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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
|
namespace quetzal {
namespace demography {
namespace strategy {
/*!
* \brief Traits class for individual based demographic history simulation, strategy suited for small number of individuals in the landscape (small populations).
* \ingroup demography
*
* \details Simulate the demographic history with an individual-based strategy: each
* individual is dispersed individually.
* \par History Initialization:
* The population size history\f$N\f$ is initialized by setting \f$N(.,t_0)\f$, the initial distribution of individuals across demes at the
* first time of the history \f$t_0\f$. Typically for a biological invasion,
* this is restricted to the introduction site(s) with the number of introduced
* individuals. For endemic species, paleoclimatic distribution can be considered
* as starting points.
*
* \par Growth:
* The offspring number \f$ \tilde{N}_{x}^{t} \f$ in each deme is freely defined by the user. It can for example be
* sampled in a distribution conditionally to a function of the local density of
* parents. Non-overlapping generations are considered (the parents die just after reproduction).
*
* \par Dispersal:
* The children dispersal is done by sampling their destination in a multinomial
* law, that defines \f$ \Phi_{x,y}^t \f$, the number of individuals going from
* \f$x\f$ to \f$y\f$ at time \f$t\f$:
*
* \f[ (\Phi_{x,y}^{t})_{y\in X} \sim \mathcal{M}(\tilde{N}_{x}^{t},(m_{xy})_y) ~. \f]
*
* The term \f$ (m_{xy})_y \f$ denotes the parameters of the multinomial law,
* giving for an individual in \f$x\f$ its proability to go to \f$y\f$.
* These probabilities are given by the dispersal law with parameter \f$\theta\f$:
*
* \f[
* \begin{array}{cclcl}
* m & : & X^2 & \mapsto & R_{+} \\
* & & (x,y) & \mapsto & m^{\theta}(x,y) ~. \\
* \end{array}
* \f]
*
* After migration, the number of individuals in deme \f$x\f$ is defined by
* the total number of individuals converging to \f$x\f$:
*
* \f[
* N(x,t+1) = \displaystyle \sum_{i\in X} \Phi_{i,x}^{t}~.
* \f]
* \section Example
* \snippet demography/test/History/Individual_based_history_test.cpp Example
* \section Output
* \include demography/test/History/Individual_based_history_test.output
*/
struct individual_based {
using value_type = unsigned int;
};
/*!
* \brief Traits class for simulating the demographic history of importnat masses of populations.
* \ingroup demography
*
* \details Simulate the demographic history by considering that populations of individuals are divisible masses,
* leading to faster simulations.
*
* \par Intialization:
* \f$N\f$ is initialized by setting \f$N(.,t_0)\f$ the initial distribution
* of individuals across demes at the first time \f$t_0\f$.
* Typically for a biological invasion, this is restricted to the introduction site(s)
* with the number of introduced individuals. For endemic species, paleoclimatic
* distribution can be considered as starting points.
*
* \par Growth:
* The offspring number \f$\tilde{N}_{x}^{t}\f$ in each deme is freely defined by
* the user. For example, it can be sampled in a distribution conditionally
* to a function of the local density of parents. Non-overlapping generations
* are considered (the parents die just after reproduction).
*
* \par Dispersal:
* The children dispersal is done by splitting the population masses according
* to their migration probabilities, defining
* \f$ \Phi_{x,y}^t \f$, the population flow going from \f$x\f$ to \f$y\f$ at time \f$t\f$:
* \f[ (\Phi_{x,y}^{t})_{y\in X} = (\tilde{N}_{x}^{t} \times m_{xy})_{y\in X} ~. \f]
* The term \f$ m_{xy} \f$ denotes the parameters of the transition kernel,
* giving for an individual in \f$x\f$ its probability to go to \f$y\f$.
* These probabilities are given by the dispersal law with parameter \f$\theta\f$:
* \f[
* \begin{array}{cclcl}
* m & : & X^2 & \mapsto & R_{+} \\
* & & (x,y) & \mapsto & m^{\theta}(x,y) ~. \\
* \end{array}
* \f]
* After migration, the population size in deme \f$x\f$ is defined by the sum of population flows converging to \f$x\f$:
* \f[
* N(x,t+1) = \displaystyle \sum_{i\in X} \Phi_{i,x}^{t}~.
* \f]
* \section Example
* \snippet demography/test/History/Mass_based_history_test.cpp Example
* \section Output
* \include demography/test/History/Mass_based_history_test.output
*/
struct mass_based {
using value_type = double;
};
}
/*!
* \brief Base class for spatially explicit and forward-in time population history simulators.
*
* \tparam Space Demes identifiers.
* \tparam Time EqualityComparable, CopyConstructible.
* \tparam Strategy Strategy use for simulating populations dynamics
*
* \details Is used as an implementation base of the specialized simulation strategies.
*
* \ingroup demography
*
*/
template<typename Space, typename Time, typename Strategy>
class BaseHistory {
public:
//! \typedef strategy used for simulating populations dynamics
using strategy_type = Strategy;
//! \typedef type of the population flows database
using flow_type = Flow<Space, Time, typename strategy_type::value_type>;
//! \typedef type of the population size database
using pop_sizes_type = PopulationSize<Space, Time, typename strategy_type::value_type>;
//! \typedef space type
using coord_type = Space;
//! \typedef time type
using time_type = Time;
//! \typedef type of the discrete distribution used inside the backward dispersal kernel
using discrete_distribution_type = quetzal::random::DiscreteDistribution<coord_type>;
//! \typedef Backward dispersal kernel type
using backward_kernel_type = quetzal::random::TransitionKernel<time_type, discrete_distribution_type>;
protected:
// Need to be accessed by the expand method
std::unique_ptr<pop_sizes_type> m_sizes;
std::unique_ptr<flow_type> m_flows;
std::vector<Time> m_times;
std::unique_ptr<backward_kernel_type> m_kernel;
private:
auto make_backward_distribution(coord_type const& x, time_type const& t) const
{
std::vector<double> weights;
std::vector<coord_type> support;
weights.reserve(m_flows->flux_to(x,t).size());
support.reserve(m_flows->flux_to(x,t).size());
for(auto const& it : m_flows->flux_to(x,t) )
{
support.push_back(it.first);
weights.push_back(static_cast<double>(it.second));
}
return discrete_distribution_type(std::move(support),std::move(weights));
}
public:
/**
* \brief Constructor initializing the demographic database.
* \param x the coordinate of introduction
* \param t the introduction time
* \param N the population size at coordinate x at time t
* \section Example
* \snippet demography/test/History/History_test.cpp Example
* \section Output
* \include demography/test/History/History_test.output
*/
BaseHistory(coord_type const& x, time_type const& t, typename strategy_type::value_type N):
m_sizes(std::make_unique<pop_sizes_type>()),
m_flows(std::make_unique<flow_type>()),
m_kernel(std::make_unique<backward_kernel_type>())
{
m_sizes->operator()(x,t) = N;
m_times.push_back(t);
}
/**
* \brief Read-only access to the demographic flows database
*/
flow_type const& flows() const
{
return *m_flows;
}
/**
* \brief Read and write access to the demographic flows database
*/
flow_type & flows()
{
return *m_flows;
}
/**
* \brief Read-only access to the demographic sizes database.
* \remark Can be used for composition into time dependent growth functions.
*/
const pop_sizes_type & pop_sizes() const
{
return *m_sizes;
}
/**
* \brief Read-and-write access to the demographic sizes database
*/
pop_sizes_type & pop_sizes()
{
return *m_sizes;
}
/**
* \brief First time recorded in the foward-in-time database history.
*/
time_type const& first_time() const
{
return m_times.front();
}
/**
* \brief Last time recorded in the foward-in-time database history.
*/
time_type const& last_time() const
{
return m_times.back();
}
/**
* \brief Samples a coordinate from the backward-in-time transition matrix
*
* \details The transition matrix is computed from the demographic flows
* database. The returned coordinate object will basically answer the question:
* when an individual is found in \f$x\f$ at time \f$t\f$, where could it
* have been at time \f$t-1\f$ ?
* Let \f$ X \f$ be the geographic space and \f$\Phi_{x,y}^t\f$ be the number of individuals going
* from coordinate \f$x \in X \f$ to coordinate \f$y\f$ at time \f$t\f$.
* Knowing that a child node \f$c\f$ is found in \f$ j \in X \f$, the probability for its parent
* \f$p\f$ to be in \f$i\in X \f$ is: \f$ P( p \in i ~|~ e \in j) = \frac{\Phi_{i, j}^{t}}{ \sum_{k} \Phi_{k, j}^{t} } ~.\f$
* \section Example
* \snippet demography/test/History/Mass_based_history_test.cpp Example
* \section Output
* \include demography/test/History/Mass_based_history_test.output
*/
template<typename Generator>
coord_type backward_kernel(coord_type const& x, time_type t, Generator& gen)
{
--t;
assert(m_flows->flux_to_is_defined(x,t));
if( ! m_kernel->has_distribution(x, t))
{
m_kernel->set(x, t, make_backward_distribution(x, t));
}
return m_kernel->operator()(gen, x, t);
}
};
/**
* \brief Demographic history simulated from an individual-based strategy (each individual is dispersed individually).
*
* \tparam Space Demes identifiers.
* \tparam Time EqualityComparable, CopyConstructible.
* \tparam Strategy Strategy used for simulating populations dynamics
*
* \details Inherit from this class and specialize the Strategy template parameter
*
* \ingroup demography
*
*/
template<typename Space, typename Time, typename Strategy>
class History : public BaseHistory<Space, Time, Strategy>
{
};
/**
* \brief Demographic history simulated from an individual-based strategy (each individual is dispersed individually).
*
* \tparam Space Demes identifiers.
* \tparam Time EqualityComparable, CopyConstructible.
*
* \ingroup demography
*
* \details $N$ is initialized by setting \f$ N(.,t_0) \f$ the initial distribution
* of individuals across demes at the first time \f$ t_0 \f$.
* Typically for a biological invasion, this is restricted to the introduction site(s)
* with the number of introduced individuals. For endemic species, paleoclimatic
* distribution can be considered as starting points. The number of descendants
* \f$ \tilde{N}_{x}^{t} \f$ in each deme is sampled in a distribution conditionally
* to a function of the the local density of parents. Non-overlapping generations
* are considered (the parents die just after reproduction). The children dispersal
* is done by sampling their destination in a multinomial law, that defines
* \f$ \Phi_{x,y}^t \f$ the number of individuals going from \f$ x \f$ to \f$ y \f$ at time \f$ t \f$:
* \f[ (\Phi_{x,y}^{t})_{y\in X} \sim \mathcal{M}(\tilde{N}_{x}^{t},(m_{xy})_y) ~. \f]
* The term \f$ (m_{xy})_y \f$ denotes the parameters of the multinomial law,
* giving for an individual in \f$x\f$ its proability to go to \f$y\f$.
* These probabilities are given by the dispersal law with parameter \f$\theta\f$:
* \f[
* \begin{array}{cclcl}
* m & : & X^2 & \mapsto & R_{+} \\
* & & (x,y) & \mapsto & m^{\theta}(x,y) ~. \\
* \end{array}
* \f]
* After migration, the number of individuals in deme \f$x\f$ is defined by the total number of individuals converging to \f$x\f$:
* \f[
* N(x,t+1) = \displaystyle \sum_{i\in X} \Phi_{i,x}^{t}~.
* \f\]
*
* \section Example
* \snippet demography/test/History/History_test.cpp Example
* \section Output
* \include demography/test/History/History_test.output
*/
template<typename Space, typename Time>
class History<Space, Time, strategy::individual_based> : public BaseHistory<Space, Time, strategy::individual_based>
{
using BaseHistory<Space, Time, strategy::individual_based>::BaseHistory;
public:
/**
* \brief Expands the demographic database.
*
* \tparam Space Demes identifiers.
* \tparam Time EqualityComparable, CopyConstructible.
*
* \exception std::domain_error if the population goes extincted before the simulation is completed.
*
* \param nb_generations the number of generations to simulate
* \param sim_growth a functor simulating \f$\tilde{N}_{x}^{t}\f$.
*
* The functor can possibly internally use a reference on the population sizes database to represent the time dependency.
* The signature of the function should be equivalent to the following:
* `unsigned int sim_growth(Generator &gen, const coord_type &x, const time_type &t);`
*
* \param kernel a functor representing the dispersal location kernel that simulates the coordinate of the next location conditionally to the current location.
* The signature should be equivalent to `coord_type kernel(Generator &gen, const coord_type &x, const time_type &t);`
*
* \section Example
* \snippet demography/test/History/History_test.cpp Example
* \section Output
* \include demography/test/History/History_test.output
*/
template<typename Growth, typename Dispersal, typename Generator>
void expand(unsigned int nb_generations, Growth sim_growth, Dispersal kernel, Generator& gen)
{
for(unsigned int g = 0; g < nb_generations; ++g)
{
auto t = this->last_time();
auto t_next = t; ++ t_next;
this->m_times.push_back(t_next);
unsigned int landscape_individuals_count = 0;
// TODO : optimize the definition_space function (for loop)
for(auto x : this->m_sizes->definition_space(t) )
{
auto N_tilde = sim_growth(gen, x, t);
landscape_individuals_count += N_tilde;
if(N_tilde >= 1)
{
for(unsigned int ind = 1; ind <= N_tilde; ++ind)
{
auto y = kernel(gen, x, t);
this->m_flows->add_to_flux_from_to(x, y, t, 1);
this->m_sizes->operator()(y, t_next) += 1;
}
}
}
if(landscape_individuals_count == 0)
{
throw std::domain_error("Landscape populations went extinct before sampling");
}
}
}
};
/**
* \brief Demographic history where populations levels are assumed high enough to be considered as divisible masses.
*
* \ingroup demography
*
* \tparam Space Demes identifiers.
* \tparam Time EqualityComparable, CopyConstructible.
*
* \details $N$ is initialized by setting \f$N(.,t_0)\f$ the initial distribution
* of individuals across demes at the first time \f$t_0\f$.
* Typically for a biological invasion, this is restricted to the introduction site(s)
* with the number of introduced individuals. For endemic species, paleoclimatic
* distribution can be considered as starting points. The number of descendants
* \f$\tilde{N}_{x}^{t}\f$ in each deme is sampled in a distribution conditionally
* to a function of the the local density of parents. Non-overlapping generations
* are considered (the parents die just after reproduction). The children dispersal
* is done by sampling their destination in a multinomial law, that defines
* \f$ \Phi_{x,y}^t \f$ the number of individuals going from \f$x\f$ to \f$y\f$ at time \f$t\f$:
* \f[ (\Phi_{x,y}^{t})_{y\in X} \sim \mathcal{M}(\tilde{N}_{x}^{t},(m_{xy})_y) ~. \f]
* The term \f$ (m_{xy})_y \f$ denotes the parameters of the multinomial law,
* giving for an individual in \f$x\f$ its probability to go to \f$y\f$.
* These probabilities are given by the dispersal law with parameter \f$\theta\f$:
* \f[
* \begin{array}{cclcl}
* m & : & X^2 & \mapsto & R_{+} \\
* & & (x,y) & \mapsto & m^{\theta}(x,y) ~. \\
* \end{array}
* \f]
* After migration, the number of individuals in deme \f$x\f$ is defined by the total number of individuals converging to \f$x\f$:
* \f[
* N(x,t+1) = \displaystyle \sum_{i\in X} \Phi_{i,x}^{t}~.
* \f\]
*
* \section Example
* \snippet demography/test/History/History_test.cpp Example
* \section Output
* \include demography/test/History/History_test.output
*/
template<typename Space, typename Time>
class History<Space, Time, strategy::mass_based> : public BaseHistory<Space, Time, strategy::mass_based>{
using BaseHistory<Space, Time, strategy::mass_based>::BaseHistory;
public:
/**
* \brief Expands the demographic database,
* \details $N$ is initialized by setting \f$N(.,t_0)\f$ the initial distribution
* of individuals across demes at the first time \f$t_0\f$.
* Typically for a biological invasion, this is restricted to the introduction site(s)
* with the number of introduced individuals. For endemic species, paleoclimatic
* distribution can be considered as starting points. The number of descendants
* \f$\tilde{N}_{x}^{t}\f$ in each deme is sampled in a distribution conditionally
* to a function of the the local density of parents. Non-overlapping generations
* are considered (the parents die just after reproduction). The children dispersal
* is done by sampling their destination in a multinomial law, that defines
* \f$ \Phi_{x,y}^t \f$ the population flow going from \f$x\f$ to \f$y\f$ at time \f$t\f$:
* \f[ (\Phi_{x,y}^{t})_{y\in X} = (\tilde{N}_{x}^{t}*m_{xy})_{y\in X} ~. \f]
* The term \f$ m_{xy} \f$ denotes the parameters of the transition kernel,
* giving for an individual in \f$x\f$ its probability to go to \f$y\f$.
* These probabilities are given by the dispersal law with parameter \f$\theta\f$:
* \f[
* \begin{array}{cclcl}
* m & : & X^2 & \mapsto & R_{+} \\
* & & (x,y) & \mapsto & m^{\theta}(x,y) ~. \\
* \end{array}
* \f]
* After migration, the population size in deme \f$x\f$ is defined by the sum of population flows converging to \f$x\f$:
* \f[
* N(x,t+1) = \displaystyle \sum_{i\in X} \Phi_{i,x}^{t}~.
* \f\]
* \param nb_generations the number of generations to simulate
* \param sim_growth a functor simulating \f$\tilde{N}_{x}^{t}\f$.
* The functor can possibly internally use a reference on the population sizes database to represent the time dependency.
* The signature of the function should be equivalent to the following:
* `double sim_growth(Generator &gen, const coord_type &x, const time_type &t);`
* \param kernel a functor representing the dispersal location kernel that simulates the coordinate of the next location conditionally to the current location.
* The signature should be equivalent to `coord_type kernel(Generator &gen, const coord_type &x, const time_type &t);`.
* The expression `kernel.support(time_type const& t)` must be valid and return an iterable container of geographic coordinates
* indicating the transition kernel state space at time \f$t\f$.
* \section Example
* \snippet demography/test/History/History_test.cpp Example
* \section Output
* \include demography/test/History/History_test.output
*/
template<typename Growth, typename Dispersal, typename Generator>
void expand(unsigned int nb_generations, Growth sim_growth, Dispersal kernel, Generator& gen)
{
for(unsigned int g = 0; g < nb_generations; ++g)
{
auto t = this->last_time();
auto t_next = t; ++ t_next;
this->m_times.push_back(t_next);
for(auto x : this->m_sizes->definition_space(t) )
{
auto N_tilde = sim_growth(gen, x, t);
for(auto y : kernel.state_space(t) )
{
auto m = kernel(x, y, t);
this->m_flows->set_flux_from_to(x, y, t, m*N_tilde);
this->m_sizes->operator()(y, t_next) += m*N_tilde;
}
}
}
}
};
} // namespace demography
} // namespace quetzal
#endif |
Heritage + specialisation partielle de patron + doxygen
Le débutant, lui, ignore qu'il ignore à ce point, il est fier de ses premiers succès, bien plus qu'il n'est conscient de l'étendue de ce qu'il ne sait pas, dès qu'il progresse en revanche, dès que s'accroît ce qu'il sait, il commence à saisir tout ce qui manque encore à son savoir. Qui sait peu ignore aussi très peu. [Roger Pol-Droit] 
Répondre avec citation
A méditer: La solution la plus simple est toujours la moins compliquée
Bon, j'attend un peu avant de mettre le sujet en résolu mais apparemment c'est tout bon ! 
).
) mais j'ai bon espoir de voir la communauté évoluer dans les années à venir (on n'aura pas trop le choix de toute manière, vu le chemin et l'importance que prend la recherche computationnelle en biologie). En Angleterre, ils ont même commencer à monter une communauté (Research Software Engineers) pour se fédérer et pouvoir communiquer sur l'importance et les particularités de notre job, qui échappe un peu à l'intuition du chercheur traditionnel, et essayer de lancer des réseaux de recrutement plus adaptés.
Envoyé par Seabirds
Mais en tout cas je vois ce que tu veux dire: les cartouches c'est pour la casquette dev, point barre et tant pis si la casquette bio bloque sur les mots "politique" ou "spécialisation partielle de template": les cartouches sont pas faites pour ça ! Capito !
C'est super dur de vendre ses priorités domaine-spécifiques a l'autre domaine 
Partager