verifyformula(n, points, k, m = 1) or activatejuliafunction(n, points, k, m = 1)

Verify if a formula is valid. If it is valid, generate and activate its Julia function(s). If not, try to find a formula for the derivative using the points.

     n: the n-th order derivative
points: in the format of a range, start : stop, or a list
     k: a list of the coefficients in a formula
     m: the coefficient in the denominator of a formula

Examples

import FiniteDifferenceFormula as fd
fd.verifyformula(1,[-1,2],[-3,4],5)  # f'(x[i]) = (-3f(x[i-1])+4f(x[i+2]))/(5h)?
fd.verifyformula(2, -3:3, [2,-27,270,-490,270,-27,2], 18)
fd.activatejuliafunction(2, -3:3, [1/90,-3/20,3/2,-49/18,3/2,-3/20,1/90])
fd.verifyformula(2, -3:3, [1//90,-3//20,3//2,-49//18,3//2,-3//20,1//90])

compute(n, points, printformulaq = false)

Compute a formula for the nth order derivative using the given points.

            n: the n-th order derivative to be found
       points: in the format of a range, start : stop, or a vector
printformulaq: print the computed formula or not

A vector can be like [1, 0, 2] or [1 0 2]. It will be rearranged so that elements are ordered from lowest to highest with duplicate ones removed.

Examples

julia> import FiniteDifferenceFormula as fd
julia> fd.compute(2, [0 1 2 3])
julia> fd.compute(2, 0:3)
julia> fd.compute(2, [-5, -2, 1, 2, 4], true)

decimalplaces() or decimalplaces(n)

Show present decimal places for generating Julia function(s) of computed formulas, if no argument is passed to decimalplaces. Otherwise, set the decimal places to n.

Examples

import FiniteDifferenceFormula as fd
fd.compute(2,-3:3)
fd.formula()  # by default, use 16 decimal places to generate a Julia function
fd.decimalplaces(4)
fd.formula()  # now, use 4 decimal places to generate a Julia function

find(n, points, printformulaq = false)

Compute a formula for the nth order derivative using the given points.

For the input, n and points (See [compute]), there may not be formulas which use the two end points like -2 and 3 in -2 : 3 or [-2, -1, 0, 1, 2, 3]. In this case, find tries to find a formula by shrinking the range to, first -1 : 3, then, -2 : 2, then -1 : 2, and so on, until a formula is found or no formulas can be found at all.

See also [compute], [findbackward], and [findforward].

Examples

import FiniteDifferenceFormula as fd
fd.find(2, -10:9)

findbackward(n, points, printformulaq = false)

Compute a formula for the nth order derivative using the given points.

For the input, n and points (See [compute]), there may not be formulas which use the two end points like -2 and 3 in -2 : 3 or [-2, -1, 0, 1, 2, 3]. In this case, findbackward tries to find a formula by shrinking the range from the right endpoint to, first -2 : 2, then, -2 : 1, then -2 : 0, and so on, until a formula is found or no formulas can be found at all.

See also [compute], [find], and [findforward].

Examples

import FiniteDifferenceFormula as fd
fd.compute(3,-100:50)
# output: ***** Warning: 3, -100:22 might be your input for which a formula is found.
fd.findbackward(3,-99:50)
# output: (3, -99:23, ...)

findforward(n, points, printformulaq = false)

Compute a formula for the nth order derivative using the given points.

For the input, n and points (See [compute]), there may not be formulas which use the two end points like -2 and 3 in -2 : 3 or [-2, -1, 0, 1, 2, 3]. In this case, findforward tries to find a formula by shrinking the range from the left endpoint to, first -1 : 3, then, 0 : 3, then 1 : 3, and so on, until a formula is found or no formulas can be found at all.

See also [compute], [find], and [findbackward].

Examples

import FiniteDifferenceFormula as fd
fd.findforward(2, -10:9)

formula()

Generate and list:

1. k[1]*f(x[i+points[1]]) + k[2]*f(x[i+points[2]]) + ... + k[len]*f(x[i+points[len]]) = m*f^(n)(x[i]) + ..., m > 0

2. The formula for f^(n)(x[i]), including estimation of accuracy in the big-O notation.

3. Julia function(s) for f^(n)(x[i]).

Calling compute(n, points, true) is the same as calling compute(n, points) and then formula().

Even if no formula can be found, it still lists the computing results from which we can see why. For example, after compute(2,1:2), try formula().

formulas(orders = 1:3, minnumofpoints = 2, maxnumofpoints = 5)

By default, the function prints all forward, backward, and central finite difference formulas for the 1st, 2nd, and 3rd derivatives, using 2 to 5 points.

Examples

# The following examples show all forward, backward, and central finite
# difference formulas for the specified derivatives, using 4 to 11 points.
julia> import FiniteDifferenceFormula as fd
julia> fd.formulas(2:5, 4, 11)       # the 2nd, 3rd, .., 5th derivatives
juliq> fd.formulas([2, 4, 7], 4, 11) # the 2nd, 4th, and 7th derivatives
julia> fd.formulas(3, 4, 11)         # the 3rd derivative

loadcomputingresults(results)

Input: 'results' is a tuple, (n, points, k[:], m). See compute(...).

Load computing results from the output of compute(...). After this command, formula(), activatepythonfunction(), truncationerror(0, etc., are available. It allows users to work on saved computing results (say, in a textfile). For example, when it takes hours to compute/find a formula, users may run commands like the following one from OS terminal

julia -e "import FiniteDifferenceFormula as fd; println(fd.compute(1, range(-200, 201)))" > data.txt

and then mannually load data from data.txt to this function later.

See also [compute].

taylor()

• Print the first few nonzero terms of the Taylor series of the linear combination k[0]f(x[i+j0]) + k[1]f(x[i+j1]) + ... for the newly computed formula (even if failed).

taylor(j, n = 10)

• Print the 1st n terms of Taylor series of f(x[i+j]) about x[i].

taylor(coefs, n = 10), or

• Print the 1st n terms of Taylor series with coefficients in 'coefs'

taylor(points, k, n::Int = 10)

• Prints the 1st n nonzero terms of the Taylor series of the linear combination: k[0]f(x[i+points[0]]) + k[1]f(x[i+points[1]]) + ...

The last two provide also another way to verify if a formula is mathematically valid or not.

See also [verifyformula], [activatejuliafunction], and [taylorcoefs].

Examples

import FiniteDifferenceFormula as fd
fd.compute(1, [0, 1, 5, 8])
fd.taylor()

fd.taylor(2)

n = 50
coefs = 2*fd.tcoefs(0, n) - 6*fd.tcoefs(1, n) + 4*fd.tcoefs(2, n)
fd.taylor(coefs, n) # this n can be any positive integer

fd.taylor(-fd.tcoefs(0) + 3*fd.tcoefs(1) - 3*fd.tcoefs(2) + fd.tcoefs(3))
fd.taylor(0:3, [-1, 3, -3, 1], 6)

taylorcoefs(j, n = 10))

Compute and return coefficients of the first n terms of the Taylor series of f(x[i + j]) = f(x[i] + jh) about x[i], where h is the increment in x.

Examples

import FiniteDifferenceFormula as fd
fd.taylorcoefs(-2)
fd.taylorcoefs(5, 4)

tcoefs(j, n = 10))

Same as taylorcoefs(j, n).

truncationerror()

Show the truncation error of the last computed formula in the big_O notation.

Examples

import FiniteDifferenceFormula as fd
fd.compute(2,-3:3)
fd.truncationerror()
fd.find(3,[-2, 1, 2, 5, 7, 15])
fd.truncationerror()

verifyformula(n, points, k, m = 1) or activatejuliafunction(n, points, k, m = 1)

Verify if a formula is valid. If it is valid, generate and activate its Julia function(s). If not, try to find a formula for the derivative using the points.

     n: the n-th order derivative
points: in the format of a range, start : stop, or a vector
     k: a list of the coefficients in a formula
     m: the coefficient in the denominator of a formula

Examples

import FiniteDifferenceFormula as fd
fd.verifyformula(1,[-1,2],[-3,4],5)  # f'(x[i]) = (-3f(x[i-1])+4f(x[i+2]))/(5h)?
fd.verifyformula(2, -3:3, [2,-27,270,-490,270,-27,2], 18)
fd.activatejuliafunction(2, -3:3, [1/90,-3/20,3/2,-49/18,3/2,-3/20,1/90])
fd.verifyformula(2, -3:3, [1//90,-3//20,3//2,-49//18,3//2,-3//20,1//90])