| extract_text(m[:content]) }
|
|
262
|
-
.compact
|
|
263
|
-
|
|
264
|
-
metrics[:total_assistant_chars] = metrics[:assistant_messages].join.length
|
|
265
|
-
|
|
266
|
-
# Cleanup
|
|
267
|
-
FileUtils.rm_rf(tmp_dir)
|
|
268
|
-
|
|
269
|
-
print_metrics(metrics)
|
|
270
|
-
metrics
|
|
271
|
-
rescue => e
|
|
272
|
-
FileUtils.rm_rf(tmp_dir) if defined?(tmp_dir) && tmp_dir
|
|
273
|
-
error_result = {
|
|
274
|
-
success: false,
|
|
275
|
-
error: e.message,
|
|
276
|
-
error_class: e.class.name,
|
|
277
|
-
iterations: defined?(agent) ? agent&.iterations : 0,
|
|
278
|
-
total_cost: defined?(agent) ? agent&.total_cost&.round(6) : 0
|
|
279
|
-
}
|
|
280
|
-
puts " ERROR: #{e.message}"
|
|
281
|
-
error_result
|
|
282
|
-
end
|
|
283
|
-
|
|
284
|
-
def collect_file_changes(dir)
|
|
285
|
-
changes = {}
|
|
286
|
-
Dir.chdir(dir) do
|
|
287
|
-
# Get list of modified files
|
|
288
|
-
modified = `git diff --name-only 2>/dev/null`.strip.split("\n").reject(&:empty?)
|
|
289
|
-
modified.each do |f|
|
|
290
|
-
next unless File.exist?(f)
|
|
291
|
-
changes[f] = File.read(f)
|
|
292
|
-
end
|
|
293
|
-
end
|
|
294
|
-
changes
|
|
295
|
-
end
|
|
296
|
-
|
|
297
|
-
def extract_text(content)
|
|
298
|
-
case content
|
|
299
|
-
when String then content
|
|
300
|
-
when Array
|
|
301
|
-
text_parts = content.select { |p| p.is_a?(Hash) && p[:type] == "text" }
|
|
302
|
-
text_parts.map { |p| p[:text] }.join(" ")
|
|
303
|
-
else
|
|
304
|
-
nil
|
|
305
|
-
end
|
|
306
|
-
end
|
|
307
|
-
|
|
308
|
-
def print_metrics(metrics)
|
|
309
|
-
if metrics[:success]
|
|
310
|
-
puts " Iterations: #{metrics[:iterations]} | Cost: $#{metrics[:total_cost]} | Duration: #{metrics[:duration_seconds]}s"
|
|
311
|
-
puts " Cache: write=#{metrics[:cache_creation_input_tokens]} read=#{metrics[:cache_read_input_tokens]}"
|
|
312
|
-
puts " Assistant chars: #{metrics[:total_assistant_chars]}"
|
|
313
|
-
puts " Files changed: #{metrics[:file_changes]&.keys&.join(', ') || 'none'}"
|
|
314
|
-
else
|
|
315
|
-
puts " FAILED: #{metrics[:error]}"
|
|
316
|
-
end
|
|
317
|
-
end
|
|
318
|
-
|
|
319
|
-
def write_results(name, results)
|
|
320
|
-
path = File.join(RESULTS_DIR, "#{name}_#{timestamp}.json")
|
|
321
|
-
File.write(path, JSON.pretty_generate(results))
|
|
322
|
-
puts "\n#{name.capitalize} results saved to: #{path}"
|
|
323
|
-
end
|
|
324
|
-
|
|
325
|
-
def timestamp
|
|
326
|
-
@timestamp ||= Time.now.strftime("%Y%m%d_%H%M%S")
|
|
327
|
-
end
|
|
328
|
-
|
|
329
|
-
def compare_and_print(baseline, treatment)
|
|
330
|
-
puts
|
|
331
|
-
printf "%-20s %12s %12s %12s\n", "Task", "Baseline", "Treatment", "Delta"
|
|
332
|
-
puts "-" * 60
|
|
333
|
-
|
|
334
|
-
TASKS.each do |task|
|
|
335
|
-
task_key = task[:name].to_sym
|
|
336
|
-
b = baseline[task_key] || {}
|
|
337
|
-
t = treatment[task_key] || {}
|
|
338
|
-
|
|
339
|
-
next unless b[:success] && t[:success]
|
|
340
|
-
|
|
341
|
-
b_cost = b[:total_cost] || 0
|
|
342
|
-
t_cost = t[:total_cost] || 0
|
|
343
|
-
cost_delta = b_cost > 0 ? "#{(t_cost / b_cost * 100).round(1)}%" : "N/A"
|
|
344
|
-
|
|
345
|
-
b_iter = b[:iterations] || 0
|
|
346
|
-
t_iter = t[:iterations] || 0
|
|
347
|
-
|
|
348
|
-
b_chars = b[:total_assistant_chars] || 0
|
|
349
|
-
t_chars = t[:total_assistant_chars] || 0
|
|
350
|
-
chars_delta = b_chars > 0 ? "#{(t_chars / b_chars.to_f * 100).round(1)}%" : "N/A"
|
|
351
|
-
|
|
352
|
-
printf "%-20s\n", task[:name]
|
|
353
|
-
printf " Cost: $%-10.6f $%-10.6f %s\n", b_cost, t_cost, cost_delta
|
|
354
|
-
printf " Iterations: %-11d %-11d %s\n", b_iter, t_iter, "#{t_iter - b_iter > 0 ? '+' : ''}#{t_iter - b_iter}"
|
|
355
|
-
printf " Assistant chars: %-11d %-11d %s\n", b_chars, t_chars, chars_delta
|
|
356
|
-
puts
|
|
357
|
-
end
|
|
358
|
-
|
|
359
|
-
# Totals
|
|
360
|
-
b_total_cost = 0
|
|
361
|
-
t_total_cost = 0
|
|
362
|
-
b_total_iter = 0
|
|
363
|
-
t_total_iter = 0
|
|
364
|
-
b_total_chars = 0
|
|
365
|
-
t_total_chars = 0
|
|
366
|
-
|
|
367
|
-
TASKS.each do |task|
|
|
368
|
-
task_key = task[:name].to_sym
|
|
369
|
-
b = baseline[task_key] || {}
|
|
370
|
-
t = treatment[task_key] || {}
|
|
371
|
-
next unless b[:success] && t[:success]
|
|
372
|
-
|
|
373
|
-
b_total_cost += b[:total_cost] || 0
|
|
374
|
-
t_total_cost += t[:total_cost] || 0
|
|
375
|
-
b_total_iter += b[:iterations] || 0
|
|
376
|
-
t_total_iter += t[:iterations] || 0
|
|
377
|
-
b_total_chars += b[:total_assistant_chars] || 0
|
|
378
|
-
t_total_chars += t[:total_assistant_chars] || 0
|
|
379
|
-
end
|
|
380
|
-
|
|
381
|
-
puts "-" * 60
|
|
382
|
-
printf "%-20s\n", "TOTALS"
|
|
383
|
-
cost_pct = b_total_cost > 0 ? (t_total_cost / b_total_cost * 100).round(1) : 0
|
|
384
|
-
printf " Total cost: $%-10.6f $%-10.6f %s%%\n", b_total_cost, t_total_cost, cost_pct
|
|
385
|
-
printf " Total iterations: %-11d %-11d %+d\n", b_total_iter, t_total_iter, t_total_iter - b_total_iter
|
|
386
|
-
chars_pct = b_total_chars > 0 ? (t_total_chars / b_total_chars.to_f * 100).round(1) : 0
|
|
387
|
-
printf " Total chars: %-11d %-11d %s%%\n", b_total_chars, t_total_chars, chars_pct
|
|
388
|
-
end
|
|
389
|
-
|
|
390
|
-
# Minimal UI that captures output without displaying
|
|
391
|
-
class BenchmarkUI
|
|
392
|
-
def log(msg, level: :info); end
|
|
393
|
-
def show_assistant_message(content, files: []); end
|
|
394
|
-
def show_tool_call(name, args); end
|
|
395
|
-
def show_tool_result(result); end
|
|
396
|
-
def show_tool_stdout(lines); end
|
|
397
|
-
def show_tool_error(error); end
|
|
398
|
-
def show_tool_args(formatted_args); end
|
|
399
|
-
def show_file_write_preview(path, is_new_file:); end
|
|
400
|
-
def show_file_edit_preview(path); end
|
|
401
|
-
def show_file_error(error_message); end
|
|
402
|
-
def show_shell_preview(command); end
|
|
403
|
-
def show_diff(old_content, new_content, max_lines: 50); end
|
|
404
|
-
def show_token_usage(token_data); end
|
|
405
|
-
def show_complete(iterations:, cost:, duration: nil, cache_stats: nil, awaiting_user_feedback: false, cost_source: nil); end
|
|
406
|
-
def append_output(content); end
|
|
407
|
-
def show_info(message, prefix_newline: true); end
|
|
408
|
-
def show_warning(message); end
|
|
409
|
-
def show_error(message); end
|
|
410
|
-
def show_success(message); end
|
|
411
|
-
def show_progress(message = nil, prefix_newline: true, progress_type: "thinking", phase: "active", metadata: {}); end
|
|
412
|
-
def start_progress(message: nil, style: :primary, quiet_on_fast_finish: false); end
|
|
413
|
-
def with_progress(message: nil, style: :primary, quiet_on_fast_finish: false)
|
|
414
|
-
yield if block_given?
|
|
415
|
-
end
|
|
416
|
-
def update_sessionbar(tasks: nil, cost: nil, cost_source: nil, status: nil, latency: nil); end
|
|
417
|
-
def update_todos(todos); end
|
|
418
|
-
def set_working_status; end
|
|
419
|
-
def set_idle_status; end
|
|
420
|
-
def request_confirmation(message, default: true); end
|
|
421
|
-
def clear_input; end
|
|
422
|
-
def set_input_tips(message, type: :info); end
|
|
423
|
-
def stop; end
|
|
424
|
-
end
|
|
425
|
-
end
|
|
426
|
-
|
|
427
|
-
if __FILE__ == $0
|
|
428
|
-
variant = ARGV[0]&.downcase
|
|
429
|
-
runner = BenchmarkRunner.new
|
|
430
|
-
|
|
431
|
-
case variant
|
|
432
|
-
when "baseline"
|
|
433
|
-
runner.run_baseline
|
|
434
|
-
when "treatment"
|
|
435
|
-
runner.run_treatment
|
|
436
|
-
when "report"
|
|
437
|
-
runner.run_report
|
|
438
|
-
else
|
|
439
|
-
runner.run
|
|
440
|
-
end
|
|
441
|
-
end
|
|
@@ -1,168 +0,0 @@
|
|
|
1
|
-
# frozen_string_literal: true
|
|
2
|
-
|
|
3
|
-
require "fileutils"
|
|
4
|
-
|
|
5
|
-
module Clacky
|
|
6
|
-
# Discovers and loads user-defined HTTP API extensions from
|
|
7
|
-
# ~/.clacky/api_ext/<name>/handler.rb. Each handler is expected to define a
|
|
8
|
-
# subclass of Clacky::ApiExtension; the subclass is auto-registered with the
|
|
9
|
-
# framework and its routes become available under /api/ext/<name>/.
|
|
10
|
-
#
|
|
11
|
-
# A broken extension (syntax error, missing base class, route conflict) is
|
|
12
|
-
# isolated: skipped with a logged warning, never aborts the load of others.
|
|
13
|
-
module ApiExtensionLoader
|
|
14
|
-
DEFAULT_DIR = File.expand_path("~/.clacky/api_ext")
|
|
15
|
-
BUILTIN_DIR = File.expand_path("../default_extensions", __FILE__)
|
|
16
|
-
DISABLED_DIR = "_disabled"
|
|
17
|
-
|
|
18
|
-
Result = Struct.new(:loaded, :skipped, keyword_init: true)
|
|
19
|
-
|
|
20
|
-
class << self
|
|
21
|
-
def load_all(dir: DEFAULT_DIR, builtin: true)
|
|
22
|
-
result = Result.new(loaded: [], skipped: [])
|
|
23
|
-
Clacky::ApiExtension.reset_registry!
|
|
24
|
-
|
|
25
|
-
# Load built-in (gem-shipped) extensions first (lowest priority)
|
|
26
|
-
if builtin && Dir.exist?(BUILTIN_DIR)
|
|
27
|
-
Dir.glob(File.join(BUILTIN_DIR, "*", "handler.rb")).sort.each do |handler_path|
|
|
28
|
-
ext_dir = File.dirname(handler_path)
|
|
29
|
-
ext_id = File.basename(ext_dir)
|
|
30
|
-
next if ext_id.start_with?("_")
|
|
31
|
-
load_one(ext_id, ext_dir, handler_path, result)
|
|
32
|
-
end
|
|
33
|
-
end
|
|
34
|
-
|
|
35
|
-
# Load user extensions (higher priority — same ext_id overwrites built-in)
|
|
36
|
-
if Dir.exist?(dir)
|
|
37
|
-
Dir.glob(File.join(dir, "*", "handler.rb")).sort.each do |handler_path|
|
|
38
|
-
ext_dir = File.dirname(handler_path)
|
|
39
|
-
ext_id = File.basename(ext_dir)
|
|
40
|
-
next if ext_id == DISABLED_DIR || ext_id.start_with?("_")
|
|
41
|
-
load_one(ext_id, ext_dir, handler_path, result)
|
|
42
|
-
end
|
|
43
|
-
end
|
|
44
|
-
|
|
45
|
-
@last_result = result
|
|
46
|
-
log_summary(result)
|
|
47
|
-
result
|
|
48
|
-
end
|
|
49
|
-
|
|
50
|
-
def last_result
|
|
51
|
-
@last_result || load_all
|
|
52
|
-
end
|
|
53
|
-
|
|
54
|
-
def load_one(ext_id, ext_dir, handler_path, result)
|
|
55
|
-
meta = read_meta(ext_dir)
|
|
56
|
-
before = Clacky::ApiExtension.pending_subclasses.size
|
|
57
|
-
|
|
58
|
-
require handler_path
|
|
59
|
-
|
|
60
|
-
new_subclasses = Clacky::ApiExtension.pending_subclasses[before..] || []
|
|
61
|
-
klass = new_subclasses.last
|
|
62
|
-
|
|
63
|
-
unless klass
|
|
64
|
-
result.skipped << [ext_id, "no Clacky::ApiExtension subclass defined in handler.rb"]
|
|
65
|
-
log_skip(ext_id, result.skipped.last[1])
|
|
66
|
-
return
|
|
67
|
-
end
|
|
68
|
-
|
|
69
|
-
klass.ext_id = ext_id
|
|
70
|
-
klass.ext_dir = ext_dir
|
|
71
|
-
klass.meta = meta
|
|
72
|
-
|
|
73
|
-
if klass.routes.empty?
|
|
74
|
-
result.skipped << [ext_id, "no routes declared (use get/post/... DSL)"]
|
|
75
|
-
log_skip(ext_id, result.skipped.last[1])
|
|
76
|
-
Clacky::ApiExtension.registry.delete(ext_id)
|
|
77
|
-
return
|
|
78
|
-
end
|
|
79
|
-
|
|
80
|
-
if (gap = validate_public_endpoints(klass, meta))
|
|
81
|
-
result.skipped << [ext_id, gap]
|
|
82
|
-
log_skip(ext_id, gap)
|
|
83
|
-
return
|
|
84
|
-
end
|
|
85
|
-
|
|
86
|
-
Clacky::ApiExtension.register(ext_id, klass)
|
|
87
|
-
result.loaded << ext_id
|
|
88
|
-
public_count = klass.public_paths.size
|
|
89
|
-
suffix = public_count > 0 ? " (#{public_count} public)" : ""
|
|
90
|
-
Clacky::Logger.info("[ApiExtensionLoader] Loaded '#{ext_id}' — #{klass.routes.size} route(s)#{suffix}")
|
|
91
|
-
rescue StandardError, ScriptError => e
|
|
92
|
-
result.skipped << [ext_id, e.message]
|
|
93
|
-
log_skip(ext_id, e.message)
|
|
94
|
-
end
|
|
95
|
-
|
|
96
|
-
private def read_meta(ext_dir)
|
|
97
|
-
path = File.join(ext_dir, "meta.yml")
|
|
98
|
-
return {} unless File.exist?(path)
|
|
99
|
-
|
|
100
|
-
YAMLCompat.load_file(path) || {}
|
|
101
|
-
rescue StandardError => e
|
|
102
|
-
Clacky::Logger.warn("[ApiExtensionLoader] Failed to read meta.yml in #{ext_dir}: #{e.message}")
|
|
103
|
-
{}
|
|
104
|
-
end
|
|
105
|
-
|
|
106
|
-
private def validate_public_endpoints(klass, meta)
|
|
107
|
-
return nil if klass.public_paths.empty?
|
|
108
|
-
return nil if meta["public"] == true
|
|
109
|
-
|
|
110
|
-
"uses public_endpoint but meta.yml is missing 'public: true'"
|
|
111
|
-
end
|
|
112
|
-
|
|
113
|
-
private def log_skip(ext_id, reason)
|
|
114
|
-
Clacky::Logger.warn("[ApiExtensionLoader] Skipped '#{ext_id}': #{reason}")
|
|
115
|
-
end
|
|
116
|
-
|
|
117
|
-
private def log_summary(result)
|
|
118
|
-
return if result.loaded.empty? && result.skipped.empty?
|
|
119
|
-
|
|
120
|
-
total_routes = result.loaded.sum { |id| Clacky::ApiExtension.registry[id]&.routes&.size || 0 }
|
|
121
|
-
Clacky::Logger.info("[ApiExtensionLoader] #{result.loaded.size} extension(s), #{total_routes} route(s); #{result.skipped.size} skipped")
|
|
122
|
-
end
|
|
123
|
-
|
|
124
|
-
# Generate a starter handler.rb at ~/.clacky/api_ext/<name>/handler.rb.
|
|
125
|
-
# Returns the path to the generated file.
|
|
126
|
-
def scaffold(name, dir: DEFAULT_DIR)
|
|
127
|
-
slug = name.to_s.strip.downcase.gsub(/[^a-z0-9_-]+/, "-").gsub(/\A-+|-+\z/, "")
|
|
128
|
-
raise ArgumentError, "invalid api_ext name: #{name.inspect}" if slug.empty?
|
|
129
|
-
|
|
130
|
-
target_dir = File.join(dir, slug)
|
|
131
|
-
path = File.join(target_dir, "handler.rb")
|
|
132
|
-
raise ArgumentError, "api_ext already exists: #{path}" if File.exist?(path)
|
|
133
|
-
|
|
134
|
-
FileUtils.mkdir_p(target_dir)
|
|
135
|
-
File.write(path, skeleton(slug))
|
|
136
|
-
path
|
|
137
|
-
end
|
|
138
|
-
|
|
139
|
-
private def skeleton(slug)
|
|
140
|
-
const = slug.split(/[-_]/).map(&:capitalize).join + "Ext"
|
|
141
|
-
<<~RUBY
|
|
142
|
-
# frozen_string_literal: true
|
|
143
|
-
|
|
144
|
-
# Custom HTTP API extension mounted at /api/ext/#{slug}/
|
|
145
|
-
# Scaffolded by `clacky api_ext_new #{slug}` — fill in the routes you need.
|
|
146
|
-
class #{const} < Clacky::ApiExtension
|
|
147
|
-
get "/hello" do
|
|
148
|
-
json(message: "hello from #{slug}")
|
|
149
|
-
end
|
|
150
|
-
|
|
151
|
-
# Examples — uncomment and adapt:
|
|
152
|
-
#
|
|
153
|
-
# post "/items" do
|
|
154
|
-
# body = json_body
|
|
155
|
-
# error!("name required", status: 422) unless body["name"]
|
|
156
|
-
# File.write(data_path("items.json"), body.to_json)
|
|
157
|
-
# json(ok: true)
|
|
158
|
-
# end
|
|
159
|
-
#
|
|
160
|
-
# get "/items/:id" do
|
|
161
|
-
# json(id: params[:id])
|
|
162
|
-
# end
|
|
163
|
-
end
|
|
164
|
-
RUBY
|
|
165
|
-
end
|
|
166
|
-
end
|
|
167
|
-
end
|
|
168
|
-
end
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
(No user profile configured yet. To personalize, create ~/.clacky/agents/USER.md)
|
|
File without changes
|