symath 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +12 -0
- data/.rspec +3 -0
- data/.travis.yml +7 -0
- data/CODE_OF_CONDUCT.md +74 -0
- data/Gemfile +8 -0
- data/LICENSE.txt +21 -0
- data/README.md +616 -0
- data/Rakefile +6 -0
- data/bin/console +14 -0
- data/bin/setup +8 -0
- data/lib/symath/definition/abs.rb +48 -0
- data/lib/symath/definition/arccos.rb +25 -0
- data/lib/symath/definition/arccot.rb +23 -0
- data/lib/symath/definition/arccsc.rb +24 -0
- data/lib/symath/definition/arcsec.rb +24 -0
- data/lib/symath/definition/arcsin.rb +25 -0
- data/lib/symath/definition/arctan.rb +23 -0
- data/lib/symath/definition/bounds.rb +39 -0
- data/lib/symath/definition/codiff.rb +31 -0
- data/lib/symath/definition/constant.rb +111 -0
- data/lib/symath/definition/cos.rb +17 -0
- data/lib/symath/definition/cot.rb +17 -0
- data/lib/symath/definition/csc.rb +17 -0
- data/lib/symath/definition/curl.rb +27 -0
- data/lib/symath/definition/d.rb +62 -0
- data/lib/symath/definition/div.rb +27 -0
- data/lib/symath/definition/exp.rb +112 -0
- data/lib/symath/definition/fact.rb +55 -0
- data/lib/symath/definition/flat.rb +31 -0
- data/lib/symath/definition/function.rb +197 -0
- data/lib/symath/definition/grad.rb +23 -0
- data/lib/symath/definition/hodge.rb +23 -0
- data/lib/symath/definition/int.rb +75 -0
- data/lib/symath/definition/laplacian.rb +23 -0
- data/lib/symath/definition/lmd.rb +97 -0
- data/lib/symath/definition/ln.rb +45 -0
- data/lib/symath/definition/number.rb +51 -0
- data/lib/symath/definition/operator.rb +228 -0
- data/lib/symath/definition/sec.rb +17 -0
- data/lib/symath/definition/sharp.rb +31 -0
- data/lib/symath/definition/sin.rb +17 -0
- data/lib/symath/definition/sqrt.rb +62 -0
- data/lib/symath/definition/tan.rb +17 -0
- data/lib/symath/definition/trig.rb +95 -0
- data/lib/symath/definition/variable.rb +284 -0
- data/lib/symath/definition/xd.rb +28 -0
- data/lib/symath/definition.rb +205 -0
- data/lib/symath/equation.rb +67 -0
- data/lib/symath/fraction.rb +177 -0
- data/lib/symath/matrix.rb +252 -0
- data/lib/symath/minus.rb +125 -0
- data/lib/symath/operation/differential.rb +167 -0
- data/lib/symath/operation/distributivelaw.rb +367 -0
- data/lib/symath/operation/exterior.rb +64 -0
- data/lib/symath/operation/integration.rb +329 -0
- data/lib/symath/operation/match.rb +166 -0
- data/lib/symath/operation/normalization.rb +458 -0
- data/lib/symath/operation.rb +36 -0
- data/lib/symath/operator.rb +163 -0
- data/lib/symath/parser.rb +473 -0
- data/lib/symath/parser.y +129 -0
- data/lib/symath/poly/dup.rb +835 -0
- data/lib/symath/poly/galois.rb +621 -0
- data/lib/symath/poly.rb +142 -0
- data/lib/symath/power.rb +224 -0
- data/lib/symath/product.rb +183 -0
- data/lib/symath/sum.rb +174 -0
- data/lib/symath/type.rb +282 -0
- data/lib/symath/value.rb +372 -0
- data/lib/symath/version.rb +3 -0
- data/lib/symath/wedge.rb +48 -0
- data/lib/symath.rb +157 -0
- data/symath.gemspec +39 -0
- metadata +160 -0
data/README.md
ADDED
@@ -0,0 +1,616 @@
|
|
1
|
+
# SyMath
|
2
|
+
|
3
|
+
Rudimentary symbolic math library for Ruby. This gem is mainly intended
|
4
|
+
as a coding excercise. The operations have not been optimized for speed.
|
5
|
+
The current state of the project is 'under construction'.
|
6
|
+
|
7
|
+
# Installation
|
8
|
+
|
9
|
+
Add this line to your application's Gemfile:
|
10
|
+
|
11
|
+
```
|
12
|
+
gem 'symath'
|
13
|
+
```
|
14
|
+
|
15
|
+
Then execute:
|
16
|
+
|
17
|
+
$ bundle
|
18
|
+
|
19
|
+
Or install it yourself as:
|
20
|
+
|
21
|
+
$ gem install symath
|
22
|
+
|
23
|
+
## Usage
|
24
|
+
|
25
|
+
Using the library:
|
26
|
+
|
27
|
+
<pre>
|
28
|
+
> require 'SyMath'
|
29
|
+
</pre>
|
30
|
+
|
31
|
+
### Simple introduction
|
32
|
+
|
33
|
+
A convenient way to explore the SyMath library is using the interactive
|
34
|
+
Ruby interpreter, irb:
|
35
|
+
|
36
|
+
<pre>
|
37
|
+
> # Load the symath library
|
38
|
+
> require 'symath'
|
39
|
+
=> false
|
40
|
+
> # Add the symbols module to your environment
|
41
|
+
> extend SyMath::Definitions
|
42
|
+
=> main
|
43
|
+
</pre>
|
44
|
+
|
45
|
+
You can now say, for example:
|
46
|
+
|
47
|
+
<pre>
|
48
|
+
> # Simplify an expression
|
49
|
+
> sin(:x) + 2*sin(:x)
|
50
|
+
=> 3*sin(:x)
|
51
|
+
> # Derivative of tan(2*y + 3)
|
52
|
+
> (d(tan(2*:y + 3))/d(:y)).evaluate
|
53
|
+
=> 2*(tan(2*y + 3)**2 + 1)
|
54
|
+
</pre>
|
55
|
+
|
56
|
+
Ruby symbols, :x and :y in the above example, are converted into
|
57
|
+
symbolic math variables and Ruby numbers are converted into symbolic
|
58
|
+
math numbers. Functions, operators and constants (e, pi, i, etc.) are
|
59
|
+
available as methods through the SyMath::Definitions module. In some cases
|
60
|
+
it is necessary to tell Ruby that your number or symbol is to be
|
61
|
+
understood as a symbolic object, and not just a Ruby number or
|
62
|
+
symbol. Use the to_m method to explicitly convert them to symbolic
|
63
|
+
bjects:
|
64
|
+
|
65
|
+
<pre>
|
66
|
+
> # Ruby integer math
|
67
|
+
> 3**4
|
68
|
+
=> 81
|
69
|
+
> # SyMath symbolic math
|
70
|
+
> 3.to_m**4
|
71
|
+
=> 3**4
|
72
|
+
> (3.to_m**4).normalize
|
73
|
+
=> 81
|
74
|
+
</pre>
|
75
|
+
|
76
|
+
An complete expression can also be converted from a string, using the
|
77
|
+
same to_m method:
|
78
|
+
|
79
|
+
<pre>
|
80
|
+
> 'ln(e) + sin(pi/2)'.to_m
|
81
|
+
=> ln(e) + sin(pi/2)
|
82
|
+
> 'ln(e) + sin(pi/2)'.to_m.normalize
|
83
|
+
=> 2
|
84
|
+
</pre>
|
85
|
+
|
86
|
+
### The SyMath::Definitions module
|
87
|
+
|
88
|
+
The module SyMath::Definitions is available to be included or extended
|
89
|
+
to your class or code block. It gives a Ruby method for each operator,
|
90
|
+
function and constant that exists, so they can be referred to by their
|
91
|
+
name, as in the code examples above. If you don't want to use the
|
92
|
+
module, functions, operators and constants must be referred to by the
|
93
|
+
fn, op and definition methods:
|
94
|
+
|
95
|
+
<pre>
|
96
|
+
> # Using the SyMath::Definitions methods
|
97
|
+
> sin(:x)
|
98
|
+
=> sin(x)
|
99
|
+
> int(:x)
|
100
|
+
=> int(x)
|
101
|
+
> e
|
102
|
+
=> e
|
103
|
+
> sin
|
104
|
+
=> sin(...)
|
105
|
+
> # Using the generic creator functions
|
106
|
+
> fn(:sin, :x)
|
107
|
+
=> sin(x)
|
108
|
+
> op(:int, :x)
|
109
|
+
=> int(x)
|
110
|
+
> definition(:e)
|
111
|
+
=> e
|
112
|
+
> definition(:sin)
|
113
|
+
=> sin(...)
|
114
|
+
</pre>
|
115
|
+
|
116
|
+
The SyMath::Definitions module is updated dynamically after the user has
|
117
|
+
defined new functions, operators and constants.
|
118
|
+
|
119
|
+
### String representaton of symbolic objects
|
120
|
+
|
121
|
+
Symbolic math objects, inheriting from SyMath::Value, all have a to_s
|
122
|
+
method which returns a string representation of the object. The string
|
123
|
+
representation is compatible with the String.to_m method which
|
124
|
+
converts a string representation into a symbolic object:
|
125
|
+
|
126
|
+
<pre>
|
127
|
+
> (ln(e) + sin(pi/2)).to_s
|
128
|
+
=> "ln(e) + sin(pi/2)"
|
129
|
+
> 'ln(e) + sin(pi/2)'.to_m
|
130
|
+
=> ln(e) + sin(pi/2)
|
131
|
+
</pre>
|
132
|
+
|
133
|
+
SyMath::Value overrides the Object.inspect method, returning the to_s
|
134
|
+
representation rather than the more verbose and less readable
|
135
|
+
Object.inspect output. This behaviour can be disabled with the setting
|
136
|
+
'inspect_to_s':
|
137
|
+
|
138
|
+
<pre>
|
139
|
+
> SyMath.setting(:inspect_to_s, false)
|
140
|
+
=> false
|
141
|
+
> ln(e) + sin(pi/2)
|
142
|
+
=> "#<SyMath::Sum:0x000055e8a1d93b38 @definition=..."
|
143
|
+
</pre>
|
144
|
+
|
145
|
+
### Simplification and normalizing
|
146
|
+
|
147
|
+
Simple reduction rules are automatically applied when composing an
|
148
|
+
expression. These can be disabled with the setting
|
149
|
+
'compose_with_simplify'. More thorough reductions are done by the use
|
150
|
+
of the normalize method.
|
151
|
+
|
152
|
+
<pre>
|
153
|
+
> e*e*e*e
|
154
|
+
=> e**4
|
155
|
+
> SyMath.setting(:compose_with_simplify, false)
|
156
|
+
=> false
|
157
|
+
> e*e*e*e
|
158
|
+
=> e*e*e*e
|
159
|
+
> sin(pi/2).normalize
|
160
|
+
=> 1
|
161
|
+
</pre>
|
162
|
+
|
163
|
+
### Functions
|
164
|
+
|
165
|
+
The library comes with a number of built-in function, which the system
|
166
|
+
knows how to derivate and integrate over. The built-in functions also
|
167
|
+
have a number of reduction rules which are applied by the reduce
|
168
|
+
method and also as part of the 'normalize' method. A list of the
|
169
|
+
defined functions is returned by the functions method. The description
|
170
|
+
method gives a small description of the function:
|
171
|
+
|
172
|
+
<pre>
|
173
|
+
> SyMath::Definition::Function.functions
|
174
|
+
=> [sqrt, sin, cos, tan, sec, csc, cot, arcsin, arccos, arctan,
|
175
|
+
arcsec, arccsc, arccot, ln, exp, abs, fact, sinh, cosh, tanh,
|
176
|
+
coth, sech, csch, arsinh, arcosh, artanh, arcoth, arsech,
|
177
|
+
arcsch]
|
178
|
+
> sin.description
|
179
|
+
=> "sin(x) - trigonometric sine"
|
180
|
+
</pre>
|
181
|
+
|
182
|
+
### Defining functions
|
183
|
+
|
184
|
+
User-defined functions can be added by the method define_fn:
|
185
|
+
|
186
|
+
<pre>
|
187
|
+
> define_fn('poly', [:x, :y], :x**3 + :y**2 + 1)
|
188
|
+
=> poly
|
189
|
+
</pre>
|
190
|
+
|
191
|
+
The user-defined function will now be available as a method in the
|
192
|
+
SyMath::Definitions module and can be used in expressions, just as the
|
193
|
+
built in functions. Functions defined by an expression can be
|
194
|
+
evaluated by the evaluate method, which returns the expression with
|
195
|
+
each free variable replaced with the input arguments to the function:
|
196
|
+
|
197
|
+
<pre>
|
198
|
+
> poly(2, 3).evaluate
|
199
|
+
=> 2**3 + 3**2 + 1
|
200
|
+
> poly(3).evaluate.normalize
|
201
|
+
=> 18
|
202
|
+
</pre>
|
203
|
+
|
204
|
+
### Lambda functions
|
205
|
+
|
206
|
+
A nameless user-defined function can be created using the lmd
|
207
|
+
method. The method returns a function object which does not have a
|
208
|
+
name, but otherwise works as a function. The lambda function has
|
209
|
+
important usages in operators. Since they eturn a function as the
|
210
|
+
result, it will typically be a lambda function. Also, the lambda
|
211
|
+
function can be used for wrapping an expression into a function before
|
212
|
+
doing an integral or derivative, in this way telling which variables
|
213
|
+
the operator should work on. The lambda function can be called using
|
214
|
+
the call method or the Ruby 'call' operator '()':
|
215
|
+
|
216
|
+
<pre>
|
217
|
+
> l = lmd(:x**3 + :y**2 + 1, :x, :y)
|
218
|
+
> l.(2, 3)
|
219
|
+
=> (x**3 + y**2 + 1).(2,3)
|
220
|
+
> l.(2, 3).evaluate
|
221
|
+
=> 2**3 + 3**2 + 1
|
222
|
+
> l.(2, 3).evaluate.normalize
|
223
|
+
=> 18
|
224
|
+
</pre>
|
225
|
+
|
226
|
+
### Operators
|
227
|
+
|
228
|
+
The library has some built-in operators, i.e. functions which take
|
229
|
+
functions as arguments and return functions. A list of the defined
|
230
|
+
operators is returned by the operators method. The description method
|
231
|
+
gives a small description of the operator:
|
232
|
+
|
233
|
+
<pre>
|
234
|
+
> SyMath::Definition::Operator.operators
|
235
|
+
=> [d(...), xd(...), int(...), [f](b,), #(), b(), hodge(...), grad(f),
|
236
|
+
curl(f), div(f), laplacian(f), codiff(f), laplace(f), fourier(f),
|
237
|
+
invfourier(f), dpart(f,t)]
|
238
|
+
> codiff.description
|
239
|
+
=> "codiff(f) - codifferential of function f"
|
240
|
+
</pre>
|
241
|
+
|
242
|
+
### Defining operators
|
243
|
+
|
244
|
+
User-defined operators can be added by the method define_op:
|
245
|
+
|
246
|
+
<pre>
|
247
|
+
> define_op('d2', [:f, :x], d(d(:f)/d(:x))/d(:x))
|
248
|
+
=> d2
|
249
|
+
> d2(:x**3 + 2, :x).evaluate
|
250
|
+
=> 6*x
|
251
|
+
</pre>
|
252
|
+
|
253
|
+
The user-defined function will now be available as a method in the
|
254
|
+
SyMath::Definitions module and can be used in expressions.
|
255
|
+
|
256
|
+
### Evaluating functions and operators
|
257
|
+
|
258
|
+
Evaluating a functions or operators which is defined by an expression
|
259
|
+
returns the expression with each free variable replaced with input
|
260
|
+
arguments. Functions which do not have an expression will typically
|
261
|
+
evaluate to itself (no reduction). Most operators which do not have an
|
262
|
+
expression has a built in evaluation, and returns a function or
|
263
|
+
expression according to the operator.
|
264
|
+
|
265
|
+
### Derivative
|
266
|
+
|
267
|
+
The d-operator returns the differential of a function or expresson. If
|
268
|
+
a function is given, the differential is made over all the free
|
269
|
+
variables of the function. If an expression is given, the operator
|
270
|
+
differentiates over the first free variable found in the
|
271
|
+
expression. Wrapping the expression into a lambda function makes it
|
272
|
+
possible to say which variables to differentiate over. Note that the
|
273
|
+
differential is an operator, so it returns the result in a a lambda
|
274
|
+
function, and not just the expression.
|
275
|
+
|
276
|
+
<pre>
|
277
|
+
> d(sin(:x)).evaluate
|
278
|
+
=> cos(x)*dx.(x)
|
279
|
+
> d(:x**2 + :y**3 + 1).evaluate.normalize
|
280
|
+
=> (2*x*dx).(x)
|
281
|
+
> d(lmd(:x**2 + :y**3 + 1, :y)).evaluate.normalize
|
282
|
+
=> (3*y**2*dy).(y)
|
283
|
+
> d(lmd(:x**2 + :y**3 + 1, :x, :y)).evaluate.normalize
|
284
|
+
=> (3*y**2*dy + 2*x*dx).(x,y)
|
285
|
+
</pre>
|
286
|
+
|
287
|
+
As a special case, the notatonal form d(f)/d(x) is recognized as the
|
288
|
+
derivative of f with regards to x. This is calculated as d(lmd(f,
|
289
|
+
x))/d(x). This evaluates to the derivative expression:
|
290
|
+
|
291
|
+
<pre>
|
292
|
+
> (d(:y**2 + :x**3 + 1)/d(:x)).evaluate
|
293
|
+
=> 3*x**2
|
294
|
+
</pre>
|
295
|
+
|
296
|
+
The partial derivative is available as well as 'syntactic sugar':
|
297
|
+
|
298
|
+
<pre>
|
299
|
+
> dpart(:x**2 + :y**3 + 1, :x).evaluate.normalize
|
300
|
+
=> 2*x
|
301
|
+
> dpart(:x**2 + :y**3 + 1, :y).evaluate.normalize
|
302
|
+
=> 3*y**2
|
303
|
+
</pre>
|
304
|
+
|
305
|
+
### Integration
|
306
|
+
|
307
|
+
Integration is available as the int-operator. With one argument, the
|
308
|
+
operator evaluates to the antiderivative of the expression:
|
309
|
+
|
310
|
+
<pre>
|
311
|
+
> int(2**:x).evaluate
|
312
|
+
=> 2**x/ln(2) + C
|
313
|
+
</pre>
|
314
|
+
|
315
|
+
The variable C is used by convention to represent the free constant
|
316
|
+
factor of the antiderivative.
|
317
|
+
|
318
|
+
With three arguments, the int-operator evaluates to the definite
|
319
|
+
integral from a to b:
|
320
|
+
|
321
|
+
<pre>
|
322
|
+
> int(2**:x, 3, 4).evaluate.normalize
|
323
|
+
=> 8/ln(2)
|
324
|
+
</pre>
|
325
|
+
|
326
|
+
### Complex numbers and quaternions
|
327
|
+
|
328
|
+
The imaginary unit, i, is available as a constant, and can be used for
|
329
|
+
composing expressions with complex numbers. Simple reduction rules are
|
330
|
+
built in which reduces i*i to -1, and so on.
|
331
|
+
|
332
|
+
The basic quaternions, i, j, k are also available as constants. The
|
333
|
+
quaternion i is identical to the complex imaginary unit. Some simple
|
334
|
+
reduction rules are available for the quaternions as well.
|
335
|
+
|
336
|
+
### Exterior algebra
|
337
|
+
|
338
|
+
Caveat: Exterior algebra and differential forms are not well
|
339
|
+
understood by the author of this code. The following has not been
|
340
|
+
reviewed by any others who understand the subject better than me, and
|
341
|
+
it may very well contain a lot of errors and misunderstandings.
|
342
|
+
|
343
|
+
D-forms can be defined in several ways. The following are equal:
|
344
|
+
|
345
|
+
<pre>
|
346
|
+
> # Using the to_d method on a scalar variable
|
347
|
+
> :x.to_m.to_d
|
348
|
+
=> dx
|
349
|
+
> # Differentiating a scalar variable
|
350
|
+
> d(:x)
|
351
|
+
=> dx
|
352
|
+
> # Creating a variable, and specifying the dform type
|
353
|
+
> :dx.to_m('dform')
|
354
|
+
=> dx
|
355
|
+
</pre>
|
356
|
+
|
357
|
+
D-forms can be wedged together, forming n-forms (note that the ^
|
358
|
+
operator has lower preceedence in Ruby than in math, so parantheses
|
359
|
+
must be used, e.g. when adding):
|
360
|
+
|
361
|
+
<pre>
|
362
|
+
> d(:x)^d(:y)^d(:z)
|
363
|
+
=> dx^dy^dz
|
364
|
+
> (d(:x)^d(:x)^d(:z)).normalize
|
365
|
+
=> 0
|
366
|
+
</pre>
|
367
|
+
|
368
|
+
A vector variable can be defined using the to_m method, and specifying
|
369
|
+
the vector type.
|
370
|
+
|
371
|
+
<pre>
|
372
|
+
> :v.to_m('vector')
|
373
|
+
=> v'
|
374
|
+
</pre>
|
375
|
+
|
376
|
+
The exterior derivative and related operators all work in a local
|
377
|
+
coordinate system defined by set of basic vectors of names :x1, :x2,
|
378
|
+
x3. The names can be changed by setting the built-in variable 'basis':
|
379
|
+
|
380
|
+
<pre>
|
381
|
+
> SyMath.get_variable('basis')
|
382
|
+
=> [x1, x2, x3]
|
383
|
+
> SyMath.assign_variable('basis', [:x, :y, :z])
|
384
|
+
=> {dx=>x', dy=>y', dz=>z'}
|
385
|
+
> SyMath.get_variable('basis')
|
386
|
+
=> [x, y, z]
|
387
|
+
</pre>
|
388
|
+
|
389
|
+
The rest of this section assumes that the following scalars, vectors
|
390
|
+
and d-forms are defined:
|
391
|
+
|
392
|
+
<pre>
|
393
|
+
> SyMath.assign_variable('basis', [:x1, :x2, :x3])
|
394
|
+
=> {dx1=>x1', dx2=>x2', dx3=>x3'}
|
395
|
+
> x1 = :x1.to_m
|
396
|
+
> x2 = :x2.to_m
|
397
|
+
> x3 = :x3.to_m
|
398
|
+
> x1v = :x1.to_m('vector')
|
399
|
+
> x2v = :x2.to_m('vector')
|
400
|
+
> x3v = :x3.to_m('vector')
|
401
|
+
> dx1 = :dx1.to_m('dform')
|
402
|
+
> dx2 = :dx2.to_m('dform')
|
403
|
+
> dx3 = :dx3.to_m('dform')
|
404
|
+
</pre>
|
405
|
+
|
406
|
+
The exterior derivative is available as the xd-operator:
|
407
|
+
|
408
|
+
<pre>
|
409
|
+
> xd(:x1 - :x1*:x2 + :x3**2).evaluate
|
410
|
+
=> dx1 - (dx1*x2 + x1*dx2) + 2*x3*dx3
|
411
|
+
</pre>
|
412
|
+
|
413
|
+
The musical isomorphisms are available as the flat and sharp operators:
|
414
|
+
|
415
|
+
<pre>
|
416
|
+
> flat(x1v^x2v).evaluate
|
417
|
+
=> dx1^dx2
|
418
|
+
> sharp(dx1^dx2).evaluate
|
419
|
+
=> x1'^x2'
|
420
|
+
</pre>
|
421
|
+
|
422
|
+
The flat and sharp operators use the metric tensor in their
|
423
|
+
calculations. This is available a the built-in 'g' variable:
|
424
|
+
|
425
|
+
<pre>
|
426
|
+
> SyMath.get_variable(:g)
|
427
|
+
=> [1, 0, 0; 0, 1, 0; 0, 0, 1]
|
428
|
+
</pre>
|
429
|
+
|
430
|
+
It can be changed using the SyMath.assign_variable(:g, [value]) method.
|
431
|
+
|
432
|
+
The hodge star operator is available as well:
|
433
|
+
|
434
|
+
<pre>
|
435
|
+
> hodge(dx1^dx2).evaluate
|
436
|
+
=> dx3
|
437
|
+
> hodge(3).evaluate
|
438
|
+
=> 3*dx1^dx2^dx3
|
439
|
+
</pre>
|
440
|
+
|
441
|
+
Gradient, curl, divergence, laplacian and co-differential are defined
|
442
|
+
from the above operators in the usual way:
|
443
|
+
|
444
|
+
<pre>
|
445
|
+
> grad(x1 - x1*x2 + x3**2).evaluate
|
446
|
+
=> 2*x3*x3' - x2*x1' - x1*x2' + x1'
|
447
|
+
> curl(-x2*x1v + x1*x2*x2v + x3*x3v).evaluate
|
448
|
+
=> x2*x3' + x3'
|
449
|
+
> div(-x2*x1v + x1*x2*x2v + x3*x3v).evaluate
|
450
|
+
=> x1 + 1
|
451
|
+
> laplacian(x1**2 + x2**2 + x3**2).evaluate
|
452
|
+
=> 6
|
453
|
+
> codiff(x1**2*(dx1^dx3) + x2**2*(dx3^dx1) + x3**2*(dx1^dx2)).evaluate
|
454
|
+
=> 2*x1*dx3
|
455
|
+
</pre>
|
456
|
+
|
457
|
+
### Vectors and matrices
|
458
|
+
|
459
|
+
Vectors can be defined in the coordinate array form by converting an
|
460
|
+
array to a math object, using the to_m method. Matrices can be created
|
461
|
+
the same way from two dimensional arrays:
|
462
|
+
|
463
|
+
<pre>
|
464
|
+
> m = [[1, 2, 3], [4, 5, 6]].to_m
|
465
|
+
> v = [-3, 4, 1].to_m
|
466
|
+
> m*v.transpose
|
467
|
+
=> [1, 2, 3; 4, 5, 6]*[- 3; 4; 1]
|
468
|
+
> (m*v.transpose).evaluate
|
469
|
+
=> [8; 14]
|
470
|
+
</pre>
|
471
|
+
|
472
|
+
The vector and matrix cells can of course contain symbolic expressions
|
473
|
+
instead of just numbers.
|
474
|
+
|
475
|
+
### Methods for manipulating expressions
|
476
|
+
|
477
|
+
The library contains a few more complex expression manipulation
|
478
|
+
methods which are available to all math expression objects inheriting
|
479
|
+
from the SyMath::Value class (the root class of the expression
|
480
|
+
components).
|
481
|
+
|
482
|
+
#### Normalization
|
483
|
+
|
484
|
+
The normalization method tries to put an expression on a normal form,
|
485
|
+
based on some heuristics.
|
486
|
+
|
487
|
+
* Expressions formed by natural numbers are calculated.
|
488
|
+
* Fractions of natural numbers are simplified as far as possible.
|
489
|
+
* Products of equal factors are collapsed to power expressions.
|
490
|
+
* Products of powers with equal base are collapsed.
|
491
|
+
* Sums of equal terms are simplified to integer products.
|
492
|
+
* Product factors are ordered if permitted by commutativity.
|
493
|
+
* Sum terms are ordered.
|
494
|
+
|
495
|
+
<pre>
|
496
|
+
> # FIXME: Find some better examples
|
497
|
+
> (:x*4*:x*3*:y*:y**10).normalize
|
498
|
+
=> 12*x**2*y**11
|
499
|
+
</pre>
|
500
|
+
|
501
|
+
#### Variable replacement
|
502
|
+
|
503
|
+
The replace method replaces takes a map of 'variable => expression' as
|
504
|
+
argument. It looks up all instances of the variables in the original
|
505
|
+
expression, and replaces them with the expressions given by the map:
|
506
|
+
|
507
|
+
<pre>
|
508
|
+
> (:x**:y).replace({:x.to_m => :a + 2, :y.to_m => :b + 3})
|
509
|
+
=> (a + 2)**(b + 3)
|
510
|
+
</pre>
|
511
|
+
|
512
|
+
#### Matching and pattern replacement
|
513
|
+
|
514
|
+
The match method can be seen as a 'reverse' operation to
|
515
|
+
replace-method covered in the last section. It compares an expression
|
516
|
+
to a template expression containing some free variables. It returns an
|
517
|
+
array of all possible maps for the free variables in the template so
|
518
|
+
that it matches the original expression:
|
519
|
+
|
520
|
+
<pre>
|
521
|
+
> (:x**2 + :y**2 + 3).match(:a + :b, [:a.to_m, :b.to_m])
|
522
|
+
=> [{a=>x**2, b=>y**2 + 3},
|
523
|
+
{a=>y**2, b=>x**2 + 3},
|
524
|
+
{a=>3, b=>x**2 + y**2},
|
525
|
+
{a=>x**2 + y**2, b=>3},
|
526
|
+
{a=>x**2 + 3, b=>y**2},
|
527
|
+
{a=>y**2 + 3, b=>x**2}]
|
528
|
+
</pre>
|
529
|
+
|
530
|
+
#### Factorization and product expansion
|
531
|
+
|
532
|
+
The factorization method has been ripped from the python
|
533
|
+
Py-library. It factorizes a polynomial of one variable:
|
534
|
+
|
535
|
+
<pre>
|
536
|
+
> (6*:x**2 + 24*:x**3 - 27*:x**4 + 18*:x**5 + 72*:x**6 - 9*:x).factorize
|
537
|
+
=> 3*x*(2*x - 1)*(4*x + 3)*(3*x**3 + 1)
|
538
|
+
</pre>
|
539
|
+
|
540
|
+
The expand method expands the polynomial:
|
541
|
+
|
542
|
+
<pre>
|
543
|
+
> (3*:x*(2*:x - 1)*(4*:x + 3)*(3*:x**3 + 1)).expand.normalize
|
544
|
+
=> 72*x**6 + 18*x**5 - 27*x**4 + 24*x**3 + 6*x**2 - 9*x
|
545
|
+
</pre>
|
546
|
+
|
547
|
+
### Settings
|
548
|
+
|
549
|
+
The library has some global settings which change the behaviour of the system:
|
550
|
+
|
551
|
+
<pre>
|
552
|
+
> # List all settings
|
553
|
+
> SyMath.settings
|
554
|
+
=> {
|
555
|
+
# Symbol used when a d-form is created
|
556
|
+
:d_symbol => "d",
|
557
|
+
# Symbol used when a vector is stringified
|
558
|
+
:vector_symbol => "'",
|
559
|
+
# Co-vector symbol used in a tensor type signature
|
560
|
+
:covector_symbol => ".",
|
561
|
+
# Show all parentheses when stringifying an expression
|
562
|
+
:expl_parentheses => false,
|
563
|
+
# Put square roots on exponent form
|
564
|
+
:sq_exponent_form => false,
|
565
|
+
# Put fractions on exponent form
|
566
|
+
:fraction_exponent_form => false,
|
567
|
+
# In latex strings, insert a product sign between the factors
|
568
|
+
:ltx_product_sign => false,
|
569
|
+
# Use simplification rules when expressions are composed
|
570
|
+
:compose_with_simplify => true,
|
571
|
+
# Use oo as complex infinity
|
572
|
+
:complex_arithmetic => true,
|
573
|
+
# Return to_s representation by the inspect method
|
574
|
+
:inspect_to_s => true,
|
575
|
+
# Maximum value of factorial which is normalized to a number
|
576
|
+
:max_calculated_factorial => 100
|
577
|
+
}
|
578
|
+
> # Show one setting
|
579
|
+
> SyMath.setting(:vector_symbol)
|
580
|
+
=> "'"
|
581
|
+
> # Change a setting
|
582
|
+
> SyMath.setting(:vector_symbol, '¤')
|
583
|
+
=> "¤"
|
584
|
+
</pre>
|
585
|
+
|
586
|
+
## Development
|
587
|
+
|
588
|
+
After checking out the repo, run `bin/setup` to install
|
589
|
+
dependencies. Then, run `rake spec` to run the tests. You can also run
|
590
|
+
`bin/console` for an interactive prompt that will allow you to
|
591
|
+
experiment.
|
592
|
+
|
593
|
+
To install this gem onto your local machine, run `bundle exec rake
|
594
|
+
install`. To release a new version, update the version number in
|
595
|
+
`version.rb`, and then run `bundle exec rake release`, which will
|
596
|
+
create a git tag for the version, push git commits and tags, and push
|
597
|
+
the `.gem` file to [rubygems.org](https://rubygems.org).
|
598
|
+
|
599
|
+
## Contributing
|
600
|
+
|
601
|
+
Bug reports and pull requests are welcome on GitHub at
|
602
|
+
https://github.com/erikoest/symath. This project is intended to be a
|
603
|
+
safe, welcoming space for collaboration, and contributors are expected
|
604
|
+
to adhere to the [Contributor
|
605
|
+
Covenant](http://contributor-covenant.org) code of conduct.
|
606
|
+
|
607
|
+
## License
|
608
|
+
|
609
|
+
The gem is available as open source under the terms of the [MIT
|
610
|
+
License](https://opensource.org/licenses/MIT).
|
611
|
+
|
612
|
+
## Code of Conduct
|
613
|
+
|
614
|
+
Everyone interacting in the SyMath project’s codebases, issue trackers,
|
615
|
+
chat rooms and mailing lists is expected to follow the [code of
|
616
|
+
conduct](https://github.com/erikoest/symath/blob/master/CODE_OF_CONDUCT.md).
|
data/Rakefile
ADDED
data/bin/console
ADDED
@@ -0,0 +1,14 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require "bundler/setup"
|
4
|
+
require "sy"
|
5
|
+
|
6
|
+
# You can add fixtures and/or initialization code here to make experimenting
|
7
|
+
# with your gem easier. You can also use a different console, if you like.
|
8
|
+
|
9
|
+
# (If you use this, don't forget to add pry to your Gemfile!)
|
10
|
+
# require "pry"
|
11
|
+
# Pry.start
|
12
|
+
|
13
|
+
require "irb"
|
14
|
+
IRB.start(__FILE__)
|
data/bin/setup
ADDED
@@ -0,0 +1,48 @@
|
|
1
|
+
require 'symath/definition/function'
|
2
|
+
|
3
|
+
module SyMath
|
4
|
+
class Definition::Abs < Definition::Function
|
5
|
+
def initialize()
|
6
|
+
super(:abs)
|
7
|
+
end
|
8
|
+
|
9
|
+
def description()
|
10
|
+
return 'abs(x) - absolute value'
|
11
|
+
end
|
12
|
+
|
13
|
+
def reduce_call(c)
|
14
|
+
arg = c.args[0]
|
15
|
+
if arg.is_nan?
|
16
|
+
return :nan.to_m
|
17
|
+
# Corner case, -oo is positive with complex arithmetic, so we need a
|
18
|
+
# specific check for that.
|
19
|
+
elsif arg.is_negative? or arg == -:oo
|
20
|
+
return - arg
|
21
|
+
elsif arg.is_positive? or arg.is_zero?
|
22
|
+
return arg
|
23
|
+
else
|
24
|
+
return c
|
25
|
+
end
|
26
|
+
end
|
27
|
+
|
28
|
+
def to_s(args = nil)
|
29
|
+
if args
|
30
|
+
arg = args[0].to_s
|
31
|
+
else
|
32
|
+
arg = '...'
|
33
|
+
end
|
34
|
+
|
35
|
+
return "|#{arg}|"
|
36
|
+
end
|
37
|
+
|
38
|
+
def to_latex(args = nil)
|
39
|
+
if args
|
40
|
+
arg = args[0].to_latex
|
41
|
+
else
|
42
|
+
arg = '...'
|
43
|
+
end
|
44
|
+
|
45
|
+
return "\\lvert#{arg}\\rvert"
|
46
|
+
end
|
47
|
+
end
|
48
|
+
end
|