Methods

Summary

method

class parent

truncated documentation

__add__

Point

ajoute un vecteur a celui-ci

__eq__

Point

retourne True si les deux points self et a sont egaux, False sinon

__eq__

NeuralTreeNode

__getitem__

GraphDistance

returns a vertex or an edge if no vertex with the given index was found

__getitem__

NeuralTreeNet

Retrieves node and attributes for node i.

__getitem__

CompletionSystem

Returns elements[i].

__getstate__

NeuralTreeNode

usual

__iadd__

Point

ajoute un vecteur à celui-ci

__init__

Edge

constructor

__init__

GraphDistance

constructor

__init__

Vertex

constructor

__init__

InformationPoint

constructeur, initialisation

__init__

LigneGradient

constructeur

__init__

SegmentNFA

segment + nombre de fausses alarmes

__init__

SegmentBord_Commun

constructeur, definit la definition de l’image

__init__

SegmentBord

initialise les dimensions et fait sorte que la classe contienne le premier segment

__init__

Point

constructeur

__init__

Segment

constructeur, pour éviter des erreurs d’etourderie, on crée des copies des extrémités a et b, comme ce sont …

__init__

NeuralTreeNode

__init__

NuagePoints

constructeur

__init__

NuagePointsLaesa

Construit la classe

__init__

MlGridBenchMark

__init__

NeuralTreeNet

__init__

ROC

Initialisation with a dataframe and two or three columns:

__init__

CompletionTrieNode

__init__

CompletionElement

constructor

__init__

CompletionSystem

fill the completion system

__init__

BaseOptimizer

__init__

SGDOptimizer

__iter__

CompletionTrieNode

Iterates on all nodes (sorted).

__iter__

CompletionSystem

Iterates over elements.

__len__

LigneGradient

Retourne le nombre de pixels dans le segment, peut etre different de la liste self.info_ligne, self.nb

__len__

NeuralTreeNet

Returns the number of nodes

__len__

ROC

usual

__len__

CompletionSystem

Number of elements.

__lt__

SegmentNFA

__repr__

Edge

usual

__repr__

GraphDistance

usual

__repr__

Vertex

usual

__repr__

Point

usuel

__repr__

NeuralTreeNode

usual

__repr__

NeuralTreeNet

usual

__repr__

ROC

Shows first elements, precision rate.

__repr__

CompletionElement

usual

__setstate__

NeuralTreeNode

usual

__str__

Edge

usual

__str__

GraphDistance

usual

__str__

Vertex

usual

__str__

InformationPoint

permet d’afficher cette classe

__str__

SegmentNFA

permet d’afficher ce segment

__str__

SegmentBord_Commun

permet d’afficher le segment

__str__

SegmentBord

permet d’afficher le segment

__str__

Point

permet d’afficher un point avec l’instruction print

__str__

Segment

permet d’afficher le segment avec l’instruction print

__str__

ROC

Shows first elements, precision rate.

__str__

CompletionTrieNode

usual

__sub__

Point

soustraction de deux de vecteurs

_add

CompletionTrieNode

Adds a child.

_common_loss_dloss

NeuralTreeNode

Common beginning to methods loss, dlossds, dlossdw.

_common_loss_dloss

NeuralTreeNet

Common beginning to methods loss, dlossds, dlossdw.

_display_progress

BaseOptimizer

Displays training progress.

_display_progress

SGDOptimizer

Displays training progress.

_evaluate_early_stopping

BaseOptimizer

_evaluate_early_stopping

SGDOptimizer

_get_output_node_attr

NeuralTreeNet

Retrieves the output nodes. nb_last is the number of expected outputs.

_get_updates

BaseOptimizer

_get_updates

SGDOptimizer

Gets the values used to update params with given gradients.

_predict

NeuralTreeNode

Computes inputs of the activation function.

_predict_one

NeuralTreeNet

_private__init__

GraphDistance

_private_string_path_matching

GraphDistance

_regularize_gradient

BaseOptimizer

Applies regularization.

_regularize_gradient

SGDOptimizer

Applies regularization.

_set_fcts

NeuralTreeNode

_update_members

NeuralTreeNet

Updates internal members.

all_completions

CompletionTrieNode

Retrieves all completions for a node, the method does not need precompute_stat() to be run first.

all_mks_completions

CompletionTrieNode

Retrieves all completions for a node, the method assumes precompute_stat() was run.

angle

Point

retourne l’angle du vecteur

append

NeuralTreeNet

Appends a node into the graph.

arrondi

Point

retourne les coordonnées arrondies à l’entier le plus proche

as_array

Point

Convertit en array.

auc

ROC

Computes the area under the curve (:epkg:`AUC`).

auc_interval

ROC

Determines a confidence interval for the :epkg:`AUC` with bootstrap.

bench_experiment

MlGridBenchMark

Calls meth fit.

calcul_bord2

SegmentBord

calcule précisément la second extrémité, parcourt la demi-droite jusqu’à sortir de l’image, le dernier point est …

calcul_vecteur

SegmentBord

En fonction de l’angle, calcule le vecteur direction du segment, ensuite fixe la première extrémité du segment self.a

calcul_vecteur_fin

SegmentBord

propose une seconde extrémité connaissant la première, beaucoup plus loin en conservant la meme orientation, …

calcule_NFA

LigneGradient

ext[ij[0]]: premier indice du segment, ext[ij[1]]: dernier indice du segment, calcule le nombre …

clean_dead_ends

GraphDistance

clear

NeuralTreeNet

Clear all nodes

common_paths

GraphDistance

compare_with_trie

CompletionSystem

Compares the results with the other implementation.

compute_metrics

CompletionSystem

Computes the metric for the completion itself.

compute_predecessor

GraphDistance

usual

compute_roc_curve

ROC

Computes a ROC curve with nb points avec nb, if nb == -1, there are as many as points as the data contains, …

compute_successor

GraphDistance

confusion

ROC

Computes the confusion matrix for a specific score or all if score is None.

connect_root_and_leave

GraphDistance

copy

SegmentBord_Commun

Copie l’instance.

copy

NeuralTreeNet

decoupe_gradient

SegmentBord_Commun

Pour un segment donne joignant deux bords de l’image, cette fonction récupère le gradient et construit une liste …

directeur

SegmentBord

retourne une copie du vecteur directeur

directeur

Segment

retourne le vecteur directeur du segment, ce vecteur est norme

distance

NuagePoints

Retourne une distance entre deux éléments.

distance_matching_graphs_paths

GraphDistance

Computes an alignment between two graphs.

dlossds

_TrainingAPI

Computes the loss derivative due to prediction error.

dlossds

NeuralTreeNode

Computes the loss derivative due to prediction error.

dlossds

NeuralTreeNet

Computes the loss derivative against the inputs.

draw_vertices_edges

GraphDistance

edit_distance_path

GraphDistance

Tries to align two paths from two graphs.

end

MlGridBenchMark

nothing to do

enumerate_all_paths

GraphDistance

enumerate_test_metric

CompletionSystem

Evaluates the completion set on a set of queries, the function returns a list of CompletionElement

extremite

LigneGradient

Comptabilise les indices des extremites possibles, les pixels choisis ont un gradient de la bonne orientation.

fill_cache

_TrainingAPI

Creates a cache with intermediate results.

fill_cache

NeuralTreeNode

Creates a cache with intermediate results. lX is the results before the activation function, aX

fill_cache

NeuralTreeNet

Creates a cache with intermediate results.

find

CompletionTrieNode

Returns the node which holds all completions starting with a given prefix.

find

CompletionSystem

Not very efficient, finds an item in a the list.

first

Segment

Retourne la première extrémité.

fit

_TrainingAPI

Fits a neuron.

fit

NuagePoints

Follows sklearn API.

fit

NuagePointsLaesa

Follows sklearn API.

fit

MlGridBenchMark

Trains a model.

get_matching_functions

GraphDistance

returns default matching functions between two vertices and two edges

get_order_vertices

GraphDistance

gradient

_TrainingAPI

Computes the gradient in X knowing the expected value y.

gradient_backward

_TrainingAPI

Computes the gradient in X.

gradient_backward

NeuralTreeNode

Computes the gradients at point X.

gradient_backward

NeuralTreeNet

Computes the gradient in X.

graphs

MlGridBenchMark

Plots multiples graphs.

has_aligned_point

LigneGradient

Dit s’il existe des points alignés sur le segment.

init_metrics

CompletionElement

initiate the metrics

is_edge

Edge

returns True

is_edge

Vertex

returns False

is_vertex

Edge

returns False

is_vertex

Vertex

returns True

items

CompletionTrieNode

Iterates on children, iterates on weight, key, child.

items

CompletionSystem

Iterates on (e.value, e).

items_list

CompletionTrieNode

All children nodes inluding itself in a list.

iter_leaves

CompletionTrieNode

Iterators on leaves sorted per weight, yield weight, value.

iteration_ends

BaseOptimizer

Performs update to learning rate and potentially other states at the end of an iteration.

iteration_ends

SGDOptimizer

Performs updates to learning rate and potential other states at the end of an iteration.

kneighbors

NuagePoints

Return the k nearest neighbors.

label

NuagePoints

Retourne le label de l’object d’indice i.

last

Segment

Retourne la seconde extrémité.

leaves

CompletionTrieNode

Iterators on leaves.

loss

_TrainingAPI

Computes the loss. Returns a float.

loss

NeuralTreeNode

Computes the loss. Returns a float.

loss

NeuralTreeNet

Computes the loss due to prediction error. Returns a float.

loss_regularization

BaseOptimizer

loss_regularization

SGDOptimizer

milieu

SegmentBord

Un autre segment, pour débugger le programme, choisit une orientation pour laquelle on sait que le résultat …

min_dynamic_keystroke

CompletionTrieNode

Returns the dynamic minimum keystrokes for a word.

min_dynamic_keystroke2

CompletionTrieNode

Returns the modified dynamic minimum keystrokes for a word.

min_keystroke

CompletionTrieNode

Returns the minimum keystrokes for a word without optimisation, this function should be used if you only have a …

min_keystroke0

CompletionTrieNode

Returns the minimum keystrokes for a word.

next

SegmentBord

passe au segment suivant dans le parcours de l’image

next_chemin

LigneGradient

Retourne le couple suivant d’extrémités possibles, None, dans le cas contraire.

normal

Segment

retourne le vecteur normal du segment, ce vecteur est norme

normalise

Point

normalise le vecteur, sa norme devient 1

norme

Point

Retourne la norme.

plot

ROC

Plots a ROC curve.

plot_graphs

MlGridBenchMark

Plots all graphs in the same graphs.

ppv

NuagePoints

Retourne l’élément le plus proche de obj et sa distance avec obj.

ppv

NuagePointsLaesa

Retourne l’élément le plus proche de obj et sa distance avec obj, utilise la sélection à l’aide pivots

precision

ROC

Computes the precision.

precompute_stat

CompletionTrieNode

Computes and stores list of completions for each node, computes mks.

predict

NeuralTreeNode

Computes neuron outputs.

predict

NeuralTreeNet

predict_score_experiment

MlGridBenchMark

Calls method score.

premier

SegmentBord

définit le premier segment, horizontal, part du bord gauche

premier_chemin

LigneGradient

Retourne la premiere d’extremite possible.

preprocess_dataset

MlGridBenchMark

Splits the dataset into train and test.

private_count_left_right

GraphDistance

private_kruskal_matrix

GraphDistance

random_cloud

ROC

Resamples among the data.

roc_intersect

ROC

The ROC curve is defined by a set of points. This function interpolates those points to determine …

roc_intersect_interval

ROC

Computes a confidence interval for the value returned by roc_intersect().

scalaire

Point

Calcule le produit scalaire.

scalairek

Point

Mulitplication par un scalaire.

score

MlGridBenchMark

Scores a model.

segments_significatifs

LigneGradient

Comptabilise le nombre de segments significatifs sur une ligne et les mémorise.

selection_pivots

NuagePointsLaesa

Sélectionne nb pivots aléatoirements.

sort_values

CompletionSystem

sort the elements by value

sort_weight

CompletionSystem

Sorts the elements by value.

str_all_completions

CompletionTrieNode

Builds a string with all completions for all prefixes along the paths.

str_all_completions

CompletionElement

builds a string with all completions for all prefixes along the paths, this is only available if parameter …

str_mks

CompletionElement

return a string with metric information

str_mks0

CompletionElement

return a string with metric information

test_metric

CompletionSystem

Evaluates the completion set on a set of queries, the function returns a dictionary with the aggregated metrics …

to_dict

CompletionSystem

Returns a dictionary.

to_dot

NeuralTreeNet

Exports the neural network into dot.

train

BaseOptimizer

Optimizes the coefficients.

train

SGDOptimizer

Optimizes the coefficients.

tuples

CompletionSystem

Iterates on (e.weight, e.value).

unsorted_iter

CompletionTrieNode

Iterates on all nodes.

update_coef

BaseOptimizer

Updates coefficients with given gradient.

update_coef

SGDOptimizer

Updates coefficients with given gradient.

update_metrics

CompletionElement

update the metrics

update_stat_dynamic

CompletionTrieNode

Must be called after precompute_stat() and computes dynamic mks (see Dynamic Minimum Keystroke). …

update_training_weights

_TrainingAPI

Updates weights.

update_training_weights

NeuralTreeNode

Updates weights.

update_training_weights

NeuralTreeNet

Updates weights.