[MAJOR UPDATE] Changed additional features to version 4, erased unnecessary input letters (unused vowels), split validation data to test data and validation data
This commit is contained in:
parent
0cc949897f
commit
f0d263e429
|
|
@ -2,8 +2,8 @@
|
|||
<project version="4">
|
||||
<component name="ChangeListManager">
|
||||
<list default="true" id="8a8ba9af-e1a4-433a-9968-475192610776" name="Default" comment="">
|
||||
<change type="MODIFICATION" beforePath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/1_epoch.h5" afterPath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/1_epoch.h5" />
|
||||
<change type="MODIFICATION" beforePath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/1_epoch_history.pkl" afterPath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/1_epoch_history.pkl" />
|
||||
<change type="NEW" beforePath="" afterPath="$PROJECT_DIR$/tex_hyphenation.py" />
|
||||
<change type="MODIFICATION" beforePath="$PROJECT_DIR$/.idea/workspace.xml" afterPath="$PROJECT_DIR$/.idea/workspace.xml" />
|
||||
<change type="MODIFICATION" beforePath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/cnn.ipynb" afterPath="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/cnn.ipynb" />
|
||||
<change type="MODIFICATION" beforePath="$PROJECT_DIR$/prepare_data.py" afterPath="$PROJECT_DIR$/prepare_data.py" />
|
||||
</list>
|
||||
|
|
@ -35,8 +35,8 @@
|
|||
<file leaf-file-name="prepare_data.py" pinned="false" current-in-tab="true">
|
||||
<entry file="file://$PROJECT_DIR$/prepare_data.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="442">
|
||||
<caret line="462" column="19" lean-forward="false" selection-start-line="462" selection-start-column="4" selection-end-line="462" selection-end-column="19" />
|
||||
<state relative-caret-position="284">
|
||||
<caret line="592" column="36" lean-forward="true" selection-start-line="592" selection-start-column="36" selection-end-line="592" selection-end-column="36" />
|
||||
<folding>
|
||||
<element signature="e#24#63#0" expanded="true" />
|
||||
</folding>
|
||||
|
|
@ -44,6 +44,16 @@
|
|||
</provider>
|
||||
</entry>
|
||||
</file>
|
||||
<file leaf-file-name="tex_hyphenation.py" pinned="false" current-in-tab="false">
|
||||
<entry file="file://$PROJECT_DIR$/tex_hyphenation.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="1206">
|
||||
<caret line="67" column="105" lean-forward="false" selection-start-line="67" selection-start-column="105" selection-end-line="67" selection-end-column="105" />
|
||||
<folding />
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
</file>
|
||||
<file leaf-file-name="workbench.py" pinned="false" current-in-tab="false">
|
||||
<entry file="file://$PROJECT_DIR$/workbench.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
|
|
@ -139,6 +149,17 @@
|
|||
<find>StringIO</find>
|
||||
<find>shuffle_inputs</find>
|
||||
<find>generator</find>
|
||||
<find>content, feature_dictionary</find>
|
||||
<find>decode</find>
|
||||
<find>create_feature_dictionary</find>
|
||||
<find>with</find>
|
||||
<find>read</find>
|
||||
<find>generate</find>
|
||||
<find>shuffle</find>
|
||||
<find>X_</find>
|
||||
<find>dictionary</find>
|
||||
<find>create_dict</find>
|
||||
<find>split_content</find>
|
||||
</findStrings>
|
||||
</component>
|
||||
<component name="Git.Settings">
|
||||
|
|
@ -157,6 +178,7 @@
|
|||
<option value="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/character_based_ffnn_keras.py" />
|
||||
<option value="$PROJECT_DIR$/workbench.py" />
|
||||
<option value="$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/character_based_ffnn_keras.ipynb" />
|
||||
<option value="$PROJECT_DIR$/tex_hyphenation.py" />
|
||||
<option value="$PROJECT_DIR$/prepare_data.py" />
|
||||
</list>
|
||||
</option>
|
||||
|
|
@ -165,7 +187,7 @@
|
|||
<option name="x" value="65" />
|
||||
<option name="y" value="24" />
|
||||
<option name="width" value="1855" />
|
||||
<option name="height" value="1056" />
|
||||
<option name="height" value="1176" />
|
||||
</component>
|
||||
<component name="ProjectView">
|
||||
<navigator currentView="ProjectPane" proportions="" version="1">
|
||||
|
|
@ -182,7 +204,6 @@
|
|||
<foldersAlwaysOnTop value="true" />
|
||||
</navigator>
|
||||
<panes>
|
||||
<pane id="Scratches" />
|
||||
<pane id="ProjectPane">
|
||||
<subPane>
|
||||
<PATH>
|
||||
|
|
@ -195,23 +216,10 @@
|
|||
<option name="myItemType" value="com.intellij.ide.projectView.impl.nodes.PsiDirectoryNode" />
|
||||
</PATH_ELEMENT>
|
||||
</PATH>
|
||||
<PATH>
|
||||
<PATH_ELEMENT>
|
||||
<option name="myItemId" value="accetuation" />
|
||||
<option name="myItemType" value="com.intellij.ide.projectView.impl.nodes.ProjectViewProjectNode" />
|
||||
</PATH_ELEMENT>
|
||||
<PATH_ELEMENT>
|
||||
<option name="myItemId" value="accetuation" />
|
||||
<option name="myItemType" value="com.intellij.ide.projectView.impl.nodes.PsiDirectoryNode" />
|
||||
</PATH_ELEMENT>
|
||||
<PATH_ELEMENT>
|
||||
<option name="myItemId" value="cnn" />
|
||||
<option name="myItemType" value="com.intellij.ide.projectView.impl.nodes.PsiDirectoryNode" />
|
||||
</PATH_ELEMENT>
|
||||
</PATH>
|
||||
</subPane>
|
||||
</pane>
|
||||
<pane id="Scope" />
|
||||
<pane id="Scratches" />
|
||||
</panes>
|
||||
</component>
|
||||
<component name="PropertiesComponent">
|
||||
|
|
@ -474,7 +482,7 @@
|
|||
<servers />
|
||||
</component>
|
||||
<component name="ToolWindowManager">
|
||||
<frame x="65" y="24" width="1855" height="1056" extended-state="6" />
|
||||
<frame x="65" y="24" width="1855" height="1176" extended-state="6" />
|
||||
<editor active="true" />
|
||||
<layout>
|
||||
<window_info id="Project" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="true" show_stripe_button="true" weight="0.16375546" sideWeight="0.5" order="0" side_tool="false" content_ui="combo" />
|
||||
|
|
@ -495,25 +503,6 @@
|
|||
<window_info id="Ant Build" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Debug" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.39979124" sideWeight="0.5" order="3" side_tool="false" content_ui="tabs" />
|
||||
</layout>
|
||||
<layout-to-restore>
|
||||
<window_info id="TODO" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="6" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Cvs" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="4" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Message" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="0" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Commander" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.4" sideWeight="0.5" order="0" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Event Log" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="7" side_tool="true" content_ui="tabs" />
|
||||
<window_info id="Inspection" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.4" sideWeight="0.5" order="5" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Version Control" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="10" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Python Console" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.32985386" sideWeight="0.5" order="8" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Run" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="true" show_stripe_button="true" weight="0.32985386" sideWeight="0.5" order="2" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Terminal" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.32985386" sideWeight="0.5" order="9" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Project" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="true" show_stripe_button="true" weight="0.16375546" sideWeight="0.5" order="0" side_tool="false" content_ui="combo" />
|
||||
<window_info id="Hierarchy" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="2" side_tool="false" content_ui="combo" />
|
||||
<window_info id="Find" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Structure" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Ant Build" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
|
||||
<window_info id="Favorites" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="2" side_tool="true" content_ui="tabs" />
|
||||
<window_info id="Debug" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.39979124" sideWeight="0.5" order="3" side_tool="false" content_ui="tabs" />
|
||||
</layout-to-restore>
|
||||
</component>
|
||||
<component name="VcsContentAnnotationSettings">
|
||||
<option name="myLimit" value="2678400000" />
|
||||
|
|
@ -529,17 +518,6 @@
|
|||
<watches-manager />
|
||||
</component>
|
||||
<component name="editorHistoryManager">
|
||||
<entry file="file://$PROJECT_DIR$/theano_tutorial/tutorial_derivates.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="0">
|
||||
<caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
|
||||
<folding>
|
||||
<element signature="e#0#18#0" expanded="true" />
|
||||
</folding>
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/theanoTest.py" />
|
||||
<entry file="file://$PROJECT_DIR$/theano_tutorial/test.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="1368">
|
||||
|
|
@ -822,16 +800,6 @@
|
|||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/cnn_test_on_other_attributes.ipynb" />
|
||||
<entry file="file://$PROJECT_DIR$/prepare_data.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="442">
|
||||
<caret line="462" column="19" lean-forward="false" selection-start-line="462" selection-start-column="4" selection-end-line="462" selection-end-column="19" />
|
||||
<folding>
|
||||
<element signature="e#24#63#0" expanded="true" />
|
||||
</folding>
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/theano_tutorial/logistic_regression.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="162">
|
||||
|
|
@ -882,6 +850,14 @@
|
|||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/theanoTest.py" />
|
||||
<entry file="file://$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/character_based_ffnn_keras.py" />
|
||||
<entry file="file://$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/character_based_ffnn_keras.ipynb" />
|
||||
<entry file="file://$PROJECT_DIR$/hyphenation">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="0">
|
||||
<caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/workbench.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="396">
|
||||
|
|
@ -892,6 +868,23 @@
|
|||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/cnn/word_accetuation/cnn_dictionary/character_based_ffnn_keras.ipynb" />
|
||||
<entry file="file://$PROJECT_DIR$/tex_hyphenation.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="1206">
|
||||
<caret line="67" column="105" lean-forward="false" selection-start-line="67" selection-start-column="105" selection-end-line="67" selection-end-column="105" />
|
||||
<folding />
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
<entry file="file://$PROJECT_DIR$/prepare_data.py">
|
||||
<provider selected="true" editor-type-id="text-editor">
|
||||
<state relative-caret-position="284">
|
||||
<caret line="592" column="36" lean-forward="true" selection-start-line="592" selection-start-column="36" selection-end-line="592" selection-end-column="36" />
|
||||
<folding>
|
||||
<element signature="e#24#63#0" expanded="true" />
|
||||
</folding>
|
||||
</state>
|
||||
</provider>
|
||||
</entry>
|
||||
</component>
|
||||
</project>
|
||||
1113
hyphenation
Normal file
1113
hyphenation
Normal file
File diff suppressed because it is too large
Load Diff
314
prepare_data.py
314
prepare_data.py
|
|
@ -7,6 +7,7 @@ import h5py
|
|||
import gc
|
||||
import math
|
||||
import keras.backend as K
|
||||
import os.path
|
||||
|
||||
|
||||
# functions for saving, loading and shuffling whole arrays to ram
|
||||
|
|
@ -34,9 +35,15 @@ def load_inputs(file_name, other_features=False):
|
|||
return X, y
|
||||
|
||||
|
||||
def shuffle_inputs(X, y, X_pure=[]):
|
||||
s = np.arange(X.shape[0])
|
||||
np.random.shuffle(s)
|
||||
def shuffle_inputs(X, y, shuffle_vector_location, X_pure=[]):
|
||||
if os.path.exists(shuffle_vector_location):
|
||||
s = load_shuffle_vector(shuffle_vector_location)
|
||||
else:
|
||||
s = np.arange(X.shape[0])
|
||||
np.random.shuffle(s)
|
||||
create_and_save_shuffle_vector(shuffle_vector_location, s)
|
||||
# s = np.arange(X.shape[0])
|
||||
# np.random.shuffle(s)
|
||||
X = X[s]
|
||||
y = y[s]
|
||||
if X_pure != []:
|
||||
|
|
@ -57,7 +64,7 @@ def create_and_save_inputs(file_name, part, X, y, X_pure):
|
|||
|
||||
|
||||
def load_extended_inputs(file_name, obtain_range):
|
||||
h5f = h5py.File(file_name,'r')
|
||||
h5f = h5py.File(file_name, 'r')
|
||||
X = h5f['X'][obtain_range[0]:obtain_range[1]]
|
||||
y = h5f['y'][obtain_range[0]:obtain_range[1]]
|
||||
X_pure = h5f['X_pure'][obtain_range[0]:obtain_range[1]]
|
||||
|
|
@ -69,16 +76,17 @@ def load_extended_inputs(file_name, obtain_range):
|
|||
# functions for creating and loading shuffle vector
|
||||
def create_and_save_shuffle_vector(file_name, shuffle_vector):
|
||||
# X, y, X_pure = generate_full_vowel_matrix_inputs()
|
||||
h5f = h5py.File(file_name + '_shuffle_vector.h5', 'w')
|
||||
adict=dict(shuffle_vector=shuffle_vector)
|
||||
h5f = h5py.File(file_name, 'w')
|
||||
adict = dict(shuffle_vector=shuffle_vector)
|
||||
for k, v in adict.items():
|
||||
h5f.create_dataset(k,data=v)
|
||||
h5f.create_dataset(k, data=v)
|
||||
h5f.close()
|
||||
|
||||
|
||||
def load_shuffle_vector(file_name):
|
||||
h5f = h5py.File(file_name,'r')
|
||||
shuffle_vector = h5f['shuffle_vector'][[179859, 385513, 893430]]
|
||||
h5f = h5py.File(file_name, 'r')
|
||||
# shuffle_vector = h5f['shuffle_vector'][[179859, 385513, 893430]]
|
||||
shuffle_vector = h5f['shuffle_vector'][:]
|
||||
|
||||
h5f.close()
|
||||
return shuffle_vector
|
||||
|
|
@ -138,7 +146,8 @@ def create_dict():
|
|||
vowels.extend(accetuated_vowels)
|
||||
vowels.extend(default_vowels)
|
||||
|
||||
dictionary = ['']
|
||||
dictionary_output = ['']
|
||||
dictionary_input = ['']
|
||||
line = 0
|
||||
max_word = 0
|
||||
# ADD 'EMPTY' VOWEL
|
||||
|
|
@ -154,12 +163,12 @@ def create_dict():
|
|||
for c in list(el[3]):
|
||||
if is_vowel(list(el[3]), i, vowels):
|
||||
num_vowels += 1
|
||||
if c not in dictionary:
|
||||
dictionary.append(c)
|
||||
if c not in dictionary_output:
|
||||
dictionary_output.append(c)
|
||||
i += 1
|
||||
for c in list(el[0]):
|
||||
if c not in dictionary:
|
||||
dictionary.append(c)
|
||||
if c not in dictionary_input:
|
||||
dictionary_input.append(c)
|
||||
if num_vowels > max_num_vowels:
|
||||
max_num_vowels = num_vowels
|
||||
except Exception:
|
||||
|
|
@ -167,10 +176,10 @@ def create_dict():
|
|||
print(el)
|
||||
break
|
||||
line += 1
|
||||
dictionary = sorted(dictionary)
|
||||
dictionary_input = sorted(dictionary_input)
|
||||
max_num_vowels += 1
|
||||
print('DICTIONARY CREATION SUCCESSFUL!')
|
||||
return dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels
|
||||
return dictionary_input, max_word, max_num_vowels, content, vowels, accetuated_vowels
|
||||
|
||||
|
||||
# GENERATE X and y
|
||||
|
|
@ -272,7 +281,22 @@ def generate_presentable_y(accetuations_list, word_list, max_num_vowels):
|
|||
# return X, y
|
||||
|
||||
|
||||
def generate_full_matrix_inputs():
|
||||
def generate_full_matrix_inputs(content_shuffle_vector_location, shuffle_vector_location):
|
||||
dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels = create_dict()
|
||||
train_content, test_content, validate_content = split_content(content, 0.2, content_shuffle_vector_location)
|
||||
feature_dictionary = create_feature_dictionary()
|
||||
|
||||
# Generate X and y
|
||||
print('GENERATING X AND y...')
|
||||
X_train, X_other_features_train, y_train = generate_X_and_y(dictionary, max_word, max_num_vowels, train_content, vowels, accetuated_vowels, feature_dictionary, shuffle_vector_location + '_train.h5')
|
||||
X_test, X_other_features_test, y_test = generate_X_and_y(dictionary, max_word, max_num_vowels, test_content, vowels, accetuated_vowels, feature_dictionary, shuffle_vector_location + '_test.h5')
|
||||
X_validate, X_other_features_validate, y_validate = generate_X_and_y(dictionary, max_word, max_num_vowels, validate_content, vowels, accetuated_vowels, feature_dictionary, shuffle_vector_location + '_validate.h5')
|
||||
print('GENERATION SUCCESSFUL!')
|
||||
return X_train, X_other_features_train, y_train, X_test, X_other_features_test, y_test, X_validate, X_other_features_validate, y_validate
|
||||
|
||||
|
||||
# generate full matrix, with old features
|
||||
def old_generate_full_matrix_inputs():
|
||||
dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels = create_dict()
|
||||
train_content, validate_content = split_content(content, 0.2)
|
||||
feature_dictionary = create_feature_dictionary(content)
|
||||
|
|
@ -286,7 +310,7 @@ def generate_full_matrix_inputs():
|
|||
|
||||
|
||||
# Generate each y as an array of 11 numbers (with possible values between 0 and 1)
|
||||
def generate_X_and_y(dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels, feature_dictionary):
|
||||
def generate_X_and_y(dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels, feature_dictionary, shuffle_vector_location):
|
||||
y = np.zeros((len(content), max_num_vowels))
|
||||
X = np.zeros((len(content), max_word, len(dictionary)))
|
||||
print('CREATING OTHER FEATURES...')
|
||||
|
|
@ -328,7 +352,7 @@ def generate_X_and_y(dictionary, max_word, max_num_vowels, content, vowels, acce
|
|||
i += 1
|
||||
|
||||
print('SHUFFELING INPUTS...')
|
||||
X, y, X_other_features = shuffle_inputs(X, y, X_pure=X_other_features)
|
||||
X, y, X_other_features = shuffle_inputs(X, y, shuffle_vector_location, X_pure=X_other_features)
|
||||
print('INPUTS SHUFFELED!')
|
||||
return X, X_other_features, y
|
||||
|
||||
|
|
@ -559,6 +583,7 @@ def shuffle_full_vowel_inputs(name, orderd_name, parts):
|
|||
|
||||
# Decoders for inputs and outputs
|
||||
def decode_X_features(feature_dictionary, X_other_features):
|
||||
final_word = []
|
||||
for word in X_other_features:
|
||||
final_word = []
|
||||
i = 0
|
||||
|
|
@ -574,6 +599,7 @@ def decode_X_features(feature_dictionary, X_other_features):
|
|||
final_word.append(feature_dictionary[z][j][k])
|
||||
i += 1
|
||||
print(u''.join(final_word))
|
||||
return u''.join(final_word)
|
||||
|
||||
|
||||
def decode_position(y, max_num_vowels):
|
||||
|
|
@ -650,7 +676,37 @@ def decode_position_from_vowel_to_final_number(y):
|
|||
|
||||
|
||||
# split content so that there is no overfitting
|
||||
def split_content(content, ratio):
|
||||
def split_content(content, test_and_validation_ratio, content_shuffle_vector_location, validation_ratio=0.5):
|
||||
expanded_content = [el[1] if el[1] != '=' else el[0] for el in content]
|
||||
# print(len(content))
|
||||
unique_content = sorted(set(expanded_content))
|
||||
|
||||
if os.path.exists(content_shuffle_vector_location):
|
||||
s = load_shuffle_vector(content_shuffle_vector_location)
|
||||
else:
|
||||
s = np.arange(len(unique_content))
|
||||
np.random.shuffle(s)
|
||||
create_and_save_shuffle_vector(content_shuffle_vector_location, s)
|
||||
|
||||
split_num = math.floor(len(unique_content) * test_and_validation_ratio)
|
||||
validation_num = math.floor(split_num * validation_ratio)
|
||||
shuffled_unique_train_content = [unique_content[i] for i in range(len(s)) if s[i] >= split_num]
|
||||
shuffled_unique_train_content_set = set(shuffled_unique_train_content)
|
||||
|
||||
shuffled_unique_test_content = [unique_content[i] for i in range(len(s)) if split_num > s[i] >= validation_num]
|
||||
shuffled_unique_test_content_set = set(shuffled_unique_test_content)
|
||||
|
||||
shuffled_unique_validate_content = [unique_content[i] for i in range(len(s)) if s[i] < validation_num]
|
||||
shuffled_unique_validate_content_set = set(shuffled_unique_validate_content)
|
||||
|
||||
train_content = [content[i] for i in range(len(content)) if expanded_content[i] in shuffled_unique_train_content_set]
|
||||
test_content = [content[i] for i in range(len(content)) if expanded_content[i] in shuffled_unique_test_content_set]
|
||||
validate_content = [content[i] for i in range(len(content)) if expanded_content[i] in shuffled_unique_validate_content_set]
|
||||
return train_content, test_content, validate_content
|
||||
|
||||
|
||||
# split content so that there is no overfitting with out split of validation and test data
|
||||
def old_split_content(content, ratio):
|
||||
expanded_content = [el[1] if el[1] != '=' else el[0] for el in content]
|
||||
# print(len(content))
|
||||
unique_content = sorted(set(expanded_content))
|
||||
|
|
@ -671,8 +727,8 @@ def split_content(content, ratio):
|
|||
return train_content, validate_content
|
||||
|
||||
|
||||
# create feature dictionary
|
||||
def create_feature_dictionary(content):
|
||||
# X features that use MULTEX v3 as their encoding
|
||||
def create_old_feature_dictionary(content):
|
||||
additional_data = [el[2] for el in content]
|
||||
possible_variants = sorted(set(additional_data))
|
||||
categories = sorted(set([el[0] for el in possible_variants]))
|
||||
|
|
@ -690,7 +746,8 @@ def create_feature_dictionary(content):
|
|||
return feature_dictionary
|
||||
|
||||
|
||||
def create_X_features(content, feature_dictionary):
|
||||
# X features that use MULTEX v3 as their encoding
|
||||
def create_old_X_features(content, feature_dictionary):
|
||||
content = content
|
||||
X_other_features = []
|
||||
for el in content:
|
||||
|
|
@ -707,4 +764,213 @@ def create_X_features(content, feature_dictionary):
|
|||
else:
|
||||
X_el_other_features.extend([0] * feature[0])
|
||||
X_other_features.append(X_el_other_features)
|
||||
return np.array(X_other_features)
|
||||
return np.array(X_other_features)
|
||||
|
||||
|
||||
def convert_to_MULTEXT_east_v4(old_features, feature_dictionary):
|
||||
new_features = ['-'] * 9
|
||||
new_features[:len(old_features)] = old_features
|
||||
if old_features[0] == 'A':
|
||||
if old_features[1] == 'f' or old_features[1] == 'o':
|
||||
new_features[1] = 'g'
|
||||
return new_features[:len(feature_dictionary[0]) - 1]
|
||||
if old_features[0] == 'C':
|
||||
return new_features[:len(feature_dictionary[1]) - 1]
|
||||
if old_features[0] == 'I':
|
||||
return new_features[:len(feature_dictionary[2]) - 1]
|
||||
if old_features[0] == 'M':
|
||||
new_features[2:6] = old_features[1:5]
|
||||
new_features[1] = old_features[5]
|
||||
if new_features[2] == 'm':
|
||||
new_features[2] = '-'
|
||||
return new_features[:len(feature_dictionary[3]) - 1]
|
||||
if old_features[0] == 'N':
|
||||
if len(old_features) > 5:
|
||||
new_features[5] = old_features[7]
|
||||
return new_features[:len(feature_dictionary[4]) - 1]
|
||||
if old_features[0] == 'P':
|
||||
if new_features[8] == 'n':
|
||||
new_features[8] = 'b'
|
||||
return new_features[:len(feature_dictionary[5]) - 1]
|
||||
if old_features[0] == 'Q':
|
||||
return new_features[:len(feature_dictionary[6]) - 1]
|
||||
if old_features[0] == 'R':
|
||||
return new_features[:len(feature_dictionary[7]) - 1]
|
||||
if old_features[0] == 'S':
|
||||
if len(old_features) == 4:
|
||||
new_features[1] = old_features[3]
|
||||
else:
|
||||
new_features[1] = '-'
|
||||
return new_features[:len(feature_dictionary[8]) - 1]
|
||||
if old_features[0] == 'V':
|
||||
if old_features[1] == 'o' or old_features[1] == 'c':
|
||||
new_features[1] = 'm'
|
||||
new_features[3] = old_features[2]
|
||||
new_features[2] = '-'
|
||||
if old_features[2] == 'i':
|
||||
new_features[3] = 'r'
|
||||
if len(old_features) > 3 and old_features[3] == 'p':
|
||||
new_features[3] = 'r'
|
||||
elif len(old_features) > 3 and old_features[3] == 'f':
|
||||
new_features[3] = 'f'
|
||||
if len(old_features) >= 9:
|
||||
new_features[7] = old_features[8]
|
||||
else:
|
||||
new_features[7] = '-'
|
||||
return new_features[:len(feature_dictionary[9]) - 1]
|
||||
return ''
|
||||
|
||||
|
||||
def create_X_features(content, feature_dictionary):
|
||||
content = content
|
||||
X_other_features = []
|
||||
for el in content:
|
||||
X_el_other_features = []
|
||||
converted_el = ''.join(convert_to_MULTEXT_east_v4(list(el[2]), feature_dictionary))
|
||||
# converted_el = el[2]
|
||||
for feature in feature_dictionary:
|
||||
if converted_el[0] == feature[1]:
|
||||
X_el_other_features.append(1)
|
||||
for i in range(2, len(feature)):
|
||||
for j in range(len(feature[i])):
|
||||
if i-1 < len(converted_el) and feature[i][j] == converted_el[i-1]:
|
||||
X_el_other_features.append(1)
|
||||
else:
|
||||
X_el_other_features.append(0)
|
||||
else:
|
||||
X_el_other_features.extend([0] * feature[0])
|
||||
X_other_features.append(X_el_other_features)
|
||||
return np.array(X_other_features)
|
||||
|
||||
|
||||
def create_feature_dictionary():
|
||||
# old: http://nl.ijs.si/ME/Vault/V3/msd/html/
|
||||
# new: http://nl.ijs.si/ME/V4/msd/html/
|
||||
# changes: http://nl.ijs.si/jos/msd/html-en/msd.diffs.html
|
||||
|
||||
return [[21,
|
||||
'A',
|
||||
['g', 's'],
|
||||
['p', 'c', 's'],
|
||||
['m', 'f', 'n'],
|
||||
['s', 'd', 'p'],
|
||||
['n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[3, 'C', ['c', 's']],
|
||||
[1, 'I'],
|
||||
[21,
|
||||
'M',
|
||||
['l'],
|
||||
['-', 'c', 'o', 's'],
|
||||
['m', 'f', 'n'],
|
||||
['s', 'd', 'p'],
|
||||
['n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[17,
|
||||
'N',
|
||||
['c'],
|
||||
['m', 'f', 'n'],
|
||||
['s', 'd', 'p'],
|
||||
['n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[40,
|
||||
'P',
|
||||
['p', 's', 'd', 'r', 'x', 'g', 'q', 'i', 'z'],
|
||||
['-', '1', '2', '3'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 'y', 'b']],
|
||||
[1, 'Q'],
|
||||
[5, 'R', ['g'], ['p', 'c', 's']],
|
||||
[7, 'S', ['-', 'g', 'd', 'a', 'l', 'i']],
|
||||
[24,
|
||||
'V',
|
||||
['m'],
|
||||
['-'],
|
||||
['n', 'u', 'p', 'r', 'f', 'c'],
|
||||
['-', '1', '2', '3'],
|
||||
['-', 's', 'p', 'd'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 'n', 'y']]
|
||||
]
|
||||
|
||||
|
||||
def complete_feature_dict():
|
||||
# old: http://nl.ijs.si/ME/Vault/V3/msd/html/
|
||||
# new: http://nl.ijs.si/ME/V4/msd/html/
|
||||
# changes: http://nl.ijs.si/jos/msd/html-en/msd.diffs.html
|
||||
return [[27,
|
||||
'A',
|
||||
['-', 'g', 's', 'p'],
|
||||
['-', 'p', 'c', 's'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[4, 'C', ['-', 'c', 's']],
|
||||
[1, 'I'],
|
||||
[28,
|
||||
'M',
|
||||
['-', 'd', 'r', 'l'],
|
||||
['-', 'c', 'o', 'p', 's'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[22,
|
||||
'N',
|
||||
['-', 'c', 'p'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 'n', 'y']],
|
||||
[41,
|
||||
'P',
|
||||
['-', 'p', 's', 'd', 'r', 'x', 'g', 'q', 'i', 'z'],
|
||||
['-', '1', '2', '3'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'n', 'g', 'd', 'a', 'l', 'i'],
|
||||
['-', 's', 'd', 'p'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 'y', 'b']],
|
||||
[1, 'Q'],
|
||||
[8, 'R', ['-', 'g', 'r'], ['-', 'p', 'c', 's']],
|
||||
[8, 'S', ['-', 'n', 'g', 'd', 'a', 'l', 'i']],
|
||||
[31,
|
||||
'V',
|
||||
['-', 'm', 'a'],
|
||||
['-', 'e', 'p', 'b'],
|
||||
['-', 'n', 'u', 'p', 'r', 'f', 'c', 'm'],
|
||||
['-', '1', '2', '3'],
|
||||
['-', 's', 'p', 'd'],
|
||||
['-', 'm', 'f', 'n'],
|
||||
['-', 'n', 'y']]
|
||||
]
|
||||
|
||||
|
||||
def check_feature_letter_usage(X_other_features, feature_dictionary):
|
||||
case_numbers = np.sum(X_other_features, axis=0)
|
||||
arrays = [1] * 164
|
||||
letters = list(decode_X_features(feature_dictionary, [arrays]))
|
||||
print(sum(case_numbers))
|
||||
for i in range(len(letters)):
|
||||
print(letters[i] + ': ' + str(case_numbers[i]))
|
||||
|
||||
|
||||
def dict_occurances_in_dataset_rate(content):
|
||||
feature_dictionary = complete_feature_dict()
|
||||
# case = 3107
|
||||
# print(content[case])
|
||||
# print(feature_dictionary)
|
||||
# X_other_features = create_X_features([content[case]], feature_dictionary)
|
||||
X_other_features = create_X_features(content, feature_dictionary)
|
||||
# print(X_other_features)
|
||||
# print(decode_X_features(feature_dictionary, X_other_features))
|
||||
X_other_features = np.array(X_other_features)
|
||||
|
||||
case_numbers = np.sum(X_other_features, axis=0)
|
||||
print(case_numbers)
|
||||
|
|
|
|||
101
tex_hyphenation.py
Normal file
101
tex_hyphenation.py
Normal file
|
|
@ -0,0 +1,101 @@
|
|||
import sys
|
||||
sys.path.insert(0, '../../../')
|
||||
from prepare_data import *
|
||||
dictionary, max_word, max_num_vowels, content, vowels, accetuated_vowels = create_dict()
|
||||
feature_dictionary = create_feature_dictionary(content)
|
||||
|
||||
|
||||
def read_hyphenation_pattern():
|
||||
with open('../../../hyphenation') as f:
|
||||
content = f.readlines()
|
||||
return [x[:-1] for x in content]
|
||||
|
||||
|
||||
def find_hyphenation_patterns_in_text(text, pattern):
|
||||
res = []
|
||||
index = 0
|
||||
while index < len(text):
|
||||
index = text.find(pattern, index)
|
||||
if index == -1:
|
||||
break
|
||||
res.append(index)
|
||||
index += 1 # +2 because len('ll') == 2
|
||||
|
||||
return res
|
||||
|
||||
|
||||
def create_hyphenation_dictionary(hyphenation_pattern):
|
||||
dictionary = []
|
||||
for el in hyphenation_pattern:
|
||||
substring = ''
|
||||
anomalies_indices = []
|
||||
digit_location = 0
|
||||
for let in list(el):
|
||||
if let.isdigit():
|
||||
anomalies_indices.append([digit_location, int(let)])
|
||||
else:
|
||||
substring += let
|
||||
digit_location += 1
|
||||
dictionary.append([substring, anomalies_indices])
|
||||
return dictionary
|
||||
|
||||
|
||||
def split_hyphenated_word(split, word):
|
||||
split = split[2:-2]
|
||||
print(split)
|
||||
word = list(word)[1:-1]
|
||||
res = []
|
||||
hyphenate = ''
|
||||
loc = 0
|
||||
for let in word:
|
||||
hyphenate += let
|
||||
if loc == len(split) or split[loc] % 2 == 1:
|
||||
res.append(hyphenate)
|
||||
hyphenate = ''
|
||||
loc += 1
|
||||
return res
|
||||
|
||||
|
||||
def hyphenate_word(word, hyphenation_dictionary):
|
||||
word = word.replace('è', 'č')
|
||||
word = '.' + word + '.'
|
||||
split = [0] * (len(word) + 1)
|
||||
for pattern in hyphenation_dictionary:
|
||||
pattern_locations = find_hyphenation_patterns_in_text(word, pattern[0])
|
||||
for pattern_location in pattern_locations:
|
||||
for el_hyphenation_dictionary in pattern[1]:
|
||||
if split[pattern_location + el_hyphenation_dictionary[0]] < el_hyphenation_dictionary[1]:
|
||||
split[pattern_location + el_hyphenation_dictionary[0]] = el_hyphenation_dictionary[1]
|
||||
return split_hyphenated_word(split, word)
|
||||
|
||||
|
||||
hyphenation_pattern = read_hyphenation_pattern()
|
||||
# ['zz', [{0:2},{1:1},{2:2}]]
|
||||
hyphenation_dictionary = create_hyphenation_dictionary(hyphenation_pattern)
|
||||
separated_word = hyphenate_word('izziv', hyphenation_dictionary)
|
||||
print(separated_word)
|
||||
|
||||
all_words = []
|
||||
i = 0
|
||||
for el in content:
|
||||
separated_word = hyphenate_word(el[0], hyphenation_dictionary)
|
||||
all_words.append([el[0], separated_word])
|
||||
if i % 10000 == 0:
|
||||
print(str(i)+'/'+str(len(content)))
|
||||
i += 1
|
||||
|
||||
errors = []
|
||||
errors2 = []
|
||||
for word in all_words:
|
||||
for hyphenated_part in word[1]:
|
||||
num_vowels = 0
|
||||
for let in list(hyphenated_part):
|
||||
if let in vowels:
|
||||
num_vowels += 1
|
||||
if num_vowels == 0:
|
||||
for let in list(hyphenated_part):
|
||||
if let == 'r':
|
||||
errors2.append(word[0])
|
||||
num_vowels += 1
|
||||
if num_vowels != 1:
|
||||
errors.append(word)
|
||||
Loading…
Reference in New Issue
Block a user