def __init__(self,
parent=None,
windowTitle="Color Palette List",
**kwargs):
super().__init__(parent, windowTitle=windowTitle, **kwargs)
self.setLayout(QVBoxLayout())
self.layout().setContentsMargins(0, 0, 0, 0)
sa = QScrollArea(horizontalScrollBarPolicy=Qt.ScrollBarAlwaysOff,
verticalScrollBarPolicy=Qt.ScrollBarAlwaysOn)
sa.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
sa.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
self.layout().addWidget(sa)
space = QWidget(self)
space.setLayout(QVBoxLayout())
sa.setWidget(space)
sa.setWidgetResizable(True)
self.buttons = []
self.setMinimumWidth(400)
box = gui.vBox(space, "Information", addSpace=True)
gui.widgetLabel(
box, '<p align="center">This dialog shows a list of predefined '
'color palettes <br>from colorbrewer.org that can be used '
'in Orange.<br/>You can select a palette by clicking on it.</p>')
box = gui.vBox(space, "Default Palette", addSpace=True)
butt = _ColorButton(DefaultRGBColors,
flat=True,
toolTip="Default color palette",
clicked=self._buttonClicked)
box.layout().addWidget(butt)
self.buttons.append(butt)
for type in [
"Qualitative", "Spectral", "Diverging", "Sequential", "Pastels"
]:
colorGroup = colorbrewer.colorSchemes.get(type.lower(), {})
if colorGroup:
box = gui.vBox(space, type + " Palettes", addSpace=True)
items = sorted(colorGroup.items())
for key, colors in items:
butt = _ColorButton(colors,
self,
toolTip=key,
flat=True,
clicked=self._buttonClicked)
box.layout().addWidget(butt)
self.buttons.append(butt)
buttons = QDialogButtonBox(QDialogButtonBox.Cancel,
rejected=self.reject)
self.layout().addWidget(buttons)
self.selectedColors = None
def __init__(self, parent=None, windowTitle="Color Palette List",
**kwargs):
super().__init__(parent, windowTitle=windowTitle, **kwargs)
self.setLayout(QVBoxLayout())
self.layout().setContentsMargins(0, 0, 0, 0)
sa = QScrollArea(
horizontalScrollBarPolicy=Qt.ScrollBarAlwaysOff,
verticalScrollBarPolicy=Qt.ScrollBarAlwaysOn
)
sa.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
sa.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
self.layout().addWidget(sa)
space = QWidget(self)
space.setLayout(QVBoxLayout())
sa.setWidget(space)
sa.setWidgetResizable(True)
self.buttons = []
self.setMinimumWidth(400)
box = gui.vBox(space, "Information", addSpace=True)
gui.widgetLabel(
box,
'<p align="center">This dialog shows a list of predefined '
'color palettes <br>from colorbrewer.org that can be used '
'in Orange.<br/>You can select a palette by clicking on it.</p>'
)
box = gui.vBox(space, "Default Palette", addSpace=True)
butt = _ColorButton(
DefaultRGBColors, flat=True, toolTip="Default color palette",
clicked=self._buttonClicked
)
box.layout().addWidget(butt)
self.buttons.append(butt)
for type in ["Qualitative", "Spectral", "Diverging", "Sequential", "Pastels"]:
colorGroup = colorbrewer.colorSchemes.get(type.lower(), {})
if colorGroup:
box = gui.vBox(space, type + " Palettes", addSpace=True)
items = sorted(colorGroup.items())
for key, colors in items:
butt = _ColorButton(colors, self, toolTip=key, flat=True,
clicked=self._buttonClicked)
box.layout().addWidget(butt)
self.buttons.append(butt)
buttons = QDialogButtonBox(
QDialogButtonBox.Cancel, rejected=self.reject
)
self.layout().addWidget(buttons)
self.selectedColors = None
class OWPreprocess(OWWidget):
name = 'Preprocess Text'
description = 'Construct a text pre-processing pipeline.'
icon = 'icons/TextPreprocess.svg'
priority = 30
class Inputs:
corpus = Input("Corpus", Corpus)
class Outputs:
corpus = Output("Corpus", Corpus)
autocommit = settings.Setting(True)
preprocessors = [
TransformationModule,
TokenizerModule,
NormalizationModule,
FilteringModule,
NgramsModule,
POSTaggingModule,
]
transformers = settings.SettingProvider(TransformationModule)
tokenizer = settings.SettingProvider(TokenizerModule)
normalizer = settings.SettingProvider(NormalizationModule)
filters = settings.SettingProvider(FilteringModule)
ngrams_range = settings.SettingProvider(NgramsModule)
pos_tagger = settings.SettingProvider(POSTaggingModule)
control_area_width = 250
buttons_area_orientation = Qt.Vertical
UserAdviceMessages = [
widget.Message(
"Some preprocessing methods require data (like word relationships, stop words, "
"punctuation rules etc.) from the NLTK package. This data was downloaded "
"to: {}".format(nltk_data_dir()), "nltk_data")
]
class Error(OWWidget.Error):
stanford_tagger = Msg("Problem while loading Stanford POS Tagger\n{}")
class Warning(OWWidget.Warning):
no_token_left = Msg(
'No tokens on output! Please, change configuration.')
def __init__(self, parent=None):
super().__init__(parent)
self.corpus = None
self.initial_ngram_range = None # initial range of input corpus — used for inplace
self.preprocessor = preprocess.Preprocessor()
# -- INFO --
info_box = gui.widgetBox(self.controlArea, 'Info')
info_box.setFixedWidth(self.control_area_width)
self.controlArea.layout().addStretch()
self.info_label = gui.label(info_box, self, '')
self.update_info()
# -- PIPELINE --
frame = QFrame()
frame.setContentsMargins(0, 0, 0, 0)
frame.setFrameStyle(QFrame.Box)
frame.setStyleSheet('.QFrame { border: 1px solid #B3B3B3; }')
frame_layout = QVBoxLayout()
frame_layout.setContentsMargins(0, 0, 0, 0)
frame_layout.setSpacing(0)
frame.setLayout(frame_layout)
self.stages = []
for stage in self.preprocessors:
widget = stage(self)
self.stages.append(widget)
setattr(self, stage.attribute, widget)
frame_layout.addWidget(widget)
widget.change_signal.connect(self.settings_invalidated)
frame_layout.addStretch()
self.scroll = QScrollArea()
self.scroll.setWidget(frame)
self.scroll.setWidgetResizable(True)
self.scroll.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
self.scroll.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
self.scroll.resize(frame_layout.sizeHint())
self.scroll.setMinimumHeight(500)
self.set_minimal_width()
self.mainArea.layout().addWidget(self.scroll)
# Buttons area
self.report_button.setFixedWidth(self.control_area_width)
commit_button = gui.auto_commit(self.buttonsArea,
self,
'autocommit',
'Commit',
box=False)
commit_button.setFixedWidth(self.control_area_width - 5)
self.buttonsArea.layout().addWidget(commit_button)
@Inputs.corpus
def set_data(self, data=None):
self.corpus = data.copy() if data is not None else None
self.initial_ngram_range = data.ngram_range if data is not None else None
self.commit()
def update_info(self, corpus=None):
if corpus is not None:
info = 'Document count: {}\n' \
'Total tokens: {}\n'\
'Total types: {}'\
.format(len(corpus), sum(map(len, corpus.tokens)), len(corpus.dictionary))
else:
info = 'No corpus.'
self.info_label.setText(info)
def commit(self):
self.Warning.no_token_left.clear()
if self.corpus is not None:
self.apply()
else:
self.update_info()
self.Outputs.corpus.send(None)
def apply(self):
self.preprocess()
@asynchronous
def preprocess(self):
for module in self.stages:
setattr(self.preprocessor, module.attribute, module.value)
self.corpus.pos_tags = None # reset pos_tags and ngrams_range
self.corpus.ngram_range = self.initial_ngram_range
return self.preprocessor(self.corpus,
inplace=True,
on_progress=self.on_progress)
@preprocess.on_start
def on_start(self):
self.progressBarInit(None)
@preprocess.callback
def on_progress(self, i):
self.progressBarSet(i, None)
@preprocess.on_result
def on_result(self, result):
self.update_info(result)
if result is not None and len(result.dictionary) == 0:
self.Warning.no_token_left()
result = None
self.Outputs.corpus.send(result)
self.progressBarFinished(None)
def set_minimal_width(self):
max_width = 250
for widget in self.stages:
if widget.enabled:
max_width = max(max_width, widget.sizeHint().width())
self.scroll.setMinimumWidth(max_width + 20)
@pyqtSlot()
def settings_invalidated(self):
self.set_minimal_width()
self.commit()
def send_report(self):
self.report_items('Preprocessor', self.preprocessor.report())
class OWPreprocess(OWWidget):
name = 'Preprocess Text'
description = 'Construct a text pre-processing pipeline.'
icon = 'icons/TextPreprocess.svg'
priority = 30
inputs = [(Input.CORPUS, Corpus, 'set_data')]
outputs = [(Output.PP_CORPUS, Corpus)]
autocommit = settings.Setting(True)
preprocessors = [
TransformationModule,
TokenizerModule,
NormalizationModule,
FilteringModule,
NgramsModule,
POSTaggingModule,
]
transformers = settings.SettingProvider(TransformationModule)
tokenizer = settings.SettingProvider(TokenizerModule)
normalizer = settings.SettingProvider(NormalizationModule)
filters = settings.SettingProvider(FilteringModule)
ngrams_range = settings.SettingProvider(NgramsModule)
pos_tagger = settings.SettingProvider(POSTaggingModule)
control_area_width = 250
buttons_area_orientation = Qt.Vertical
UserAdviceMessages = [
widget.Message(
"Some preprocessing methods require data (like word relationships, stop words, "
"punctuation rules etc.) from the NLTK package. This data, if you didn't have it "
"already, was downloaded to: {}".format(Downloader().default_download_dir()),
"nltk_data")]
class Error(OWWidget.Error):
stanford_tagger = Msg("Problem while loading Stanford POS Tagger\n{}")
class Warning(OWWidget.Warning):
no_token_left = Msg('No tokens on output! Please, change configuration.')
def __init__(self, parent=None):
super().__init__(parent)
self.corpus = None
self.initial_ngram_range = None # initial range of input corpus — used for inplace
self.preprocessor = preprocess.Preprocessor()
# -- INFO --
info_box = gui.widgetBox(self.controlArea, 'Info')
info_box.setFixedWidth(self.control_area_width)
self.controlArea.layout().addStretch()
self.info_label = gui.label(info_box, self, '')
self.update_info()
# -- PIPELINE --
frame = QFrame()
frame.setContentsMargins(0, 0, 0, 0)
frame.setFrameStyle(QFrame.Box)
frame.setStyleSheet('.QFrame { border: 1px solid #B3B3B3; }')
frame_layout = QVBoxLayout()
frame_layout.setContentsMargins(0, 0, 0, 0)
frame_layout.setSpacing(0)
frame.setLayout(frame_layout)
self.stages = []
for stage in self.preprocessors:
widget = stage(self)
self.stages.append(widget)
setattr(self, stage.attribute, widget)
frame_layout.addWidget(widget)
widget.change_signal.connect(self.settings_invalidated)
frame_layout.addStretch()
self.scroll = QScrollArea()
self.scroll.setWidget(frame)
self.scroll.setWidgetResizable(True)
self.scroll.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
self.scroll.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
self.scroll.resize(frame_layout.sizeHint())
self.scroll.setMinimumHeight(500)
self.set_minimal_width()
self.mainArea.layout().addWidget(self.scroll)
# Buttons area
self.report_button.setFixedWidth(self.control_area_width)
commit_button = gui.auto_commit(self.buttonsArea, self, 'autocommit',
'Commit', box=False)
commit_button.setFixedWidth(self.control_area_width - 5)
self.buttonsArea.layout().addWidget(commit_button)
def set_data(self, data=None):
self.corpus = data.copy() if data is not None else None
self.initial_ngram_range = data.ngram_range if data is not None else None
self.commit()
def update_info(self, corpus=None):
if corpus is not None:
info = 'Document count: {}\n' \
'Total tokens: {}\n'\
'Total types: {}'\
.format(len(corpus), sum(map(len, corpus.tokens)), len(corpus.dictionary))
else:
info = 'No corpus.'
self.info_label.setText(info)
def commit(self):
self.Warning.no_token_left.clear()
if self.corpus is not None:
self.apply()
else:
self.update_info()
self.send(Output.PP_CORPUS, None)
def apply(self):
self.preprocess()
@asynchronous
def preprocess(self):
for module in self.stages:
setattr(self.preprocessor, module.attribute, module.value)
self.corpus.pos_tags = None # reset pos_tags and ngrams_range
self.corpus.ngram_range = self.initial_ngram_range
return self.preprocessor(self.corpus, inplace=True, on_progress=self.on_progress)
@preprocess.on_start
def on_start(self):
self.progressBarInit(None)
@preprocess.callback
def on_progress(self, i):
self.progressBarSet(i, None)
@preprocess.on_result
def on_result(self, result):
self.update_info(result)
if result is not None and len(result.dictionary) == 0:
self.Warning.no_token_left()
result = None
self.send(Output.PP_CORPUS, result)
self.progressBarFinished(None)
def set_minimal_width(self):
max_width = 250
for widget in self.stages:
if widget.enabled:
max_width = max(max_width, widget.sizeHint().width())
self.scroll.setMinimumWidth(max_width + 20)
@pyqtSlot()
def settings_invalidated(self):
self.set_minimal_width()
self.commit()
def send_report(self):
self.report_items('Preprocessor', self.preprocessor.report())
class OWPreprocess(OWWidget):
name = '文本预处理'
description = '构建文本预处理的管道'
icon = 'icons/TextPreprocess.svg'
priority = 200
class Inputs:
corpus = Input("Corpus", Corpus)
class Outputs:
corpus = Output("Corpus", Corpus)
autocommit = settings.Setting(True)
preprocessors = [
TransformationModule,
TokenizerModule,
NormalizationModule,
FilteringModule,
NgramsModule,
POSTaggingModule,
]
transformers = settings.SettingProvider(TransformationModule)
tokenizer = settings.SettingProvider(TokenizerModule)
normalizer = settings.SettingProvider(NormalizationModule)
filters = settings.SettingProvider(FilteringModule)
ngrams_range = settings.SettingProvider(NgramsModule)
pos_tagger = settings.SettingProvider(POSTaggingModule)
control_area_width = 250
buttons_area_orientation = Qt.Vertical
UserAdviceMessages = [
widget.Message("部分预处理所需要的数据（例如词汇关系、停用词、标点符号规则等）是从NLTK包中获取的，",
"这些数据可以从{}下载。".format(nltk_data_dir()))
]
class Error(OWWidget.Error):
stanford_tagger = Msg("无法加载Stanford POS Tagger\n{}")
stopwords_encoding = Msg("停用词表编码不正确，请使用 UTF-8 再试一次。")
lexicon_encoding = Msg("词典编码不正确，请使用 UTF-8 再试一次。")
error_reading_stopwords = Msg("读取文件错误: {}")
error_reading_lexicon = Msg("读取文件错误: {}")
class Warning(OWWidget.Warning):
no_token_left = Msg('没有标记输出，请重新配置')
udpipe_offline = Msg('没有网络连接，UDPipe 只加载本地模型')
udpipe_offline_no_models = Msg('没有网络连接，UDPipe无本地模型')
def __init__(self, parent=None):
super().__init__(parent)
self.corpus = None
self.initial_ngram_range = None # initial range of input corpus — used for inplace
self.preprocessor = preprocess.Preprocessor()
# -- INFO --
info_box = gui.widgetBox(self.controlArea, '基本信息')
info_box.setFixedWidth(self.control_area_width)
self.controlArea.layout().addStretch()
self.info_label = gui.label(info_box, self, '')
self.update_info()
# -- PIPELINE --
frame = QFrame()
frame.setContentsMargins(0, 0, 0, 0)
frame.setFrameStyle(QFrame.Box)
frame.setStyleSheet('.QFrame { border: 1px solid #B3B3B3; }')
frame_layout = QVBoxLayout()
frame_layout.setContentsMargins(0, 0, 0, 0)
frame_layout.setSpacing(0)
frame.setLayout(frame_layout)
self.stages = []
for stage in self.preprocessors:
widget = stage(self)
self.stages.append(widget)
setattr(self, stage.attribute, widget)
frame_layout.addWidget(widget)
widget.change_signal.connect(self.settings_invalidated)
frame_layout.addStretch()
self.scroll = QScrollArea()
self.scroll.setWidget(frame)
self.scroll.setWidgetResizable(True)
self.scroll.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
self.scroll.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
self.scroll.resize(frame_layout.sizeHint())
self.scroll.setMinimumHeight(500)
self.set_minimal_width()
self.mainArea.layout().addWidget(self.scroll)
# Buttons area
self.report_button.setFixedWidth(self.control_area_width)
commit_button = gui.auto_commit(self.buttonsArea,
self,
'autocommit',
'提交',
'自动提交',
box=False)
commit_button.setFixedWidth(self.control_area_width - 5)
self.buttonsArea.layout().addWidget(commit_button)
@Inputs.corpus
def set_data(self, data=None):
self.corpus = data.copy() if data is not None else None
self.initial_ngram_range = data.ngram_range if data is not None else None
self.commit()
def update_info(self, corpus=None):
if corpus is not None:
info = '文档数量: {}\n' \
'标记数量: {}\n'\
'类型数量: {}'\
.format(len(corpus), sum(map(len, corpus.tokens)), len(corpus.dictionary))
else:
info = '没有数据集'
self.info_label.setText(info)
def commit(self):
self.Warning.no_token_left.clear()
if self.corpus is not None:
self.apply()
else:
self.update_info()
self.Outputs.corpus.send(None)
def apply(self):
self.preprocess()
@asynchronous
def preprocess(self):
for module in self.stages:
setattr(self.preprocessor, module.attribute, module.value)
self.corpus.pos_tags = None # reset pos_tags and ngrams_range
self.corpus.ngram_range = self.initial_ngram_range
return self.preprocessor(self.corpus,
inplace=True,
on_progress=self.on_progress)
@preprocess.on_start
def on_start(self):
self.progressBarInit(None)
@preprocess.callback
def on_progress(self, i):
self.progressBarSet(i, None)
@preprocess.on_result
def on_result(self, result):
self.update_info(result)
if result is not None and len(result.dictionary) == 0:
self.Warning.no_token_left()
result = None
self.Outputs.corpus.send(result)
self.progressBarFinished(None)
def set_minimal_width(self):
max_width = 250
for widget in self.stages:
if widget.enabled:
max_width = max(max_width, widget.sizeHint().width())
self.scroll.setMinimumWidth(max_width + 20)
@pyqtSlot()
def settings_invalidated(self):
self.set_minimal_width()
self.commit()
def send_report(self):
self.report_items('Preprocessor', self.preprocessor.report())
def __init__(self):
super().__init__()
self.query_edits = []
self.remove_buttons = []
# GUI
scrollArea = QScrollArea()
#### header
head_box = gui.hBox(self.controlArea)
head_box.setMaximumHeight(200)
info_box = gui.widgetBox(head_box, 'Info')
self.info = gui.widgetLabel(info_box,
("Import and/or create queries\n\n" +
"If a Table with a dictionary is connected as input,\n" +
"it can be imported, in which case multiple query\n" +
"terms with the same label will be combined.\n\n" +
"If the label is a number, or if negative weights are\n" +
"used (e.g. sentiment dictionaries), the query will be\n" +
"split into positive and negative parts.\n\n"
"If a row in the dictionary contains multiple terms\n" +
"separated with spaces, it will be seen as a phrase\n" +
"(by adding quotes). This can be turned of in case\n" +
"rows are already boolean queries."))
## from input
input_box = gui.widgetBox(head_box, "Create queries from dictionary file")
input_box.setMaximumWidth(350)
gui.button(input_box, self, 'multiple words are phrases', toggleButton=True, value='add_quotes',
buttonType=QCheckBox)
inputline_box = gui.hBox(input_box)
inputline_box.setMinimumHeight(70)
gui.listBox(inputline_box, self, 'query_in', labels='querytable_vars', box = 'Query column', callback=self.update_if_sync)
gui.listBox(inputline_box, self, 'label_in', labels='querytable_vars2', box = 'Label column', callback=self.update_if_sync)
self.weight_select = gui.listBox(inputline_box, self, 'weight_in', labels='querytable_vars2', box = 'Weight column', callback=self.update_if_sync)
input_button_box = gui.hBox(input_box)
gui.button(input_button_box, self, 'Keep synchronized', self.sync_on_off, toggleButton=True, value='sync', buttonType=QCheckBox)
gui.button(input_button_box, self, 'Import', self.import_queries)
gui.button(input_button_box, self, 'Append', self.append_queries)
## query field
query_box = gui.widgetBox(self.controlArea)
self.queries_box = QGridLayout()
query_box.layout().addLayout(self.queries_box)
scroll = QScrollArea()
scroll.setWidgetResizable(True)
scroll.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOn)
scroll.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
scroll.setWidget(query_box)
querygridbox = gui.widgetBox(self.controlArea, 'Query')
querygridbox.layout().addWidget(scroll)
querygridbox.setMinimumHeight(200)
querygridbox.setMinimumWidth(500)
gui.rubber(query_box)
self.queries_box.setColumnMinimumWidth(0, 5)
self.queries_box.setColumnMinimumWidth(1, 60)
self.queries_box.setColumnMinimumWidth(2, 350)
self.queries_box.setColumnStretch(1, 0)
self.queries_box.setColumnStretch(2, 100)
self.queries_box.addWidget(QLabel("Label"), 0, 1)
self.queries_box.addWidget(QLabel("Query"), 0, 2)
self.update_queries()
gui.button(query_box, self, "add query", callback=self.add_row, autoDefault=False)
## buttons
scarybuttonbox = gui.hBox(self.controlArea)
scarybuttonbox.layout().setAlignment(Qt.AlignRight)
gui.button(scarybuttonbox, self, "remove all queries", callback=self.remove_all, width=150)
QTimer.singleShot(0, self.send_queries) ## for send on startup
