File size: 12,359 Bytes
b562f4c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#!/usr/bin/env python3
"""Trained Router v3: Combined best approach.
Uses per-tier P(success) classifiers + safety floors + 
cost-aware routing with ASYMMETRIC penalties (underkill penalized 5x harder than overkill).
"""
import json, os, sys, random, pickle, uuid
import numpy as np
from datetime import datetime
from collections import defaultdict
from typing import Dict, List, Tuple, Any, Optional

TASK_TYPES = ["quick_answer","coding","research","document_drafting",
              "legal_regulated","tool_heavy","retrieval_heavy",
              "long_horizon","unknown_ambiguous"]
TT2IDX = {t:i for i,t in enumerate(TASK_TYPES)}

CODE_KW = ["python","javascript","code","function","bug","debug","refactor",
           "implement","test","compile","runtime","class","module","async","thread"]
LEGAL_KW = ["contract","legal","compliance","gdpr","privacy","policy","regulatory","liability"]
RESEARCH_KW = ["research","find sources","literature","investigate","compare","analyze","survey"]
TOOL_KW = ["search","fetch","retrieve","query","api","database","scrape","aggregate"]
LONG_KW = ["plan","project","roadmap","orchestrate","multi-step","migrate","pipeline","deploy"]
MATH_KW = ["calculate","compute","solve","equation","formula","optimize","probability"]

TIER_STR = {1:0.35,2:0.55,3:0.80,4:0.93,5:0.97}
TIER_COST = {1:0.05,2:0.15,3:0.75,4:1.0,5:1.5}
TASK_FLOOR = {"legal_regulated":4,"long_horizon":3,"research":3,"coding":3,
              "unknown_ambiguous":3,"quick_answer":1,"document_drafting":2,
              "tool_heavy":2,"retrieval_heavy":2}

TASK_TEMPLATES = {
    "quick_answer":["What is the capital of France?","Explain quantum computing briefly.",
        "What is 237*452?","Define photosynthesis.","Who wrote Hamlet?",
        "What is the speed of light?","List the primary colors.","What is GDP?"],
    "coding":["Write a Python function to reverse a linked list.",
        "Fix the bug in this React component.","Refactor auth module to JWT.",
        "Implement LRU cache in Go.","Debug segfault in C++ thread pool.",
        "Add unit tests for the payment module.","Optimize this SQL query.",
        "Create a REST API for user management.","Implement binary search in Rust."],
    "research":["Research latest transformer advances.",
        "Find sources comparing LoRA and full FT.",
        "Investigate data center climate impact.",
        "Survey privacy-preserving ML techniques.",
        "Compare reinforcement learning algorithms for robotics."],
    "document_drafting":["Draft project proposal for ML pipeline.",
        "Write email to team about deployment.","Create technical report on performance."],
    "legal_regulated":["Review this contract for liability clauses.",
        "Check GDPR compliance for data pipeline.","Draft privacy policy section.",
        "Verify regulatory compliance for medical device software."],
    "tool_heavy":["Search open issues and create summary.",
        "Fetch API docs and generate client code.","Query Q3 sales and produce chart."],
    "retrieval_heavy":["Answer based on 50-page document.",
        "Find all payment processing mentions.","Retrieve relevant cases for legal query."],
    "long_horizon":["Plan 3-month roadmap.","Orchestrate multi-region deployment.",
        "Redesign data architecture end-to-end.","Migrate monolith to microservices."],
    "unknown_ambiguous":["Help me with this thing.",
        "I need something about the server.","Can you look into that issue?"],
}

def tsp(tier, diff):
    return TIER_STR[tier] ** (diff * 0.6)

def extract_features(request, task_type, difficulty=3):
    r = request.lower()
    f = {
        "req_len": len(request), "num_words": len(request.split()),
        "has_code": int(any(k in r for k in CODE_KW)),
        "n_code": sum(1 for k in CODE_KW if k in r),
        "has_legal": int(any(k in r for k in LEGAL_KW)),
        "n_legal": sum(1 for k in LEGAL_KW if k in r),
        "has_research": int(any(k in r for k in RESEARCH_KW)),
        "n_research": sum(1 for k in RESEARCH_KW if k in r),
        "has_tool": int(any(k in r for k in TOOL_KW)),
        "n_tool": sum(1 for k in TOOL_KW if k in r),
        "has_long": int(any(k in r for k in LONG_KW)),
        "has_math": int(any(k in r for k in MATH_KW)),
        "tt_idx": TT2IDX.get(task_type, 8), "difficulty": difficulty,
    }
    for tt in TASK_TYPES:
        f[f"tt_{tt}"] = int(task_type == tt)
    return f

def gen_trace(idx, rng):
    tt = rng.choice(list(TASK_TEMPLATES.keys()))
    diff = {"quick_answer":1,"document_drafting":2,"tool_heavy":2,"retrieval_heavy":2,
            "research":3,"coding":3,"unknown_ambiguous":3,"long_horizon":4,"legal_regulated":5}[tt]
    tier_out = {}
    for t in range(1,6):
        tier_out[t] = rng.random() < tsp(t, diff)
    opt = 5
    for t in range(1,6):
        if tier_out[t]: opt = t; break
    if diff <= 2: actual = rng.choices([1,2,3,4,5],weights=[3,4,2,1,0.5])[0]
    elif diff == 3: actual = rng.choices([1,2,3,4,5],weights=[1,2,4,2,1])[0]
    elif diff == 4: actual = rng.choices([1,2,3,4,5],weights=[0.5,1,2,4,2])[0]
    else: actual = rng.choices([1,2,3,4,5],weights=[0.2,0.5,1,3,4])[0]
    outcome = "success" if tier_out[actual] else "failure"
    req = rng.choice(TASK_TEMPLATES[tt])
    feats = extract_features(req, tt, diff)
    return {"feats":feats,"opt":opt,"actual":actual,"outcome":outcome,
            "tier_out":tier_out,"tt":tt,"diff":diff,"req":req}

print("="*80)
print("AGENT COST OPTIMIZER - TRAINED ROUTER v3")
print("Asymmetric cost scoring: underkill 5x penalty")
print("="*80)

# ─── Generate ────────────────────────────────────────────────────────
print("\n[1] Generating 50K training traces...")
rng = random.Random(42)
traces = [gen_trace(i, rng) for i in range(50000)]
FEAT_KEYS = sorted(traces[0]["feats"].keys())
def f2v(feats):
    return np.array([float(feats.get(k, 0.0)) for k in FEAT_KEYS], dtype=np.float32)

X_all = np.array([f2v(t["feats"]) for t in traces])
y_opt = np.array([t["opt"] for t in traces])

per_tier_labels = {}
for tier in range(1,6):
    per_tier_labels[tier] = np.array([1 if t["tier_out"].get(tier,False) else 0 for t in traces])

from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, f1_score
import xgboost as xgb

X_train, X_test, idx_train, idx_test = train_test_split(X_all, range(len(traces)), test_size=0.2, random_state=42, stratify=y_opt)
print(f"  Train: {len(X_train)}, Test: {len(X_test)}")

# ─── Train Per-Tier Classifiers ─────────────────────────────────────
print("\n[2] Training per-tier P(success) classifiers...")
tier_clfs = {}
for tier in range(1,6):
    y_tr = per_tier_labels[tier][idx_train]
    neg = (y_tr==0).sum(); pos = (y_tr==1).sum()
    spw = neg / max(pos,1)
    clf = xgb.XGBClassifier(n_estimators=150, max_depth=5, learning_rate=0.1,
        subsample=0.8, colsample_bytree=0.8, scale_pos_weight=min(spw,5.0),
        objective="binary:logistic", eval_metric="logloss", random_state=42, verbosity=0)
    clf.fit(X_train, y_tr)
    y_pred = clf.predict(X_test)
    acc = accuracy_score(per_tier_labels[tier][idx_test], y_pred)
    f1 = f1_score(per_tier_labels[tier][idx_test], y_pred, zero_division=0)
    tier_clfs[tier] = clf
    print(f"  Tier {tier}: acc={acc:.3f}, f1={f1:.3f}")

# ─── Asymmetric Cost Router ──────────────────────────────────────────
print("\n[3] Building asymmetric cost router...")

def route_asymmetric(x, task_type, tier_clfs, underkill_penalty=5.0, overkill_penalty=1.0):
    """Score each tier with asymmetric penalties.
    
    score(tier) = P(failure@tier) * underkill_penalty * cost_of_failure
                + cost(tier) * overkill_penalty
    
    Underkill (routing too low) is penalized 5x more than overkill.
    """
    if x.ndim == 1:
        x = x.reshape(1, -1)
    floor = TASK_FLOOR.get(task_type, 2)

    best_tier = floor
    best_score = float("inf")

    for tier in range(floor, 6):
        p_fail = 1.0 - tier_clfs[tier].predict_proba(x)[0, 1]
        cost_norm = TIER_COST[tier] / TIER_COST[5]  # [0.03, 1.0]
        
        # Expected cost of failure (cheap model on hard task)
        failure_cost = p_fail * underkill_penalty
        
        # Cost of using this tier (overkill penalty)
        tier_cost = cost_norm * overkill_penalty
        
        score = failure_cost + tier_cost
        
        if score < best_score:
            best_score = score
            best_tier = tier

    return best_tier

# ─── Evaluate ─────────────────────────────────────────────────────────
print("\n[4] Evaluating all routers...")
n_test = len(idx_test)
results = {}

def eval_router(name, route_fn):
    succ = 0; cost = 0.0; unsafe = 0; fd = 0
    td = defaultdict(int)
    for i in idx_test:
        t = traces[i]
        x = f2v(t["feats"]).reshape(1,-1)
        pred = route_fn(x, t)
        td[pred] += 1
        if t["tier_out"].get(pred, False): succ += 1
        elif pred < t["opt"]: unsafe += 1
        elif pred >= t["opt"]: fd += 1
        cost += TIER_COST[pred]
    results[name] = {"success":succ/n_test, "avg_cost":cost/n_test,
                     "unsafe_rate":unsafe/n_test, "false_done":fd/n_test,
                     "tier_dist":dict(td)}

eval_router("always_frontier", lambda x,t: 4)
eval_router("always_cheap", lambda x,t: 1)
eval_router("heuristic_diff+1", lambda x,t: min(t["diff"]+1,5))
eval_router("heuristic_floor", lambda x,t: TASK_FLOOR.get(t["tt"],2))

for ukp in [3.0, 5.0, 8.0, 10.0, 15.0]:
    eval_router(f"asymmetric_uk{ukp:.0f}", lambda x,t,uk=ukp: route_asymmetric(x, t["tt"], tier_clfs, underkill_penalty=uk))

eval_router("oracle", lambda x,t: t["opt"])

# Print comparison
print(f"\n{'Router':<25} {'Success':>10} {'AvgCost':>10} {'CostRed':>10} {'Unsafe':>10} {'F-DONE':>10}")
print("-"*75)
fc = results["always_frontier"]["avg_cost"]
for name, r in sorted(results.items(), key=lambda x: (-x[1]["success"], x[1]["avg_cost"])):
    cr = (1 - r["avg_cost"]/fc)*100
    print(f"{name:<25} {r['success']:>10.3f} {r['avg_cost']:>10.4f} {cr:>9.1f}% {r['unsafe_rate']:>10.3f} {r['false_done']:>10.3f}")

# Pareto frontier
print("\nPARETO FRONTIER:")
pareto = []
for name, r in results.items():
    if name == "always_cheap": continue
    dominated = False
    for name2, r2 in results.items():
        if name == name2: continue
        if r2["success"] >= r["success"] and r2["avg_cost"] <= r["avg_cost"]:
            if r2["success"] > r["success"] or r2["avg_cost"] < r["avg_cost"]:
                dominated = True; break
    if not dominated:
        pareto.append((name, r))
        cr = (1 - r["avg_cost"]/fc)*100
        print(f"  {name:<25} success={r['success']:.3f} cost={r['avg_cost']:.4f} costRed={cr:.1f}% unsafe={r['unsafe_rate']:.3f}")

# ─── Save Best Model ─────────────────────────────────────────────────
print("\n[5] Saving models...")
os.makedirs("/app/router_models", exist_ok=True)
for tier, clf in tier_clfs.items():
    clf.save_model(f"/app/router_models/tier_{tier}_success.json")
with open("/app/router_models/feat_keys.json","w") as f:
    json.dump(FEAT_KEYS, f)
with open("/app/router_models/tier_config.json","w") as f:
    json.dump({"tier_cost":TIER_COST,"tier_str":TIER_STR,"task_floor":TASK_FLOOR}, f, indent=2)
with open("/app/router_models/eval_results_v3.json","w") as f:
    json.dump(results, f, indent=2, default=str)
print(f"  Saved all models to /app/router_models/")

# Find best config
best_name = ""
best_score = -float("inf")
for name, r in results.items():
    if name in ("oracle","always_cheap"): continue
    # Composite: success*20 - cost*50 - unsafe*100
    score = r["success"]*20 - r["avg_cost"]*50 - r["unsafe_rate"]*100
    if score > best_score:
        best_score = score
        best_name = name
print(f"\n  BEST CONFIG: {best_name} (composite score: {best_score:.2f})")
print(f"\nDONE!")