prodigy 0.4.4

Turn ad-hoc Claude sessions into reproducible development pipelines with parallel AI agents
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
//! Performance benchmarks for MapReduce phase modules
//!
//! These benchmarks verify that the phase-based architecture maintains
//! performance requirements specified in Spec 131:
//! - < 2% performance regression vs original implementation
//! - Setup phase execution overhead
//! - Map phase scaling characteristics
//! - Reduce phase aggregation performance

use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion};
use prodigy::cook::execution::mapreduce::phases::{
    coordinator::PhaseCoordinator, PhaseContext, PhaseExecutor,
};
use prodigy::cook::execution::mapreduce::{MapPhase, ReducePhase};
use prodigy::cook::execution::SetupPhase;
use prodigy::cook::orchestrator::ExecutionEnvironment;
use prodigy::cook::workflow::WorkflowStep;
use prodigy::subprocess::SubprocessManager;
use std::collections::HashMap;
use std::sync::Arc;
use std::time::Duration;
use tempfile::TempDir;
use tokio::runtime::Runtime;

/// Helper to create a test execution environment
fn create_test_env(temp_dir: &TempDir) -> ExecutionEnvironment {
    ExecutionEnvironment {
        working_dir: Arc::new(temp_dir.path().to_path_buf()),
        project_dir: Arc::new(temp_dir.path().to_path_buf()),
        worktree_name: Some(Arc::from("bench-worktree")),
        session_id: Arc::from("bench-session"),
    }
}

/// Helper to create a subprocess manager
fn create_subprocess_manager() -> Arc<SubprocessManager> {
    Arc::new(SubprocessManager::production())
}

/// Benchmark setup phase execution with varying numbers of commands
fn bench_setup_phase_scaling(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_setup");
    group.measurement_time(Duration::from_secs(10));

    for num_commands in &[1, 5, 10, 20] {
        group.bench_with_input(
            BenchmarkId::new("execute_commands", num_commands),
            num_commands,
            |b, &count| {
                b.to_async(&rt).iter(|| async move {
                    let temp_dir = TempDir::new().unwrap();
                    let env = create_test_env(&temp_dir);
                    let subprocess = create_subprocess_manager();

                    let commands: Vec<_> = (0..count)
                        .map(|i| WorkflowStep {
                            shell: Some(format!("echo 'command {}' > /dev/null", i)),
                            ..Default::default()
                        })
                        .collect();

                    let setup_phase = SetupPhase {
                        commands,
                        timeout: Some(60),
                        capture_outputs: HashMap::new(),
                    };

                    let executor =
                        prodigy::cook::execution::mapreduce::phases::setup::SetupPhaseExecutor::new(
                            setup_phase,
                        );

                    let mut context = PhaseContext::new(env, subprocess);

                    let _ = executor.execute(&mut context).await;
                });
            },
        );
    }

    group.finish();
}

/// Benchmark map phase work item processing
fn bench_map_phase_work_items(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_map");
    group.measurement_time(Duration::from_secs(15));

    for num_items in &[10, 50, 100] {
        group.bench_with_input(
            BenchmarkId::new("process_items", num_items),
            num_items,
            |b, &count| {
                b.to_async(&rt).iter(|| async move {
                    let temp_dir = TempDir::new().unwrap();
                    let env = create_test_env(&temp_dir);
                    let subprocess = create_subprocess_manager();

                    // Create a JSON input file
                    let input_file = temp_dir.path().join("items.json");
                    let items: Vec<_> = (0..count)
                        .map(|i| {
                            serde_json::json!({
                                "id": i,
                                "name": format!("item_{}", i),
                                "data": format!("data_{}", i)
                            })
                        })
                        .collect();
                    std::fs::write(&input_file, serde_json::to_string(&items).unwrap()).unwrap();

                    let map_phase = MapPhase {
                        config: prodigy::cook::execution::mapreduce::MapReduceConfig {
                            input: input_file.to_string_lossy().to_string(),
                            max_parallel: 5,
                            ..Default::default()
                        },
                        agent_template: vec![WorkflowStep {
                            shell: Some("echo '${item.name}' > /dev/null".to_string()),
                            ..Default::default()
                        }],
                        json_path: None,
                        filter: None,
                        sort_by: None,
                        max_items: Some(count),
                        distinct: None,
                        timeout_config: None,
                        workflow_env: std::collections::HashMap::new(),
                    };

                    let executor =
                        prodigy::cook::execution::mapreduce::phases::map::MapPhaseExecutor::new(
                            map_phase,
                        );

                    let context = PhaseContext::new(env, subprocess);

                    // Note: Map executor may not be fully functional in new architecture
                    // This measures the pure planning/coordination overhead
                    let _ = executor.validate_context(&context);
                });
            },
        );
    }

    group.finish();
}

/// Benchmark reduce phase aggregation performance
fn bench_reduce_phase_aggregation(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_reduce");
    group.measurement_time(Duration::from_secs(10));

    for num_commands in &[1, 5, 10] {
        group.bench_with_input(
            BenchmarkId::new("aggregate_results", num_commands),
            num_commands,
            |b, &count| {
                b.to_async(&rt).iter(|| async move {
                    let temp_dir = TempDir::new().unwrap();
                    let env = create_test_env(&temp_dir);
                    let subprocess = create_subprocess_manager();

                    let commands: Vec<_> = (0..count)
                        .map(|i| WorkflowStep {
                            shell: Some(format!("echo 'aggregate {}' > /dev/null", i)),
                            ..Default::default()
                        })
                        .collect();

                    let reduce_phase = ReducePhase {
                        commands,
                        timeout_secs: Some(60),
                    };

                    let executor =
                        prodigy::cook::execution::mapreduce::phases::reduce::ReducePhaseExecutor::new(
                            reduce_phase,
                        );

                    let mut context = PhaseContext::new(env, subprocess);

                    let _ = executor.execute(&mut context).await;
                });
            },
        );
    }

    group.finish();
}

/// Benchmark full workflow execution (Setup -> Map -> Reduce)
fn bench_full_workflow_execution(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_workflow");
    group.measurement_time(Duration::from_secs(20));

    group.bench_function("complete_workflow", |b| {
        b.to_async(&rt).iter(|| async move {
            let temp_dir = TempDir::new().unwrap();
            let env = create_test_env(&temp_dir);
            let subprocess = create_subprocess_manager();

            // Setup phase: create test data
            let setup_phase = SetupPhase {
                commands: vec![
                    WorkflowStep {
                        shell: Some("echo 'setup 1' > /dev/null".to_string()),
                        ..Default::default()
                    },
                    WorkflowStep {
                        shell: Some("echo 'setup 2' > /dev/null".to_string()),
                        ..Default::default()
                    },
                ],
                timeout: Some(30),
                capture_outputs: HashMap::new(),
            };

            // Map phase: minimal processing
            let map_phase = MapPhase {
                config: prodigy::cook::execution::mapreduce::MapReduceConfig {
                    input: "[]".to_string(),
                    max_parallel: 1,
                    ..Default::default()
                },
                agent_template: vec![],
                json_path: None,
                filter: None,
                sort_by: None,
                max_items: None,
                distinct: None,
                timeout_config: None,
                workflow_env: std::collections::HashMap::new(),
            };

            // Reduce phase: aggregate results
            let reduce_phase = ReducePhase {
                commands: vec![WorkflowStep {
                    shell: Some("echo 'reduce' > /dev/null".to_string()),
                    ..Default::default()
                }],
                timeout_secs: Some(30),
            };

            let coordinator = PhaseCoordinator::new(
                Some(setup_phase),
                map_phase,
                Some(reduce_phase),
                subprocess.clone(),
            );

            let _ = coordinator.execute_workflow(env, subprocess).await;
        });
    });

    group.finish();
}

/// Benchmark phase context creation and initialization
fn bench_phase_context_creation(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();

    c.benchmark_group("phase_context")
        .bench_function("create_context", |b| {
            b.to_async(&rt).iter(|| async move {
                let temp_dir = TempDir::new().unwrap();
                let env = create_test_env(&temp_dir);
                let subprocess = create_subprocess_manager();

                let _context = PhaseContext::new(env, subprocess);
            });
        });
}

/// Benchmark phase transition logic (pure planning)
fn bench_phase_transitions(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_transitions");

    group.bench_function("transition_overhead", |b| {
        b.to_async(&rt).iter(|| async move {
            let temp_dir = TempDir::new().unwrap();
            let env = create_test_env(&temp_dir);
            let subprocess = create_subprocess_manager();

            // Minimal workflow to measure transition overhead
            let map_phase = MapPhase {
                config: prodigy::cook::execution::mapreduce::MapReduceConfig {
                    input: "[]".to_string(),
                    max_parallel: 1,
                    ..Default::default()
                },
                agent_template: vec![],
                json_path: None,
                filter: None,
                sort_by: None,
                max_items: None,
                distinct: None,
                timeout_config: None,
                workflow_env: std::collections::HashMap::new(),
            };

            let coordinator = PhaseCoordinator::new(None, map_phase, None, subprocess.clone());

            // This measures pure coordination overhead without actual work
            let _ = coordinator.execute_workflow(env, subprocess).await;
        });
    });

    group.finish();
}

/// Benchmark parallel phase scaling (multiple phases in sequence)
fn bench_phase_scaling(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();
    let mut group = c.benchmark_group("phase_scaling");
    group.measurement_time(Duration::from_secs(15));

    for parallelism in &[1, 2, 5, 10] {
        group.bench_with_input(
            BenchmarkId::new("parallel_agents", parallelism),
            parallelism,
            |b, &max_parallel| {
                b.to_async(&rt).iter(|| async move {
                    let temp_dir = TempDir::new().unwrap();
                    let env = create_test_env(&temp_dir);
                    let subprocess = create_subprocess_manager();

                    let map_phase = MapPhase {
                        config: prodigy::cook::execution::mapreduce::MapReduceConfig {
                            input: "[]".to_string(),
                            max_parallel,
                            ..Default::default()
                        },
                        agent_template: vec![],
                        json_path: None,
                        filter: None,
                        sort_by: None,
                        max_items: None,
                        distinct: None,
                        timeout_config: None,
                        workflow_env: std::collections::HashMap::new(),
                    };

                    let coordinator =
                        PhaseCoordinator::new(None, map_phase, None, subprocess.clone());

                    let _ = coordinator.execute_workflow(env, subprocess).await;
                });
            },
        );
    }

    group.finish();
}

/// Benchmark phase executor trait overhead
fn bench_executor_trait_overhead(c: &mut Criterion) {
    let rt = Runtime::new().unwrap();

    c.benchmark_group("phase_executor")
        .bench_function("trait_dispatch", |b| {
            b.to_async(&rt).iter(|| async move {
                let temp_dir = TempDir::new().unwrap();
                let env = create_test_env(&temp_dir);
                let subprocess = create_subprocess_manager();

                let setup_phase = SetupPhase {
                    commands: vec![WorkflowStep {
                        shell: Some("true".to_string()),
                        ..Default::default()
                    }],
                    timeout: Some(30),
                    capture_outputs: HashMap::new(),
                };

                let executor =
                    prodigy::cook::execution::mapreduce::phases::setup::SetupPhaseExecutor::new(
                        setup_phase,
                    );

                // Measure trait method dispatch overhead
                let context = PhaseContext::new(env, subprocess);
                let _ = executor.phase_type();
                let _ = executor.can_skip(&context);
                let _ = executor.validate_context(&context);
            });
        });
}

criterion_group!(
    benches,
    bench_setup_phase_scaling,
    bench_map_phase_work_items,
    bench_reduce_phase_aggregation,
    bench_full_workflow_execution,
    bench_phase_context_creation,
    bench_phase_transitions,
    bench_phase_scaling,
    bench_executor_trait_overhead
);

criterion_main!(benches);