calculator

Calculator

When To Use

• The user wants a deterministic numeric calculation rather than a conceptual math explanation.
• The task involves expression evaluation, high-precision constants or large numbers, matrices, descriptive statistics, correlations, regressions, distribution pdf/cdf values, scalar optimization, roots, definite integrals, simple ODE samples, FFT/IFFT, or convolution.
• The user asks to verify arithmetic, formulas, generated answers, or intermediate numeric results.

Do not use for K12 tutoring, grading, curriculum-aligned explanations, or error diagnosis; those require a dedicated learning-support workflow. Do not use for open-ended mathematical modeling; keep modeling assumptions and model selection in a dedicated modeling workflow when available.

How To Call

Use the bundled script through Orkas:

"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- eval "2**10 + sin(pi/4)"

Choose the subcommand by computation type:

| Need | Subcommand | Example | | --- | --- | --- | | Daily arithmetic, trig, logs, powers | eval | eval "sin(pi/4) + sqrt(2)" | | Arbitrary precision, factorial, gamma, zeta | precision | precision "factorial(100)" --precision 50 | | Matrix determinant, inverse, solve, eigenvalues | matrix | matrix det --matrix "[[1,2],[3,4]]" | | Descriptive stats, correlation, regression, pdf/cdf | stats | stats describe --data "[1,2,3,4,5]" | | Scalar minimization, maximization, roots | optimize | optimize root --expr "x**3 - 2*x - 5" --bounds "[1,3]" | | Definite integrals or simple ODE samples | integrate | integrate definite --expr "sin(x)" --bounds "[0,pi]" | | FFT, IFFT, convolution | transform | transform fft --data "[1,0,1,0]" |

Core examples:

"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- eval "2**10 + sin(pi/4)"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- eval "pi" --precision 50
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- precision "factorial(1000)" --precision 50
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- matrix inverse --matrix "[[1,2],[3,4]]"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- matrix solve --matrix "[[2,1],[1,-1]]" --matrix2 "[5,1]"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- stats regression --data "[1,2,3]" --data2 "[2,4,6]"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- optimize minimize --expr "x**2 + 2*x + 1" --bounds "[-10,10]"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- integrate definite --expr "sin(x)" --bounds "[0,pi]"
"$ORKAS_NODE" "$ORKAS_PC_DIR/bin/run-skill.cjs" calculator calc -- transform convolve --data "[1,2,3]" --data2 "[4,5]"

Workflow

1. Translate the user's calculation request into the smallest matching subcommand.
2. Run the script before reporting numeric answers when exact numeric verification matters.
3. Include the expression, subcommand, and result in the response when the user may need to audit the calculation.
4. For high precision, large factorials, special functions, or results sensitive to floating-point rounding, use precision instead of eval.
5. For expressions starting with -, pass the value with equals syntax, such as --expr="-x**2+4", so argument parsing treats it as a value.
6. If the script errors because a dependency is missing, report the missing package and either compute by another reliable available method or ask before installing anything.

Output Shape

The script prints:

=== <subcommand> ===
<result>

When responding to the user, keep the answer compact:

Result: ...
Check: ...

For multi-step verification:

**Calculation**
- Command:
- Result:
- Interpretation:

Dependencies

• Python 3.
• Basic scalar eval can run with Python stdlib math functions.
• numpy is required for matrix, statistics, transforms, and array-oriented expression behavior.
• mpmath is required for arbitrary precision.
• scipy is required for distribution functions, optimization, and integration.

Limits And Safety

• Expressions are evaluated with restricted namespaces for math functions and constants, but still treat user expressions as calculation input only.
• eval uses float64-style numeric behavior unless --precision is supplied.
• Matrix and data inputs use Python literal syntax such as [[1,2],[3,4]] or [1,2,3].
• This is not a symbolic algebra system. For proofs, derivations, or curriculum-aligned explanation, use reasoning or a dedicated learning-support workflow alongside calculator verification.