started branch for analysis v4 dev

deprecated old modules
This commit is contained in:
Arthur Lu 2021-05-26 07:27:09 +00:00
parent 8a365b784e
commit 23329da2b5
12 changed files with 21 additions and 1111 deletions

View File

@ -7,10 +7,12 @@
# current benchmark of optimization: 1.33 times faster
# setup:
__version__ = "3.0.2"
__version__ = "3.0.3"
# changelog should be viewed using print(analysis.__changelog__)
__changelog__ = """changelog:
3.0.3:
- fixed spelling of deprecate
3.0.2:
- fixed __all__
3.0.1:
@ -58,7 +60,7 @@ __changelog__ = """changelog:
- cycle sort
- cocktail sort
- tested all sorting algorithms with both lists and numpy arrays
- depreciated sort function from Array class
- deprecated sort function from Array class
- added warnings as an import
2.1.4:
- added sort and search functions to Array class
@ -136,7 +138,7 @@ __changelog__ = """changelog:
1.12.4:
- renamed gliko to glicko
1.12.3:
- removed depreciated code
- removed deprecated code
1.12.2:
- removed team first time trueskill instantiation in favor of integration in superscript.py
1.12.1:
@ -248,10 +250,10 @@ __changelog__ = """changelog:
1.0.0:
- removed c_entities,nc_entities,obstacles,objectives from __all__
- applied numba.jit to all functions
- depreciated and removed stdev_z_split
- deprecated and removed stdev_z_split
- cleaned up histo_analysis to include numpy and numba.jit optimizations
- depreciated and removed all regression functions in favor of future pytorch optimizer
- depreciated and removed all nonessential functions (basic_analysis, benchmark, strip_data)
- deprecated and removed all regression functions in favor of future pytorch optimizer
- deprecated and removed all nonessential functions (basic_analysis, benchmark, strip_data)
- optimized z_normalize using sklearn.preprocessing.normalize
- TODO: implement kernel/function based pytorch regression optimizer
0.9.0:
@ -270,8 +272,8 @@ __changelog__ = """changelog:
- refactors
- bugfixes
0.8.0:
- depreciated histo_analysis_old
- depreciated debug
- deprecated histo_analysis_old
- deprecated debug
- altered basic_analysis to take array data instead of filepath
- refactor
- optimization
@ -319,7 +321,7 @@ __changelog__ = """changelog:
0.3.5:
- major bug fixes
- updated historical analysis
- depreciated old historical analysis
- deprecated old historical analysis
0.3.4:
- added __version__, __author__, __all__
- added polynomial regression

View File

@ -4,9 +4,11 @@
# this should be imported as a python module using 'from tra_analysis import Array'
# setup:
__version__ = "1.0.3"
__version__ = "1.0.4"
__changelog__ = """changelog:
1.0.4:
- fixed spelling of deprecate
1.0.3:
- fixed __all__
1.0.2:
@ -135,8 +137,8 @@ class Array(): # tests on nd arrays independent of basic_stats
return Array(np.transpose(self.array))
def sort(self, array): # depreciated
warnings.warn("Array.sort has been depreciated in favor of Sort")
def sort(self, array): # deprecated
warnings.warn("Array.sort has been deprecated in favor of Sort")
array_length = len(array)
if array_length <= 1:
return array

View File

@ -1,41 +0,0 @@
# Only included for backwards compatibility! Do not update, CorrelationTest is preferred and supported.
import scipy
from scipy import stats
class CorrelationTest:
def anova_oneway(self, *args): #expects arrays of samples
results = scipy.stats.f_oneway(*args)
return {"f-value": results[0], "p-value": results[1]}
def pearson(self, x, y):
results = scipy.stats.pearsonr(x, y)
return {"r-value": results[0], "p-value": results[1]}
def spearman(self, a, b = None, axis = 0, nan_policy = 'propagate'):
results = scipy.stats.spearmanr(a, b = b, axis = axis, nan_policy = nan_policy)
return {"r-value": results[0], "p-value": results[1]}
def point_biserial(self, x,y):
results = scipy.stats.pointbiserialr(x, y)
return {"r-value": results[0], "p-value": results[1]}
def kendall(self, x, y, initial_lexsort = None, nan_policy = 'propagate', method = 'auto'):
results = scipy.stats.kendalltau(x, y, initial_lexsort = initial_lexsort, nan_policy = nan_policy, method = method)
return {"tau": results[0], "p-value": results[1]}
def kendall_weighted(self, x, y, rank = True, weigher = None, additive = True):
results = scipy.stats.weightedtau(x, y, rank = rank, weigher = weigher, additive = additive)
return {"tau": results[0], "p-value": results[1]}
def mgc(self, x, y, compute_distance = None, reps = 1000, workers = 1, is_twosamp = False, random_state = None):
results = scipy.stats.multiscale_graphcorr(x, y, compute_distance = compute_distance, reps = reps, workers = workers, is_twosamp = is_twosamp, random_state = random_state)
return {"k-value": results[0], "p-value": results[1], "data": results[2]} # unsure if MGC test returns a k value

View File

@ -1,25 +0,0 @@
# Only included for backwards compatibility! Do not update, NaiveBayes is preferred and supported.
import sklearn
from sklearn import model_selection, neighbors
from . import ClassificationMetric, RegressionMetric
class KNN:
def knn_classifier(self, data, labels, n_neighbors, test_size = 0.3, algorithm='auto', leaf_size=30, metric='minkowski', metric_params=None, n_jobs=None, p=2, weights='uniform'): #expects *2d data and 1d labels post-scaling
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
model = sklearn.neighbors.KNeighborsClassifier()
model.fit(data_train, labels_train)
predictions = model.predict(data_test)
return model, ClassificationMetric(predictions, labels_test)
def knn_regressor(self, data, outputs, n_neighbors, test_size = 0.3, weights = "uniform", algorithm = "auto", leaf_size = 30, p = 2, metric = "minkowski", metric_params = None, n_jobs = None):
data_train, data_test, outputs_train, outputs_test = sklearn.model_selection.train_test_split(data, outputs, test_size=test_size, random_state=1)
model = sklearn.neighbors.KNeighborsRegressor(n_neighbors = n_neighbors, weights = weights, algorithm = algorithm, leaf_size = leaf_size, p = p, metric = metric, metric_params = metric_params, n_jobs = n_jobs)
model.fit(data_train, outputs_train)
predictions = model.predict(data_test)
return model, RegressionMetric(predictions, outputs_test)

View File

@ -1,43 +0,0 @@
# Only included for backwards compatibility! Do not update, NaiveBayes is preferred and supported.
import sklearn
from sklearn import model_selection, naive_bayes
from . import ClassificationMetric, RegressionMetric
class NaiveBayes:
def guassian(self, data, labels, test_size = 0.3, priors = None, var_smoothing = 1e-09):
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
model = sklearn.naive_bayes.GaussianNB(priors = priors, var_smoothing = var_smoothing)
model.fit(data_train, labels_train)
predictions = model.predict(data_test)
return model, ClassificationMetric(predictions, labels_test)
def multinomial(self, data, labels, test_size = 0.3, alpha=1.0, fit_prior=True, class_prior=None):
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
model = sklearn.naive_bayes.MultinomialNB(alpha = alpha, fit_prior = fit_prior, class_prior = class_prior)
model.fit(data_train, labels_train)
predictions = model.predict(data_test)
return model, ClassificationMetric(predictions, labels_test)
def bernoulli(self, data, labels, test_size = 0.3, alpha=1.0, binarize=0.0, fit_prior=True, class_prior=None):
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
model = sklearn.naive_bayes.BernoulliNB(alpha = alpha, binarize = binarize, fit_prior = fit_prior, class_prior = class_prior)
model.fit(data_train, labels_train)
predictions = model.predict(data_test)
return model, ClassificationMetric(predictions, labels_test)
def complement(self, data, labels, test_size = 0.3, alpha=1.0, fit_prior=True, class_prior=None, norm=False):
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
model = sklearn.naive_bayes.ComplementNB(alpha = alpha, fit_prior = fit_prior, class_prior = class_prior, norm = norm)
model.fit(data_train, labels_train)
predictions = model.predict(data_test)
return model, ClassificationMetric(predictions, labels_test)

View File

@ -1,25 +0,0 @@
# Only included for backwards compatibility! Do not update, RandomForest is preferred and supported.
import sklearn
from sklearn import ensemble, model_selection
from . import ClassificationMetric, RegressionMetric
class RandomForest:
def random_forest_classifier(self, data, labels, test_size, n_estimators, criterion="gini", max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=None, random_state=None, verbose=0, warm_start=False, class_weight=None):
data_train, data_test, labels_train, labels_test = sklearn.model_selection.train_test_split(data, labels, test_size=test_size, random_state=1)
kernel = sklearn.ensemble.RandomForestClassifier(n_estimators = n_estimators, criterion = criterion, max_depth = max_depth, min_samples_split = min_samples_split, min_samples_leaf = min_samples_leaf, min_weight_fraction_leaf = min_weight_fraction_leaf, max_leaf_nodes = max_leaf_nodes, min_impurity_decrease = min_impurity_decrease, bootstrap = bootstrap, oob_score = oob_score, n_jobs = n_jobs, random_state = random_state, verbose = verbose, warm_start = warm_start, class_weight = class_weight)
kernel.fit(data_train, labels_train)
predictions = kernel.predict(data_test)
return kernel, ClassificationMetric(predictions, labels_test)
def random_forest_regressor(self, data, outputs, test_size, n_estimators, criterion="mse", max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features="auto", max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=None, random_state=None, verbose=0, warm_start=False):
data_train, data_test, outputs_train, outputs_test = sklearn.model_selection.train_test_split(data, outputs, test_size=test_size, random_state=1)
kernel = sklearn.ensemble.RandomForestRegressor(n_estimators = n_estimators, criterion = criterion, max_depth = max_depth, min_samples_split = min_samples_split, min_weight_fraction_leaf = min_weight_fraction_leaf, max_features = max_features, max_leaf_nodes = max_leaf_nodes, min_impurity_decrease = min_impurity_decrease, min_impurity_split = min_impurity_split, bootstrap = bootstrap, oob_score = oob_score, n_jobs = n_jobs, random_state = random_state, verbose = verbose, warm_start = warm_start)
kernel.fit(data_train, outputs_train)
predictions = kernel.predict(data_test)
return kernel, RegressionMetric(predictions, outputs_test)

View File

@ -1,391 +0,0 @@
# Only included for backwards compatibility! Do not update, Sort is preferred and supported.
class Sort: # if you haven't used a sort, then you've never lived
def quicksort(self, a):
def sort(array):
less = []
equal = []
greater = []
if len(array) > 1:
pivot = array[0]
for x in array:
if x < pivot:
less.append(x)
elif x == pivot:
equal.append(x)
elif x > pivot:
greater.append(x)
return sort(less)+equal+sort(greater)
else:
return array
return np.array(sort(a))
def mergesort(self, a):
def sort(array):
array = array
if len(array) >1:
middle = len(array) // 2
L = array[:middle]
R = array[middle:]
sort(L)
sort(R)
i = j = k = 0
while i < len(L) and j < len(R):
if L[i] < R[j]:
array[k] = L[i]
i+= 1
else:
array[k] = R[j]
j+= 1
k+= 1
while i < len(L):
array[k] = L[i]
i+= 1
k+= 1
while j < len(R):
array[k] = R[j]
j+= 1
k+= 1
return array
return sort(a)
def introsort(self, a):
def sort(array, start, end, maxdepth):
array = array
if end - start <= 1:
return
elif maxdepth == 0:
heapsort(array, start, end)
else:
p = partition(array, start, end)
sort(array, start, p + 1, maxdepth - 1)
sort(array, p + 1, end, maxdepth - 1)
return array
def partition(array, start, end):
pivot = array[start]
i = start - 1
j = end
while True:
i = i + 1
while array[i] < pivot:
i = i + 1
j = j - 1
while array[j] > pivot:
j = j - 1
if i >= j:
return j
swap(array, i, j)
def swap(array, i, j):
array[i], array[j] = array[j], array[i]
def heapsort(array, start, end):
build_max_heap(array, start, end)
for i in range(end - 1, start, -1):
swap(array, start, i)
max_heapify(array, index=0, start=start, end=i)
def build_max_heap(array, start, end):
def parent(i):
return (i - 1)//2
length = end - start
index = parent(length - 1)
while index >= 0:
max_heapify(array, index, start, end)
index = index - 1
def max_heapify(array, index, start, end):
def left(i):
return 2*i + 1
def right(i):
return 2*i + 2
size = end - start
l = left(index)
r = right(index)
if (l < size and array[start + l] > array[start + index]):
largest = l
else:
largest = index
if (r < size and array[start + r] > array[start + largest]):
largest = r
if largest != index:
swap(array, start + largest, start + index)
max_heapify(array, largest, start, end)
maxdepth = (len(a).bit_length() - 1)*2
return sort(a, 0, len(a), maxdepth)
def heapsort(self, a):
def sort(array):
array = array
n = len(array)
for i in range(n//2 - 1, -1, -1):
heapify(array, n, i)
for i in range(n-1, 0, -1):
array[i], array[0] = array[0], array[i]
heapify(array, i, 0)
return array
def heapify(array, n, i):
array = array
largest = i
l = 2 * i + 1
r = 2 * i + 2
if l < n and array[i] < array[l]:
largest = l
if r < n and array[largest] < array[r]:
largest = r
if largest != i:
array[i],array[largest] = array[largest],array[i]
heapify(array, n, largest)
return array
return sort(a)
def insertionsort(self, a):
def sort(array):
array = array
for i in range(1, len(array)):
key = array[i]
j = i-1
while j >=0 and key < array[j] :
array[j+1] = array[j]
j -= 1
array[j+1] = key
return array
return sort(a)
def timsort(self, a, block = 32):
BLOCK = block
def sort(array, n):
array = array
for i in range(0, n, BLOCK):
insertionsort(array, i, min((i+31), (n-1)))
size = BLOCK
while size < n:
for left in range(0, n, 2*size):
mid = left + size - 1
right = min((left + 2*size - 1), (n-1))
merge(array, left, mid, right)
size = 2*size
return array
def insertionsort(array, left, right):
array = array
for i in range(left + 1, right+1):
temp = array[i]
j = i - 1
while j >= left and array[j] > temp :
array[j+1] = array[j]
j -= 1
array[j+1] = temp
return array
def merge(array, l, m, r):
len1, len2 = m - l + 1, r - m
left, right = [], []
for i in range(0, len1):
left.append(array[l + i])
for i in range(0, len2):
right.append(array[m + 1 + i])
i, j, k = 0, 0, l
while i < len1 and j < len2:
if left[i] <= right[j]:
array[k] = left[i]
i += 1
else:
array[k] = right[j]
j += 1
k += 1
while i < len1:
array[k] = left[i]
k += 1
i += 1
while j < len2:
array[k] = right[j]
k += 1
j += 1
return sort(a, len(a))
def selectionsort(self, a):
array = a
for i in range(len(array)):
min_idx = i
for j in range(i+1, len(array)):
if array[min_idx] > array[j]:
min_idx = j
array[i], array[min_idx] = array[min_idx], array[i]
return array
def shellsort(self, a):
array = a
n = len(array)
gap = n//2
while gap > 0:
for i in range(gap,n):
temp = array[i]
j = i
while j >= gap and array[j-gap] >temp:
array[j] = array[j-gap]
j -= gap
array[j] = temp
gap //= 2
return array
def bubblesort(self, a):
def sort(array):
for i, num in enumerate(array):
try:
if array[i+1] < num:
array[i] = array[i+1]
array[i+1] = num
sort(array)
except IndexError:
pass
return array
return sort(a)
def cyclesort(self, a):
def sort(array):
array = array
writes = 0
for cycleStart in range(0, len(array) - 1):
item = array[cycleStart]
pos = cycleStart
for i in range(cycleStart + 1, len(array)):
if array[i] < item:
pos += 1
if pos == cycleStart:
continue
while item == array[pos]:
pos += 1
array[pos], item = item, array[pos]
writes += 1
while pos != cycleStart:
pos = cycleStart
for i in range(cycleStart + 1, len(array)):
if array[i] < item:
pos += 1
while item == array[pos]:
pos += 1
array[pos], item = item, array[pos]
writes += 1
return array
return sort(a)
def cocktailsort(self, a):
def sort(array):
array = array
n = len(array)
swapped = True
start = 0
end = n-1
while (swapped == True):
swapped = False
for i in range (start, end):
if (array[i] > array[i + 1]) :
array[i], array[i + 1]= array[i + 1], array[i]
swapped = True
if (swapped == False):
break
swapped = False
end = end-1
for i in range(end-1, start-1, -1):
if (array[i] > array[i + 1]):
array[i], array[i + 1] = array[i + 1], array[i]
swapped = True
start = start + 1
return array
return sort(a)

View File

@ -1,170 +0,0 @@
# Only included for backwards compatibility! Do not update, StatisticalTest is preferred and supported.
import scipy
from scipy import stats
class StatisticalTest:
def ttest_onesample(self, a, popmean, axis = 0, nan_policy = 'propagate'):
results = scipy.stats.ttest_1samp(a, popmean, axis = axis, nan_policy = nan_policy)
return {"t-value": results[0], "p-value": results[1]}
def ttest_independent(self, a, b, equal = True, nan_policy = 'propagate'):
results = scipy.stats.ttest_ind(a, b, equal_var = equal, nan_policy = nan_policy)
return {"t-value": results[0], "p-value": results[1]}
def ttest_statistic(self, o1, o2, equal = True):
results = scipy.stats.ttest_ind_from_stats(o1["mean"], o1["std"], o1["nobs"], o2["mean"], o2["std"], o2["nobs"], equal_var = equal)
return {"t-value": results[0], "p-value": results[1]}
def ttest_related(self, a, b, axis = 0, nan_policy='propagate'):
results = scipy.stats.ttest_rel(a, b, axis = axis, nan_policy = nan_policy)
return {"t-value": results[0], "p-value": results[1]}
def ks_fitness(self, rvs, cdf, args = (), N = 20, alternative = 'two-sided', mode = 'approx'):
results = scipy.stats.kstest(rvs, cdf, args = args, N = N, alternative = alternative, mode = mode)
return {"ks-value": results[0], "p-value": results[1]}
def chisquare(self, f_obs, f_exp = None, ddof = None, axis = 0):
results = scipy.stats.chisquare(f_obs, f_exp = f_exp, ddof = ddof, axis = axis)
return {"chisquared-value": results[0], "p-value": results[1]}
def powerdivergence(self, f_obs, f_exp = None, ddof = None, axis = 0, lambda_ = None):
results = scipy.stats.power_divergence(f_obs, f_exp = f_exp, ddof = ddof, axis = axis, lambda_ = lambda_)
return {"powerdivergence-value": results[0], "p-value": results[1]}
def ks_twosample(self, x, y, alternative = 'two_sided', mode = 'auto'):
results = scipy.stats.ks_2samp(x, y, alternative = alternative, mode = mode)
return {"ks-value": results[0], "p-value": results[1]}
def es_twosample(self, x, y, t = (0.4, 0.8)):
results = scipy.stats.epps_singleton_2samp(x, y, t = t)
return {"es-value": results[0], "p-value": results[1]}
def mw_rank(self, x, y, use_continuity = True, alternative = None):
results = scipy.stats.mannwhitneyu(x, y, use_continuity = use_continuity, alternative = alternative)
return {"u-value": results[0], "p-value": results[1]}
def mw_tiecorrection(self, rank_values):
results = scipy.stats.tiecorrect(rank_values)
return {"correction-factor": results}
def rankdata(self, a, method = 'average'):
results = scipy.stats.rankdata(a, method = method)
return results
def wilcoxon_ranksum(self, a, b): # this seems to be superceded by Mann Whitney Wilcoxon U Test
results = scipy.stats.ranksums(a, b)
return {"u-value": results[0], "p-value": results[1]}
def wilcoxon_signedrank(self, x, y = None, zero_method = 'wilcox', correction = False, alternative = 'two-sided'):
results = scipy.stats.wilcoxon(x, y = y, zero_method = zero_method, correction = correction, alternative = alternative)
return {"t-value": results[0], "p-value": results[1]}
def kw_htest(self, *args, nan_policy = 'propagate'):
results = scipy.stats.kruskal(*args, nan_policy = nan_policy)
return {"h-value": results[0], "p-value": results[1]}
def friedman_chisquare(self, *args):
results = scipy.stats.friedmanchisquare(*args)
return {"chisquared-value": results[0], "p-value": results[1]}
def bm_wtest(self, x, y, alternative = 'two-sided', distribution = 't', nan_policy = 'propagate'):
results = scipy.stats.brunnermunzel(x, y, alternative = alternative, distribution = distribution, nan_policy = nan_policy)
return {"w-value": results[0], "p-value": results[1]}
def combine_pvalues(self, pvalues, method = 'fisher', weights = None):
results = scipy.stats.combine_pvalues(pvalues, method = method, weights = weights)
return {"combined-statistic": results[0], "p-value": results[1]}
def jb_fitness(self, x):
results = scipy.stats.jarque_bera(x)
return {"jb-value": results[0], "p-value": results[1]}
def ab_equality(self, x, y):
results = scipy.stats.ansari(x, y)
return {"ab-value": results[0], "p-value": results[1]}
def bartlett_variance(self, *args):
results = scipy.stats.bartlett(*args)
return {"t-value": results[0], "p-value": results[1]}
def levene_variance(self, *args, center = 'median', proportiontocut = 0.05):
results = scipy.stats.levene(*args, center = center, proportiontocut = proportiontocut)
return {"w-value": results[0], "p-value": results[1]}
def sw_normality(self, x):
results = scipy.stats.shapiro(x)
return {"w-value": results[0], "p-value": results[1]}
def shapiro(self, x):
return "destroyed by facts and logic"
def ad_onesample(self, x, dist = 'norm'):
results = scipy.stats.anderson(x, dist = dist)
return {"d-value": results[0], "critical-values": results[1], "significance-value": results[2]}
def ad_ksample(self, samples, midrank = True):
results = scipy.stats.anderson_ksamp(samples, midrank = midrank)
return {"d-value": results[0], "critical-values": results[1], "significance-value": results[2]}
def binomial(self, x, n = None, p = 0.5, alternative = 'two-sided'):
results = scipy.stats.binom_test(x, n = n, p = p, alternative = alternative)
return {"p-value": results}
def fk_variance(self, *args, center = 'median', proportiontocut = 0.05):
results = scipy.stats.fligner(*args, center = center, proportiontocut = proportiontocut)
return {"h-value": results[0], "p-value": results[1]} # unknown if the statistic is an h value
def mood_mediantest(self, *args, ties = 'below', correction = True, lambda_ = 1, nan_policy = 'propagate'):
results = scipy.stats.median_test(*args, ties = ties, correction = correction, lambda_ = lambda_, nan_policy = nan_policy)
return {"chisquared-value": results[0], "p-value": results[1], "m-value": results[2], "table": results[3]}
def mood_equalscale(self, x, y, axis = 0):
results = scipy.stats.mood(x, y, axis = axis)
return {"z-score": results[0], "p-value": results[1]}
def skewtest(self, a, axis = 0, nan_policy = 'propogate'):
results = scipy.stats.skewtest(a, axis = axis, nan_policy = nan_policy)
return {"z-score": results[0], "p-value": results[1]}
def kurtosistest(self, a, axis = 0, nan_policy = 'propogate'):
results = scipy.stats.kurtosistest(a, axis = axis, nan_policy = nan_policy)
return {"z-score": results[0], "p-value": results[1]}
def normaltest(self, a, axis = 0, nan_policy = 'propogate'):
results = scipy.stats.normaltest(a, axis = axis, nan_policy = nan_policy)
return {"z-score": results[0], "p-value": results[1]}

View File

@ -7,10 +7,14 @@
# current benchmark of optimization: 1.33 times faster
# setup:
__version__ = "3.0.0"
__version__ = "4.0.0-dev"
# changelog should be viewed using print(analysis.__changelog__)
__changelog__ = """changelog:
4.0.0:
- deprecated all *_obj.py compatibility modules
- deprecated titanlearn.py
- deprecated visualization.py
3.0.0:
- incremented version to release 3.0.0
3.0.0-rc2:
@ -53,7 +57,6 @@ __all__ = [
]
from . import Analysis as Analysis
from . import Analysis as analysis
from .Array import Array
from .ClassificationMetric import ClassificationMetric
from . import CorrelationTest

View File

@ -1,222 +0,0 @@
# Titan Robotics Team 2022: CUDA-based Regressions Module
# Not actively maintained, may be removed in future release
# Written by Arthur Lu & Jacob Levine
# Notes:
# this module has been automatically inegrated into analysis.py, and should be callable as a class from the package
# this module is cuda-optimized (as appropriate) and vectorized (except for one small part)
# setup:
__version__ = "0.0.4"
# changelog should be viewed using print(analysis.regression.__changelog__)
__changelog__ = """
0.0.4:
- bug fixes
- fixed changelog
0.0.3:
- bug fixes
0.0.2:
-Added more parameters to log, exponential, polynomial
-Added SigmoidalRegKernelArthur, because Arthur apparently needs
to train the scaling and shifting of sigmoids
0.0.1:
-initial release, with linear, log, exponential, polynomial, and sigmoid kernels
-already vectorized (except for polynomial generation) and CUDA-optimized
"""
__author__ = (
"Jacob Levine <jlevine@imsa.edu>",
"Arthur Lu <learthurgo@gmail.com>",
)
__all__ = [
'factorial',
'take_all_pwrs',
'num_poly_terms',
'set_device',
'LinearRegKernel',
'SigmoidalRegKernel',
'LogRegKernel',
'PolyRegKernel',
'ExpRegKernel',
'SigmoidalRegKernelArthur',
'SGDTrain',
'CustomTrain',
'CircleFit'
]
import torch
global device
device = "cuda:0" if torch.cuda.is_available() else "cpu"
#todo: document completely
def set_device(self, new_device):
device=new_device
class LinearRegKernel():
parameters= []
weights=None
bias=None
def __init__(self, num_vars):
self.weights=torch.rand(num_vars, requires_grad=True, device=device)
self.bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.bias]
def forward(self,mtx):
long_bias=self.bias.repeat([1,mtx.size()[1]])
return torch.matmul(self.weights,mtx)+long_bias
class SigmoidalRegKernel():
parameters= []
weights=None
bias=None
sigmoid=torch.nn.Sigmoid()
def __init__(self, num_vars):
self.weights=torch.rand(num_vars, requires_grad=True, device=device)
self.bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.bias]
def forward(self,mtx):
long_bias=self.bias.repeat([1,mtx.size()[1]])
return self.sigmoid(torch.matmul(self.weights,mtx)+long_bias)
class SigmoidalRegKernelArthur():
parameters= []
weights=None
in_bias=None
scal_mult=None
out_bias=None
sigmoid=torch.nn.Sigmoid()
def __init__(self, num_vars):
self.weights=torch.rand(num_vars, requires_grad=True, device=device)
self.in_bias=torch.rand(1, requires_grad=True, device=device)
self.scal_mult=torch.rand(1, requires_grad=True, device=device)
self.out_bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.in_bias, self.scal_mult, self.out_bias]
def forward(self,mtx):
long_in_bias=self.in_bias.repeat([1,mtx.size()[1]])
long_out_bias=self.out_bias.repeat([1,mtx.size()[1]])
return (self.scal_mult*self.sigmoid(torch.matmul(self.weights,mtx)+long_in_bias))+long_out_bias
class LogRegKernel():
parameters= []
weights=None
in_bias=None
scal_mult=None
out_bias=None
def __init__(self, num_vars):
self.weights=torch.rand(num_vars, requires_grad=True, device=device)
self.in_bias=torch.rand(1, requires_grad=True, device=device)
self.scal_mult=torch.rand(1, requires_grad=True, device=device)
self.out_bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.in_bias, self.scal_mult, self.out_bias]
def forward(self,mtx):
long_in_bias=self.in_bias.repeat([1,mtx.size()[1]])
long_out_bias=self.out_bias.repeat([1,mtx.size()[1]])
return (self.scal_mult*torch.log(torch.matmul(self.weights,mtx)+long_in_bias))+long_out_bias
class ExpRegKernel():
parameters= []
weights=None
in_bias=None
scal_mult=None
out_bias=None
def __init__(self, num_vars):
self.weights=torch.rand(num_vars, requires_grad=True, device=device)
self.in_bias=torch.rand(1, requires_grad=True, device=device)
self.scal_mult=torch.rand(1, requires_grad=True, device=device)
self.out_bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.in_bias, self.scal_mult, self.out_bias]
def forward(self,mtx):
long_in_bias=self.in_bias.repeat([1,mtx.size()[1]])
long_out_bias=self.out_bias.repeat([1,mtx.size()[1]])
return (self.scal_mult*torch.exp(torch.matmul(self.weights,mtx)+long_in_bias))+long_out_bias
class PolyRegKernel():
parameters= []
weights=None
bias=None
power=None
def __init__(self, num_vars, power):
self.power=power
num_terms=self.num_poly_terms(num_vars, power)
self.weights=torch.rand(num_terms, requires_grad=True, device=device)
self.bias=torch.rand(1, requires_grad=True, device=device)
self.parameters=[self.weights,self.bias]
def num_poly_terms(self,num_vars, power):
if power == 0:
return 0
return int(self.factorial(num_vars+power-1) / self.factorial(power) / self.factorial(num_vars-1)) + self.num_poly_terms(num_vars, power-1)
def factorial(self,n):
if n==0:
return 1
else:
return n*self.factorial(n-1)
def take_all_pwrs(self, vec, pwr):
#todo: vectorize (kinda)
combins=torch.combinations(vec, r=pwr, with_replacement=True)
out=torch.ones(combins.size()[0]).to(device).to(torch.float)
for i in torch.t(combins).to(device).to(torch.float):
out *= i
if pwr == 1:
return out
else:
return torch.cat((out,self.take_all_pwrs(vec, pwr-1)))
def forward(self,mtx):
#TODO: Vectorize the last part
cols=[]
for i in torch.t(mtx):
cols.append(self.take_all_pwrs(i,self.power))
new_mtx=torch.t(torch.stack(cols))
long_bias=self.bias.repeat([1,mtx.size()[1]])
return torch.matmul(self.weights,new_mtx)+long_bias
def SGDTrain(self, kernel, data, ground, loss=torch.nn.MSELoss(), iterations=1000, learning_rate=.1, return_losses=False):
optim=torch.optim.SGD(kernel.parameters, lr=learning_rate)
data_cuda=data.to(device)
ground_cuda=ground.to(device)
if (return_losses):
losses=[]
for i in range(iterations):
with torch.set_grad_enabled(True):
optim.zero_grad()
pred=kernel.forward(data_cuda)
ls=loss(pred,ground_cuda)
losses.append(ls.item())
ls.backward()
optim.step()
return [kernel,losses]
else:
for i in range(iterations):
with torch.set_grad_enabled(True):
optim.zero_grad()
pred=kernel.forward(data_cuda)
ls=loss(pred,ground_cuda)
ls.backward()
optim.step()
return kernel
def CustomTrain(self, kernel, optim, data, ground, loss=torch.nn.MSELoss(), iterations=1000, return_losses=False):
data_cuda=data.to(device)
ground_cuda=ground.to(device)
if (return_losses):
losses=[]
for i in range(iterations):
with torch.set_grad_enabled(True):
optim.zero_grad()
pred=kernel.forward(data)
ls=loss(pred,ground)
losses.append(ls.item())
ls.backward()
optim.step()
return [kernel,losses]
else:
for i in range(iterations):
with torch.set_grad_enabled(True):
optim.zero_grad()
pred=kernel.forward(data_cuda)
ls=loss(pred,ground_cuda)
ls.backward()
optim.step()
return kernel

View File

@ -1,122 +0,0 @@
# Titan Robotics Team 2022: ML Module
# Written by Arthur Lu & Jacob Levine
# Notes:
# this should be imported as a python module using 'import titanlearn'
# this should be included in the local directory or environment variable
# this module is optimized for multhreaded computing
# this module learns from its mistakes far faster than 2022's captains
# setup:
__version__ = "1.1.1"
#changelog should be viewed using print(analysis.__changelog__)
__changelog__ = """changelog:
1.1.1:
- removed matplotlib import
- removed graphloss()
1.1.0:
- added net, dataset, dataloader, and stdtrain template definitions
- added graphloss function
1.0.1:
- added clear functions
1.0.0:
- complete rewrite planned
- depreciated 1.0.0.xxx versions
- added simple training loop
0.0.x:
-added generation of ANNS, basic SGD training
"""
__author__ = (
"Arthur Lu <arthurlu@ttic.edu>,"
"Jacob Levine <jlevine@ttic.edu>,"
)
__all__ = [
'clear',
'net',
'dataset',
'dataloader',
'train',
'stdtrainer',
]
import torch
from os import system, name
import numpy as np
def clear():
if name == 'nt':
_ = system('cls')
else:
_ = system('clear')
class net(torch.nn.Module): #template for standard neural net
def __init__(self):
super(Net, self).__init__()
def forward(self, input):
pass
class dataset(torch.utils.data.Dataset): #template for standard dataset
def __init__(self):
super(torch.utils.data.Dataset).__init__()
def __getitem__(self, index):
pass
def __len__(self):
pass
def dataloader(dataset, batch_size, num_workers, shuffle = True):
return torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers)
def train(device, net, epochs, trainloader, optimizer, criterion): #expects standard dataloader, whch returns (inputs, labels)
dataset_len = trainloader.dataset.__len__()
iter_count = 0
running_loss = 0
running_loss_list = []
for epoch in range(epochs): # loop over the dataset multiple times
for i, data in enumerate(trainloader, 0):
inputs = data[0].to(device)
labels = data[1].to(device)
optimizer.zero_grad()
outputs = net(inputs)
loss = criterion(outputs, labels.to(torch.float))
loss.backward()
optimizer.step()
# monitoring steps below
iter_count += 1
running_loss += loss.item()
running_loss_list.append(running_loss)
clear()
print("training on: " + device)
print("iteration: " + str(i) + "/" + str(int(dataset_len / trainloader.batch_size)) + " | " + "epoch: " + str(epoch) + "/" + str(epochs))
print("current batch loss: " + str(loss.item))
print("running loss: " + str(running_loss / iter_count))
return net, running_loss_list
print("finished training")
def stdtrainer(net, criterion, optimizer, dataloader, epochs, batch_size):
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
net = net.to(device)
criterion = criterion.to(device)
optimizer = optimizer.to(device)
trainloader = dataloader
return train(device, net, epochs, trainloader, optimizer, criterion)

View File

@ -1,58 +0,0 @@
# Titan Robotics Team 2022: Visualization Module
# Written by Arthur Lu & Jacob Levine
# Notes:
# this should be imported as a python module using 'import visualization'
# this should be included in the local directory or environment variable
# fancy
# setup:
__version__ = "0.0.1"
#changelog should be viewed using print(analysis.__changelog__)
__changelog__ = """changelog:
0.0.1:
- added graphhistogram function as a fragment of visualize_pit.py
0.0.0:
- created visualization.py
- added graphloss()
- added imports
"""
__author__ = (
"Arthur Lu <arthurlu@ttic.edu>,"
"Jacob Levine <jlevine@ttic.edu>,"
)
__all__ = [
'graphloss',
]
import matplotlib.pyplot as plt
import numpy as np
def graphloss(losses):
x = range(0, len(losses))
plt.plot(x, losses)
plt.show()
def graphhistogram(data, figsize, sharey = True): # expects library with key as variable and contents as occurances
fig, ax = plt.subplots(1, len(data), sharey=sharey, figsize=figsize)
i = 0
for variable in data:
ax[i].hist(data[variable])
ax[i].invert_xaxis()
ax[i].set_xlabel('Variable')
ax[i].set_ylabel('Frequency')
ax[i].set_title(variable)
plt.yticks(np.arange(len(data[variable])))
i+=1
plt.show()