File size: 7,710 Bytes
9936912
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#!/usr/bin/env python3
"""Base STEM and Mathematical Reasoning replay dataset generator.

Prevents catastrophic forgetting by mixing in step-by-step Chain of Thought (CoT)
linear algebra, calculus, ODE, Laplace transforms, and physical systems reasoning.
"""

from __future__ import annotations

import math
import sys
from pathlib import Path

import numpy as np

PROJECT_ROOT = Path(__file__).resolve().parents[1]
if str(PROJECT_ROOT) not in sys.path:
    sys.path.insert(0, str(PROJECT_ROOT))

from controlai_data.schema import make_record


def n(val: float) -> str:
    if abs(val) < 1e-12:
        return "0.0"
    return f"{val:.6g}"


def mat_str(arr: np.ndarray) -> str:
    return repr(np.asarray(arr, dtype=float).tolist())


def generate_laplace_partial_fractions(count: int, seed: int) -> list[dict]:
    """Generate partial fraction expansion of rational transfer functions with step-by-step math."""
    records = []
    for i in range(count):
        p1 = -(1.0 + 0.5 * (i % 4))
        p2 = -(3.0 + 0.5 * (i % 5))
        z = 2.0 + 0.25 * (i % 3)
        # F(s) = (s + z) / ((s - p1)(s - p2)) = A / (s - p1) + B / (s - p2)
        # A = (p1 + z) / (p1 - p2)
        # B = (p2 + z) / (p2 - p1)
        A_res = (p1 + z) / (p1 - p2)
        B_res = (p2 + z) / (p2 - p1)

        prompt = (
            f"Compute the inverse Laplace transform $f(t) = \\mathcal{{L}}^{{-1}}\\{{F(s)\\}}$ for $t \\ge 0$, where "
            f"$$F(s) = \\frac{{s + {n(z)}}}{{(s + {n(-p1)})(s + {n(-p2)})}}$$. "
            f"Show the step-by-step partial fraction expansion and state the time-domain signal."
        )

        answer = (
            f"### 1. Partial Fraction Decomposition\n"
            f"Write $F(s)$ in terms of first-order partial fractions with distinct poles $s_1 = {n(p1)}$ and $s_2 = {n(p2)}$:\n"
            f"$$F(s) = \\frac{{s + {n(z)}}}{{(s - ({n(p1)}))(s - ({n(p2)}))}} = \\frac{{C_1}}{{s - ({n(p1)})}} + \\frac{{C_2}}{{s - ({n(p2)})}}$$\n\n"
            f"### 2. Residue Calculation\n"
            f"Applying the residue formula:\n"
            f"$$C_1 = \\lim_{{s \\to {n(p1)}}} (s - ({n(p1)})) F(s) = \\frac{{{n(p1)} + {n(z)}}}{{{n(p1)} - ({n(p2)})}} = \\frac{{{n(p1 + z)}}}{{{n(p1 - p2)}}} = {n(A_res)}$$\n"
            f"$$C_2 = \\lim_{{s \\to {n(p2)}}} (s - ({n(p2)})) F(s) = \\frac{{{n(p2)} + {n(z)}}}{{{n(p2)} - ({n(p1)})}} = \\frac{{{n(p2 + z)}}}{{{n(p2 - p1)}}} = {n(B_res)}$$\n\n"
            f"Thus, the expanded expression is:\n"
            f"$$F(s) = \\frac{{{n(A_res)}}}{{s + {n(-p1)}}} + \\frac{{{n(B_res)}}}{{s + {n(-p2)}}}$$\n\n"
            f"### 3. Inverse Laplace Transform\n"
            f"Using the standard pair $\\mathcal{{L}}^{{-1}}\\{{\\frac{{1}}{{s + a}}\\}} = e^{{-at}} u(t)$:\n"
            f"$$f(t) = ({n(A_res)} e^{{{n(p1)} t}} + {n(B_res)} e^{{{n(p2)} t}}) u(t)$$\n"
            f"**Conclusion:** The causal time-domain signal is $f(t) = {n(A_res)} e^{{{n(p1)} t}} + {n(B_res)} e^{{{n(p2)} t}}$ for $t \\ge 0$."
        )

        record = make_record(
            record_id=f"stem_laplace_partial_fraction_{i:04d}",
            domain="classical_optimal_control",
            family="stem_laplace_partial_fraction",
            task_type="derivation",
            difficulty="foundation",
            template_id=f"laplace_partial_fraction_prompt_{i % 4}",
            prompt=prompt,
            answer=answer,
            ground_truth={
                "kind": "laplace_inversion",
                "p1": p1,
                "p2": p2,
                "z": z,
                "C1": A_res,
                "C2": B_res,
            },
            source_refs=["stanford_ee263_course_reader", "astrom_murray_feedback_systems_1e"],
            verifier="verify_laplace_inversion",
            tool="requirements_analysis",
        )
        records.append(record)
    return records


def generate_rlc_circuit_cases(count: int, seed: int) -> list[dict]:
    """Generate series RLC circuit frequency response & damping ratio analysis."""
    records = []
    for i in range(count):
        R = 10.0 + 2.0 * (i % 5)  # Ohms
        L = 0.05 + 0.01 * (i % 4)  # Henry
        C = 0.001 + 0.0002 * (i % 3)  # Farad

        # Series RLC: L * ddot(q) + R * dot(q) + (1/C)*q = v(t)
        # Characteristic eq: s^2 + (R/L)s + 1/(LC) = s^2 + 2*zeta*wn*s + wn^2
        wn = 1.0 / math.sqrt(L * C)
        zeta = R / (2.0 * math.sqrt(L / C))
        wd = wn * math.sqrt(abs(1.0 - zeta**2)) if zeta < 1.0 else 0.0

        if zeta < 1.0:
            regime = "underdamped"
        elif abs(zeta - 1.0) < 1e-4:
            regime = "critically damped"
        else:
            regime = "overdamped"

        prompt = (
            f"Analyze a series RLC circuit with resistor $R = {n(R)}$ $\\Omega$, inductor $L = {n(L)}$ H, "
            f"and capacitor $C = {n(C)}$ F. "
            f"Find the natural frequency $\\omega_n$, damping ratio $\\zeta$, damped natural frequency $\\omega_d$, "
            f"and characterize the transient response regime."
        )

        answer = (
            f"### 1. Differential Equation and Standard Form\n"
            f"Applying Kirchhoff's Voltage Law (KVL) around the series loop with charge $q(t)$:\n"
            f"$$L \\frac{{d^2q}}{{dt^2}} + R \\frac{{dq}}{{dt}} + \\frac{{1}}{{C}} q = v(t)$$\n\n"
            f"Dividing by $L$ puts the equation into canonical second-order form $\\ddot{{q}} + 2\\zeta \\omega_n \\dot{{q}} + \\omega_n^2 q = \\frac{{1}}{{L}} v(t)$:\n"
            f"$$\\ddot{{q}} + \\frac{{{n(R)}}}{{{n(L)}}} \\dot{{q}} + \\frac{{1}}{{{n(L)} \\times {n(C)}}} q = \\frac{{1}}{{{n(L)}}} v(t)$$\n\n"
            f"### 2. Parameter Extraction\n"
            f"1. **Natural frequency $\\omega_n$**:\n"
            f"$$\\omega_n = \\frac{{1}}{{\\sqrt{{LC}}}} = \\frac{{1}}{{\\sqrt{{{n(L)} \\times {n(C)}}}}} = {n(wn)}\\text{{ rad/s}}$$\n\n"
            f"2. **Damping ratio $\\zeta$**:\n"
            f"$$2\\zeta \\omega_n = \\frac{{R}}{{L}} \\implies \\zeta = \\frac{{R}}{{2 \\sqrt{{L/C}}}} = \\frac{{{n(R)}}}{{2 \\sqrt{{{n(L)} / {n(C)}}}}} = {n(zeta)}$$\n\n"
            f"3. **Damped frequency $\\omega_d$**:\n"
            f"$$\\omega_d = \\omega_n \\sqrt{{|1 - \\zeta^2|}} = {n(wd)}\\text{{ rad/s}}$$\n\n"
            f"### 3. Conclusion\n"
            f"Since $\\zeta = {n(zeta)}$ ({'< 1' if zeta < 1.0 else '> 1'}), the system is **{regime}**."
        )

        record = make_record(
            record_id=f"stem_rlc_circuit_{i:04d}",
            domain="classical_optimal_control",
            family="stem_rlc_circuit_transient",
            task_type="numerical",
            difficulty="foundation",
            template_id=f"rlc_circuit_prompt_{i % 4}",
            prompt=prompt,
            answer=answer,
            ground_truth={
                "kind": "second_order_transient",
                "R": R,
                "L": L,
                "C": C,
                "wn": wn,
                "zeta": zeta,
                "wd": wd,
                "regime": regime,
            },
            source_refs=["stanford_ee263_course_reader", "astrom_murray_feedback_systems_1e"],
            verifier="verify_second_order_transient",
            tool="requirements_analysis",
        )
        records.append(record)
    return records


def build_all_stem_records(count_per_case: int = 50, seed: int = 20260902) -> list[dict]:
    """Compile all STEM replay records."""
    records = []
    records.extend(generate_laplace_partial_fractions(count_per_case, seed))
    records.extend(generate_rlc_circuit_cases(count_per_case, seed + 1000))
    return records


if __name__ == "__main__":
    cases = build_all_stem_records(40)
    print(f"Generated {len(cases)} STEM replay SFT records.")