def getFullURL(self):
from six import string_types
url = self.get_url()
domain = self.get_domain()
if self.useHttps:
protocol = b"https:/"
else:
protocol = b"http:/"
fullURL = b"/".join([protocol, domain, url])
if six.PY2:
queryString = b"?visualFeatures={2}&language={0}&details={1}".format(
self._language.decode("utf-8"),
self._detail,
self.getFeatures(),
)
else:
queryString = "?visualFeatures={2}&language={0}&details={1}".format(
self._language,
self._detail,
self.getFeatures(),
)
queryString = queryString.encode('utf-8')
fullURL = fullURL + queryString
# Unicode URL cause urllib3 to decode raw image data as if they were unicode.
if isinstance(fullURL, string_types):
if not isinstance(fullURL, str):
# URL is unicode in py2
log.io("Decode URL to str")
return fullURL.decode("utf-8")
else:
# is str in py2
return fullURL
else:
# is bytes in py3
return fullURL
def prepareImageContent(self, image):
imageContent = self.serializeImage(image)
imageSize = len(imageContent)
newImage = None
while imageSize >= self.maxSize and image.width >= self.minWidth and image.height >= self.minHeight:
self.imageInfo.resizeFactor = self.imageInfo.resizeFactor * self.asymptoticResizeFactor
newImage = self.getResizedImage()
imageContent = self.serializeImage(newImage)
imageSize = len(imageContent)
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
msg = "newWidth\n{0}\nnewHeight\n{1}\nsize\n{2}".format(
newImage.width, newImage.height, imageSize)
log.io(msg)
if imageSize > self.maxSize:
# Translators: Reported when error occurred during image serialization
ui.message(_(u"Image content is too big to upload."))
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
newImage = self.getResizedImage()
msg = "newWidth\n{0}\nnewHeight\n{1}\nsize\n{2}".format(
newImage.width, newImage.height, imageSize)
log.io(msg)
return False
else:
if self.uploadBase64EncodeImage:
return base64.standard_b64encode(imageContent)
else:
return imageContent
def extract_text(apiResult):
lineResult = []
resultSets = apiResult["ParsedResults"]
for result in resultSets:
lineResult.append(result["ParsedText"])
log.io(lineResult)
return u" ".join(lineResult)
def doHTTPRequest(callback, method, url, **kwargs):
"""
Call this method in a separate thread to avoid blocking.
@param callback:
@type callback:
@param method:
@type method:
@param url:
@type url: bytes
"""
refreshConnectionPool()
try:
if isinstance(url, bytes) and six.PY3:
url = url.decode('utf-8')
r = httpConnectionPool.request(method, url, **kwargs)
except urllib3.exceptions.TimeoutError as e:
# Translators: Message announced when network error occurred
showMessageInNetworkThread(_(u"Internet connection timeout.Recognition failed."))
callback(None)
return
except urllib3.exceptions.HTTPError as e:
log.error(e)
# Translators: Message announced when network error occurred
showMessageInNetworkThread(_(u"Recognition failed. Internet connection error."))
callback(None)
return
log.io(r.data)
callback(r.data)
def _speakSpellingGen(text,locale,useCharacterDescriptions):
textLength=len(text)
synth=getSynth()
synthConfig=config.conf["speech"][synth.name]
for count,char in enumerate(text):
uppercase=char.isupper()
charDesc=None
if useCharacterDescriptions:
from characterProcessing import getCharacterDescription
charDesc=getCharacterDescription(locale,char.lower())
if charDesc:
char=charDesc
else:
char=processSymbol(char)
if uppercase and synthConfig["sayCapForCapitals"]:
char=_("cap %s")%char
if uppercase and synth.isSupported("pitch") and synthConfig["raisePitchForCapitals"]:
oldPitch=synthConfig["pitch"]
synth.pitch=max(0,min(oldPitch+synthConfig["capPitchChange"],100))
index=count+1
log.io("Speaking character %r"%char)
if len(char) == 1 and synthConfig["useSpellingFunctionality"]:
synth.speakCharacter(char,index=index)
else:
synth.speakText(char,index=index)
if uppercase and synth.isSupported("pitch") and synthConfig["raisePitchForCapitals"]:
synth.pitch=oldPitch
while textLength>1 and (isPaused or getLastSpeechIndex()!=index):
yield
yield
if uppercase and synthConfig["beepForCapitals"]:
tones.beep(2000,50)
def recognizeHyperLink(self, link, onResult):
"""
Setup data for recognition then send request
@param link:
@param onResult: Result callback for result viewer
@return: None
"""
if self.networkThread:
# Translators: Error message
ui.message(
_(u"There is another recognition ongoing. Please wait."))
return
payloads = self.getPayloadForHyperLink(link)
fullURL = self.getFullURL()
headers = self.getHTTPHeaders("")
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
log.io(type(fullURL))
msg = u"{0}\n{1}\n{2}".format(
fullURL,
headers,
payloads,
)
log.io(msg)
self.sendRequest(self.callback, fullURL, payloads, headers)
def recognize(self, pixels, imageInfo, onResult):
"""
Setup data for recognition then send request
@param pixels:
@param imageInfo:
@param onResult: Result callback for result viewer
@return: None
"""
if self.networkThread:
# Translators: Error message
ui.message(
_(u"There is another recognition ongoing. Please wait."))
return
self.pixels = pixels
self.onResult = onResult
self.imageInfo = imageInfo
imageObject = self.prepareImageObject(pixels, imageInfo)
if not imageObject:
return
imageContent = self.prepareImageContent(imageObject)
if not imageContent:
return
payloads = self.getPayload(imageContent)
fullURL = self.getFullURL()
headers = self.getHTTPHeaders(imageContent)
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
log.io(type(fullURL))
msg = u"{0}\n{1}\n{2}".format(
fullURL,
headers,
payloads,
)
log.io(msg)
self.sendRequest(self.callback, fullURL, payloads, headers)
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (focus.sleepMode and not getattr(script, 'allowInSleepMode', False)):
raise NoInputGestureAction
wasInSayAll=False
if gesture.isModifier:
if not self.lastModifierWasInSayAll:
wasInSayAll=self.lastModifierWasInSayAll=sayAllHandler.isRunning()
elif self.lastModifierWasInSayAll:
wasInSayAll=True
self.lastModifierWasInSayAll=False
else:
wasInSayAll=sayAllHandler.isRunning()
if wasInSayAll:
gesture.wasInSayAll=True
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue, speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE, gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(queueHandler.eventQueue, speech.pauseSpeech, speechEffect == gesture.SPEECHEFFECT_PAUSE)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
log.io("Input: %s" % gesture.logIdentifier)
if self._captureFunc:
try:
if self._captureFunc(gesture) is False:
return
except:
log.error("Error in capture function, disabling", exc_info=True)
self._captureFunc = None
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"]["speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue, speech.speakMessage, gesture.displayName)
gesture.reportExtra()
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (
focus.sleepMode
and not getattr(script, 'allowInSleepMode', False)):
raise NoInputGestureAction
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE,
gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(
queueHandler.eventQueue, speech.pauseSpeech,
speechEffect == gesture.SPEECHEFFECT_PAUSE)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
log.io("Input: %s" % gesture.logIdentifier)
if self.isInputHelpActive:
bypass = getattr(script, "bypassInputHelp", False)
queueHandler.queueFunction(queueHandler.eventQueue,
self._handleInputHelp,
gesture,
onlyLog=bypass)
if not bypass:
return
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"][
"speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.speakMessage,
gesture.displayName)
gesture.reportExtra()
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def logTimeSinceInput():
"""Log the time since the last input was received.
This does nothing if time since input logging is disabled.
"""
if (not log.isEnabledFor(log.IO)
or not config.conf["debugLog"]["timeSinceInput"] or not manager
or not manager._lastInputTime):
return
log.io("%.3f sec since input" % (time.time() - manager._lastInputTime))
def logTimeSinceInput():
"""Log the time since the last input was received.
This does nothing if time since input logging is disabled.
"""
if (not log.isEnabledFor(log.IO)
or not config.conf["debugLog"]["timeSinceInput"]
or not manager or not manager._lastInputTime
):
return
log.io("%.3f sec since input" % (time.time() - manager._lastInputTime))
def _pushNextSpeech(self, doneSpeaking: bool):
queue = self._getNextPriority()
if not queue:
# No more speech.
log._speechManagerDebug("No more speech")
self._curPriQueue = None
return
if not self._curPriQueue:
# First utterance after no speech.
self._curPriQueue = queue
elif queue.priority > self._curPriQueue.priority:
# Preempted by higher priority speech.
if self._curPriQueue.enteredProfileTriggers:
if not doneSpeaking:
# Wait for the synth to finish speaking.
# _handleDoneSpeaking will call us again.
self._shouldPushWhenDoneSpeaking = True
return
self._exitProfileTriggers(
self._curPriQueue.enteredProfileTriggers)
self._curPriQueue = queue
elif queue.priority < self._curPriQueue.priority:
# Resuming a preempted, lower priority queue.
if queue.enteredProfileTriggers:
if not doneSpeaking:
# Wait for the synth to finish speaking.
# _handleDoneSpeaking will call us again.
self._shouldPushWhenDoneSpeaking = True
return
self._restoreProfileTriggers(queue.enteredProfileTriggers)
self._curPriQueue = queue
while queue.pendingSequences and isinstance(
queue.pendingSequences[0][0], ConfigProfileTriggerCommand):
if not doneSpeaking:
# Wait for the synth to finish speaking.
# _handleDoneSpeaking will call us again.
self._shouldPushWhenDoneSpeaking = True
return
self._switchProfile()
if not queue.pendingSequences:
# The last commands in this queue were profile switches.
# Call this method again in case other queues are waiting.
return self._pushNextSpeech(True)
seq = self._buildNextUtterance()
if seq:
log.io(f"Sending to synth: {seq}")
# So that we can handle any accidentally skipped indexes.
for item in seq:
if isinstance(item, IndexCommand):
self._indexesSpeaking.append(item.index)
getSynth().speak(seq)
def nvdaKgsHandleKeyInfoProc(lpKeys):
keys = (lpKeys[0], lpKeys[1], lpKeys[2])
log.io("keyInfo %d %d %d" % keys)
log.io("keyInfo hex %x %x %x" % keys)
names = []
routingIndex = None
if keys[0] == 0:
if keys[1] & 1: names.append('lf')
if keys[1] & 2: names.append('bk')
if keys[1] & 4: names.append('sr')
if keys[1] & 8: names.append('sl')
if keys[1] & 16: names.append('func1')
if keys[1] & 32: names.append('func2')
if keys[1] & 64: names.append('func3')
if keys[1] & 128: names.append('func4')
else:
tCode = 240
if keys[0] & 1 + tCode: names.append('func1')
if keys[0] & 2 + tCode: names.append('func2')
if keys[0] & 4 + tCode: names.append('func3')
if keys[0] & 8 + tCode: names.append('func4')
names.append('route')
routingIndex = keys[1] - 1
if routingIndex is not None:
log.io("names %s %d" % ('+'.join(names), routingIndex))
else:
log.io("names %s" % '+'.join(names))
if len(names):
inputCore.manager.executeGesture(InputGesture(names, routingIndex))
return True
return False
def getSignature(ak, sk, url, method='POST'):
import hashlib
import time
import hmac
import base64
pre = "sac-auth-v1/" + ak + '/' + str(int(time.time())) + "/3600"
part2 = "\nPOST\napi.ai.sogou.com\n/pub/ocr\n"
message = pre + part2
if not isinstance(message, bytes):
message = message.encode('utf-8')
pre = pre.encode('utf-8')
sk = sk.encode('utf-8')
signature = hmac.new(sk, message, digestmod=hashlib.sha256).digest()
headerSig = pre + b'/' + base64.b64encode(signature)
log.io(headerSig)
return str(headerSig)
def form_encode(options):
"""
generate www-form
@param options: request parameters
@type options: dict
@return:
"""
encoded_option = ""
key_sequence = sorted(iterkeys(options), reverse=False)
for k in key_sequence:
query = "&" + urlencode({k: options[k]})
encoded_option += query
payload = encoded_option[1:]
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
log.io(payload)
return payload
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (focus.sleepMode and not getattr(script, "allowInSleepMode", False)):
raise NoInputGestureAction
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue, speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE, gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(
queueHandler.eventQueue, speech.pauseSpeech, speechEffect == gesture.SPEECHEFFECT_PAUSE
)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
log.io("Input: %s" % gesture.logIdentifier)
if self.isInputHelpActive:
bypass = getattr(script, "bypassInputHelp", False)
queueHandler.queueFunction(queueHandler.eventQueue, self._handleInputHelp, gesture, onlyLog=bypass)
if not bypass:
return
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"]["speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue, speech.speakMessage, gesture.displayName)
gesture.reportExtra()
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def convert_to_line_result_format(self, apiResult):
lineResult = []
resultSets = apiResult["ParsedResults"]
for result in resultSets:
for line in result["TextOverlay"]["Lines"]:
currentLine = []
for word in line['Words']:
currentLine.append({
"x": word["Left"],
"y": word["Top"],
"width": word["Width"],
"height": word["Height"],
"text": word["WordText"],
})
lineResult.append(currentLine)
log.io(lineResult)
return lineResult
def convert_to_line_result_format(self, apiResult):
def extractCoordinate(coord):
groups = coord.split(',')
return int(groups[0]), int(groups[1])
lineResult = []
for items in apiResult[u"result"]:
rightXCoord, yCoord = extractCoordinate(items[u"frame"][1])
xCoord, downYCoord = extractCoordinate(items[u"frame"][2])
lineResult.append([{
"x": xCoord,
"y": yCoord,
"width": rightXCoord - xCoord,
"height": downYCoord - yCoord,
"text": items[u"content"],
}])
log.io(lineResult)
return lineResult
def nvdaKgsHandleKeyInfoProc(lpKeys):
keys = (lpKeys[0], lpKeys[1], lpKeys[2], lpKeys[3])
log.io("keyInfo %d %d %d %d" % keys)
log.io("keyInfo hex %x %x %x %x" % keys)
names = set()
routingIndex = None
if keys[2] & 1: names.add('func1')
if keys[2] & 2: names.add('func4')
if keys[2] & 4: names.add('ctrl')
if keys[2] & 8: names.add('alt')
if keys[2] & 16: names.add('select')
if keys[2] & 32: names.add('read')
if keys[2] & 64: names.add('func2')
if keys[2] & 128: names.add('func3')
if keys[0] == 1:
if keys[1] & 1: names.add('space')
if keys[1] & 2: names.add('dot6')
if keys[1] & 4: names.add('dot5')
if keys[1] & 8: names.add('dot4')
if keys[1] & 16: names.add('enter')
if keys[1] & 32: names.add('dot3')
if keys[1] & 64: names.add('dot2')
if keys[1] & 128: names.add('dot1')
elif keys[0] == 2:
if keys[1] & 1: names.add('esc')
if keys[1] & 2: names.add('inf')
if keys[1] & 4: names.add('bs')
if keys[1] & 8: names.add('del')
if keys[1] & 16: names.add('ins')
if keys[1] & 32: names.add('chng')
if keys[1] & 64: names.add('ok')
if keys[1] & 128: names.add('set')
elif keys[0] == 3:
if keys[1] & 1: names.add('upArrow')
if keys[1] & 2: names.add('downArrow')
if keys[1] & 4: names.add('leftArrow')
if keys[1] & 8: names.add('rightArrow')
elif keys[0] == 4:
names.add('route')
routingIndex = keys[1] - 1
elif keys[0] == 6:
if keys[1] & 1: names.add('bw')
if keys[1] & 2: names.add('fw')
if keys[1] & 4: names.add('ls')
if keys[1] & 8: names.add('rs')
if routingIndex is not None:
log.io("names %s %d" % ('+'.join(names), routingIndex))
else:
log.io("names %s" % '+'.join(names))
if len(names):
inputCore.manager.executeGesture(InputGesture(names, routingIndex))
return True
return False
def isValid(self):
oldProxy = config.conf["onlineOCRGeneral"]["proxyAddress"]
oldProxyType = config.conf["onlineOCRGeneral"]["proxyType"]
newProxyType = self.PROXY_TYPES[self.proxyTypeList.GetSelection()][0]
if newProxyType != u"noProxy":
# Translators: Reported when save proxy settings in online ocr panel
ui.message(_(u"Checking your proxy settings"))
config.conf["onlineOCRGeneral"]["proxyType"] = newProxyType
config.conf["onlineOCRGeneral"][
"proxyAddress"] = self.proxyAddressTextCtrl.GetValue()
from .winHttp import httpConnectionPool, refreshConnectionPool
try:
refreshConnectionPool()
log.io(httpConnectionPool)
r = httpConnectionPool.request(b'GET',
b"http://www.example.com")
msg = u"pool:\n{0}\nHeaders:\n{1}\nResponse:\n{2}".format(
httpConnectionPool,
r.headers,
r.data,
)
log.io(msg)
gui.messageBox(
# Translators: Reported when proxy verification fails in online ocr settings panel
caption=_(u"Proxy is valid, settings is saved."),
message=_(r.data),
)
return True
except Exception as e:
config.conf["onlineOCRGeneral"]["proxyType"] = oldProxyType
config.conf["onlineOCRGeneral"]["proxyAddress"] = oldProxy
refreshConnectionPool()
gui.messageBox(
# Translators: Reported when proxy verification fails in online ocr settings panel
caption=
_(u"Proxy is not valid, please check your proxy type and address."
),
message=_(e),
)
return False
else:
return True
def create_payload(self, png_string, app_id="", app_secret="", use_nvda_cn=True):
if use_nvda_cn:
options = {
b"image": png_string
}
else:
import time
import random
options = {
b"app_id": app_id,
b"time_stamp": int(time.time()),
b"nonce_str": str(random.randint(10000, 99999)),
b"image": png_string
}
sign = self.calculate_signature(options, app_secret)
sign = sign.upper()
options[b"sign"] = sign
log.io(options)
return options
def beep(hz,length,left=50,right=50):
"""Plays a tone at the given hz, length, and stereo balance.
@param hz: pitch in hz of the tone
@type hz: float
@param length: length of the tone in ms
@type length: integer
@param left: volume of the left channel (0 to 100)
@type left: integer
@param right: volume of the right channel (0 to 100)
@type right: integer
"""
log.io("Beep at pitch %s, for %s ms, left volume %s, right volume %s"%(hz,length,left,right))
if not player:
return
from NVDAHelper import generateBeep
bufSize=generateBeep(None,hz,length,left,right)
buf=create_string_buffer(bufSize)
generateBeep(buf,hz,length,left,right)
player.stop()
player.feed(buf.raw)
def speakTypedCharacters(ch):
global curWordChars;
if api.isTypingProtected():
realChar="*"
else:
realChar=ch
if ch.isalnum():
curWordChars.append(realChar)
elif ch=="\b":
# Backspace, so remove the last character from our buffer.
del curWordChars[-1:]
elif len(curWordChars)>0:
typedWord="".join(curWordChars)
curWordChars=[]
if log.isEnabledFor(log.IO):
log.io("typed word: %s"%typedWord)
if config.conf["keyboard"]["speakTypedWords"]:
speakText(typedWord)
if config.conf["keyboard"]["speakTypedCharacters"] and ord(ch)>=32:
speakSpelling(realChar)
def convert_to_line_result_format(self, apiResult):
lineResult = []
height = apiResult["Response"]["TextDetections"][0]["ItemPolygon"]["Y"]
wordResult = []
for words in apiResult["Response"]["TextDetections"]:
newHeight = words["ItemPolygon"]["Y"]
if newHeight > height:
height = newHeight
lineResult.append(wordResult)
wordResult = []
wordResult.append({
"x": words["ItemPolygon"]["X"],
"y": words["ItemPolygon"]["Y"],
"width": words["ItemPolygon"]["Width"],
"height": words["ItemPolygon"]["Height"],
"text": words["DetectedText"],
})
lineResult.append(wordResult)
log.io(lineResult)
return lineResult
def speakText(text,index=None,reason=REASON_MESSAGE,expandPunctuation=None):
"""Speaks some text.
@param text: The text to speak.
@type text: str
@param index: The index to mark this text with, which can be used later to determine whether this piece of text has been spoken.
@type index: int
@param reason: The reason for this speech; one of the REASON_* constants.
@param expandPunctuation: Whether to speak punctuation; C{None} (default) to use the user's configuration.
@param expandPunctuation: bool
"""
import speechViewer
if speechViewer.isActive:
speechViewer.appendText(text)
from brailleDisplayDrivers.DirectBM_drv import DirectBM_drv # Masataka.Shinke
DirectBM_drv.sp(text) # Masataka.Shinke
#if speechViewer.isActive: # Masataka.Shinke
# speechViewer.appendText(u'('+DirectBM_drv.wakach(text)+u')') # Masataka.Shinke
global beenCanceled, curWordChars
curWordChars=[]
if speechMode==speechMode_off:
return
elif speechMode==speechMode_beeps:
tones.beep(config.conf["speech"]["beepSpeechModePitch"],speechMode_beeps_ms)
return
if isPaused:
cancelSpeech()
beenCanceled=False
log.io("Speaking %r" % text)
if expandPunctuation is None:
expandPunctuation=config.conf["speech"]["speakPunctuation"]
if text is None:
text=""
else:
text=processText(text,expandPunctuation=expandPunctuation)
if not text or not text.isspace():
getSynth().speakText(text,index=index)
def convert_to_line_result_format(self, apiResult):
def extractCoordinate(coord):
groups = coord.split(',')
return int(groups[0]), int(groups[1]), int(groups[2]), int(
groups[3])
lineResult = []
resultSets = apiResult["regions"]
for result in resultSets:
for line in result["lines"]:
currentLine = []
for word in line['words']:
x, y, w, h = extractCoordinate(word["boundingBox"])
currentLine.append({
"x": x,
"y": y,
"width": w,
"height": h,
"text": word["text"]
})
lineResult.append(currentLine)
log.io(lineResult)
return lineResult
def getConnectionPool():
proxyType = config.conf["onlineOCRGeneral"]["proxyType"]
proxyAddress = config.conf["onlineOCRGeneral"]["proxyAddress"]
msg = u"type:\n{0}\naddress:\n{1}".format(
proxyType,
proxyAddress
)
if proxyType == u"http":
pool = urllib3.ProxyManager(
proxyAddress,
# cert_reqs='CERT_REQUIRED',
# ca_certs=certifi.where(),
timeout=urllib3.Timeout(connect=10, read=10)
)
msg += u"\nHTTP proxy\n{0}".format(
pool
)
log.io(msg)
return pool
elif proxyType == u"socks":
pool = SOCKSProxyManager(
proxyAddress,
# cert_reqs='CERT_REQUIRED',
# ca_certs=certifi.where(),
timeout=urllib3.Timeout(connect=10, read=10)
)
msg += u"\nSocks proxy\n{0}".format(
pool
)
log.io(msg)
return pool
else:
pool = urllib3.PoolManager(
# cert_reqs='CERT_REQUIRED',
# ca_certs=certifi.where(),
timeout=urllib3.Timeout(connect=10, read=10)
)
msg += u"\nNo proxy\n{0}".format(
pool
)
log.io(msg)
return pool
def extract_text(self, apiResult):
words = []
for items in apiResult["Response"]["TextDetections"]:
words.append(items["DetectedText"])
log.io(words)
return u" ".join(words)
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (
focus.sleepMode
and not getattr(script, 'allowInSleepMode', False)):
raise NoInputGestureAction
wasInSayAll = False
if gesture.isModifier:
if not self.lastModifierWasInSayAll:
wasInSayAll = self.lastModifierWasInSayAll = sayAllHandler.isRunning(
)
elif self.lastModifierWasInSayAll:
wasInSayAll = True
self.lastModifierWasInSayAll = False
else:
wasInSayAll = sayAllHandler.isRunning()
if wasInSayAll:
gesture.wasInSayAll = True
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE,
gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(
queueHandler.eventQueue, speech.pauseSpeech,
speechEffect == gesture.SPEECHEFFECT_PAUSE)
if gesture.shouldPreventSystemIdle:
winKernel.SetThreadExecutionState(winKernel.ES_SYSTEM_REQUIRED)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
self._lastInputTime = time.time()
log.io("Input: %s" % gesture.identifiers[0])
if self._captureFunc:
try:
if self._captureFunc(gesture) is False:
return
except:
log.error("Error in capture function, disabling",
exc_info=True)
self._captureFunc = None
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"][
"speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.speakMessage,
gesture.displayName)
gesture.reportExtra()
# #2953: if an intercepted command Script (script that sends a gesture) is queued
# then queue all following gestures (that don't have a script) with a fake script so that they remain in order.
if not script and scriptHandler._numIncompleteInterceptedCommandScripts:
script = lambda gesture: gesture.send()
if script:
scriptHandler.queueScript(script, gesture)
return
else:
# Clear memorized last script to avoid getLastScriptRepeatCount detect a repeat
# in case an unbound gesture is executed between two identical bound gestures.
queueHandler.queueFunction(queueHandler.eventQueue,
scriptHandler.clearLastScript)
raise NoInputGestureAction
def convert_to_line_result_format(self, apiResult):
lineResult = [[{
"x": 0,
"y": 0,
"width": 1,
"height": 1,
"text": self.extract_text(apiResult),
}]]
if "categories" in apiResult:
for category in apiResult["categories"]:
if "detail" in category:
entries = []
if "celebrities" in category["detail"]:
for result in category["detail"]["celebrities"]:
entries.append({
"x":
result["faceRectangle"]["x"],
"y":
result["faceRectangle"]["y"],
"width":
result["faceRectangle"]["w"],
"height":
result["faceRectangle"]["h"],
"text":
result["name"]
})
if len(entries) > 0:
lineResult.append(entries)
if "objects" in apiResult and len(apiResult["objects"]) > 0:
objectResult = []
resultSets = apiResult["objects"]
for result in resultSets:
objectResult.append({
"x": result["rectangle"]["x"],
"y": result["rectangle"]["y"],
"width": result["rectangle"]["w"],
"height": result["rectangle"]["h"],
"text": result["object"]
})
lineResult.append(objectResult)
if "brands" in apiResult and len(apiResult["brands"]) > 0:
brandResult = []
resultSets = apiResult["brands"]
for result in resultSets:
brandResult.append({
"x": result["rectangle"]["x"],
"y": result["rectangle"]["y"],
"width": result["rectangle"]["w"],
"height": result["rectangle"]["h"],
"text": result["name"]
})
lineResult.append(brandResult)
if "faces" in apiResult and len(apiResult["faces"]) > 0:
faceResult = []
resultSets = apiResult["faces"]
for result in resultSets:
faceResult.append({
"x": result["faceRectangle"]["x"],
"y": result["faceRectangle"]["y"],
"width": result["faceRectangle"]["w"],
"height": result["faceRectangle"]["h"],
"text": self.getFaceDescription(result)
})
lineResult.append(faceResult)
log.io(lineResult)
return lineResult
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (
focus.sleepMode
and not getattr(script, 'allowInSleepMode', False)):
raise NoInputGestureAction
wasInSayAll = False
if gesture.isModifier:
if not self.lastModifierWasInSayAll:
wasInSayAll = self.lastModifierWasInSayAll = sayAllHandler.isRunning(
)
elif self.lastModifierWasInSayAll:
wasInSayAll = True
self.lastModifierWasInSayAll = False
else:
wasInSayAll = sayAllHandler.isRunning()
if wasInSayAll:
gesture.wasInSayAll = True
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE,
gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(
queueHandler.eventQueue, speech.pauseSpeech,
speechEffect == gesture.SPEECHEFFECT_PAUSE)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
log.io("Input: %s" % gesture.logIdentifier)
if self._captureFunc:
try:
if self._captureFunc(gesture) is False:
return
except:
log.error("Error in capture function, disabling",
exc_info=True)
self._captureFunc = None
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"][
"speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.speakMessage,
gesture.displayName)
# nvdajp begin
if hasattr(gesture, "vkCode") and gesture.vkCode == winUser.VK_RETURN:
dummy = winUser.getAsyncKeyState(winUser.VK_BACK)
# nvdajp end
gesture.reportExtra()
# #2953: if an intercepted command Script (script that sends a gesture) is queued
# then queue all following gestures (that don't have a script) with a fake script so that they remain in order.
if not script and scriptHandler._numIncompleteInterceptedCommandScripts:
script = lambda gesture: gesture.send()
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def getImageFromClipboard(cls):
CF_DIB = 8
CF_HDROP = 15
CF_UNICODETEXT = 13
clipboardImage = None
formats = cls.enumerateClipboardFormat()
if CF_DIB in formats:
clipboardImage = ImageGrab.grabclipboard()
elif CF_HDROP in formats:
try:
filePathList = []
with winUser.openClipboard(gui.mainFrame.Handle):
rawData = windll.user32.GetClipboardData(CF_HDROP)
if not rawData:
ui.message(_("Error occurred while getting pasted file."))
rawData = winKernel.HGLOBAL(rawData, autoFree=False)
with rawData.lock() as addr:
fileCount = windll.shell32.DragQueryFileW(
c_uint32(addr),
c_uint32(0xFFFFFFFF),
c_uint32(0),
c_uint32(0)
)
for c in range(fileCount):
BUFFER_SIZE = 4096
filePath = create_unicode_buffer(BUFFER_SIZE)
windll.shell32.DragQueryFileW(
c_uint32(addr),
c_uint32(c),
c_uint32(filePath),
c_uint32(BUFFER_SIZE)
)
filePathList.append(wstring_at(filePath, size=BUFFER_SIZE).rstrip('\x00'))
log.debug("filePathList\n{0}".format(filePathList))
for fileName in filePathList:
# TODO Add a prompt for users to choose from
import os
if os.path.isfile(fileName):
clipboardImage = Image.open(rawData[0])
clipboardImage = clipboardImage.convert("RGB")
return clipboardImage
except TypeError as e:
log.io(e)
elif CF_UNICODETEXT in formats:
# TODO extract url or file path from text then grab an image from it.
try:
from api import getClipData
import os
text = getClipData()
if os.path.exists(text):
if os.path.isfile(text):
clipboardImage = Image.open(text)
else:
# Translators: Reported when text in clipboard is not a valid path
ui.message(_(u"Text in clipboard is the name of a directory."))
else:
# Translators: Reported when text in clipboard is not a valid path
ui.message(_(u"Text in clipboard is not a valid path."))
except IOError:
# Translators: Reported when cannot get content of the path specified
errMsg = _("The file specified in clipboard is not an image")
ui.message(errMsg)
return clipboardImage
def executeKeyboardGesture(self, gesture, bypassRemanence=False):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if not hasattr(gesture, "noAction"):
gesture.noAction = False
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
newGesture = self.manageRemanence(
gesture) if not bypassRemanence else None
if newGesture is not None:
queueHandler.queueFunction(queueHandler.eventQueue,
self.executeNewGesture, newGesture)
return
newGesture = self.getNumpadKeyReplacement(gesture)
if newGesture is not None:
queueHandler.queueFunction(queueHandler.eventQueue,
self.executeNewGesture, newGesture)
return
script = gesture.script
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL\
or (focus.sleepMode and not getattr(script, 'allowInSleepMode', False)):
raise NoInputGestureAction
wasInSayAll = False
if gesture.isModifier:
if not _NVDA_InputManager.lastModifierWasInSayAll:
wasInSayAll = _NVDA_InputManager.lastModifierWasInSayAll = sayAllHandler.isRunning(
)
elif _NVDA_InputManager.lastModifierWasInSayAll:
wasInSayAll = True
_NVDA_InputManager.lastModifierWasInSayAll = False
else:
wasInSayAll = sayAllHandler.isRunning()
if wasInSayAll:
gesture.wasInSayAll = True
speechEffect = gesture.speechEffectWhenExecuted
if speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE,
gesture.SPEECHEFFECT_RESUME): # noqa:E501
queueHandler.queueFunction(
queueHandler.eventQueue, speech.pauseSpeech,
speechEffect == gesture.SPEECHEFFECT_PAUSE)
if py3 and gesture.shouldPreventSystemIdle:
winKernel.SetThreadExecutionState(winKernel.ES_SYSTEM_REQUIRED
| winKernel.ES_DISPLAY_REQUIRED)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
_NVDA_InputManager._lastInputTime = time.time()
log.io("Input: %s" % gesture.identifiers[0])
if _NVDA_InputManager._captureFunc:
try:
if _NVDA_InputManager._captureFunc(gesture) is False:
return
except: # noqa:E722
log.error("Error in capture function, disabling",
exc_info=True)
_NVDA_InputManager._captureFunc = None
if gesture.isModifier:
if gesture.noAction:
gesture.normalizedModifiers = []
return
raise NoInputGestureAction
self.speakGesture(gesture)
if not script:
gesture.reportExtra()
# then queue all following gestures
# (that don't have a script
# ) with a fake script so that they remain in order.
if not script and (
bypassRemanence
or scriptHandler._numIncompleteInterceptedCommandScripts):
script = lambda gesture: gesture.send() # noqa:E731
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def checkAndResizeImage(self, image):
"""
Check Image Size to meet requirement of API
@param image:
@type image: PIL.Image.Image
@return: Resized image
@rtype: PIL.Image.Image or bool
"""
isImageValid = True
width = image.width
height = image.height
widthResizeFactor = 1
heightResizeFactor = 1
msg = u"Original size\nwidth:\n{w}\nheight:\n{h}".format(w=width,
h=height)
if width >= height:
aspectRatio = width / height
else:
aspectRatio = height / width
widthStatus = 0
heightStatus = 0
if aspectRatio > (self.maxHeight / self.minWidth):
isImageValid = False
# Translators: Reported when error occurred during image resizing
errorMsg = _(
u"Image aspect ratio is too big. Cannot resize properly for this engine."
)
else:
if self.minWidth <= width <= self.maxWidth:
widthResizeFactor = 1
elif width < self.minWidth:
widthResizeFactor = (float(self.minWidth) / width) + 1
else:
widthResizeFactor = (float(self.maxWidth) / width) + 1
if self.minHeight <= height <= self.maxHeight:
heightResizeFactor = 1
elif height < self.minHeight:
heightResizeFactor = (float(self.minHeight) / height) + 1
else:
heightResizeFactor = (float(self.maxHeight) / height) + 1
msg += u"\nwidthResizeFactor:\n{0}\nheightResizeFactor:\n{1}".format(
widthResizeFactor, heightResizeFactor)
# Translators: Reported when error occurred during image conversion
errorMsg = _(u"Error occurred when converting images")
if widthResizeFactor >= self.resizeUpperLimit:
# Translators: Reported when image size is not valid
errorMsg = _(u"Image width is too big for this engine")
isImageValid = False
if heightResizeFactor <= self.resizeLowerLimit:
isImageValid = False
# Translators: Reported when error occurred during image conversion
errorMsg = _(u"Image height is too small for this engine")
if widthResizeFactor >= self.resizeUpperLimit:
# Translators: Reported when image size is not valid
errorMsg = _(u"Image width is too big for this engine")
isImageValid = False
if heightResizeFactor <= self.resizeLowerLimit:
isImageValid = False
# Translators: Reported when error occurred during image conversion
errorMsg = _(u"Image height is too small for this engine")
self.imageInfo.resizeFactor = int(
max(widthResizeFactor, heightResizeFactor))
if heightResizeFactor != 1 and widthResizeFactor != 1 and config.conf[
"onlineOCRGeneral"]["notifyIfResizeRequired"]:
if gui.messageBox(
# Translators: The confirmation prompt displayed when the image need to be resized.
_("Image size is not proper for recognition. Do you want to resize? Press OK to continue. Press Cancel to cancel recognition"
),
# Translators: The title of the confirmation dialog for the image need to be resized.
_("Confirm Resize"),
wx.OK | wx.CANCEL | wx.ICON_QUESTION,
gui.mainFrame) != wx.OK:
return
if isImageValid:
image = self.getResizedImage()
width = image.width
height = image.height
msg += u"\nSize after resizing\nwidth:\n{w}\nheight:\n{h}".format(
w=width, h=height)
log.io(msg)
if width * height > self.maxPixels:
pixelCount = image.width * image.height
while pixelCount >= self.maxPixels and image.width >= self.minWidth and image.height >= self.minHeight:
self.imageInfo.resizeFactor = self.imageInfo.resizeFactor * self.asymptoticResizeFactor
image = self.getResizedImage()
if config.conf["onlineOCRGeneral"]["verboseDebugLogging"]:
msg = "newWidth\n{0}\nnewHeight\n{1}\npixelCount\n{2}".format(
image.width, image.height, pixelCount)
log.io(msg)
if image.width * image.height > self.maxPixels:
isImageValid = False
# Translators: Reported when error occurred during image resizing
ui.message(_(u"Image has too many pixels."))
return False
else:
return self.getResizedImage()
else:
return self.getResizedImage()
else:
log.io(msg)
ui.message(errorMsg)
return False
def executeGesture(self, gesture):
"""Perform the action associated with a gesture.
@param gesture: The gesture to execute.
@type gesture: L{InputGesture}
@raise NoInputGestureAction: If there is no action to perform.
"""
if watchdog.isAttemptingRecovery:
# The core is dead, so don't try to perform an action.
# This lets gestures pass through unhindered where possible,
# as well as stopping a flood of actions when the core revives.
raise NoInputGestureAction
script = gesture.script
if "brailleDisplayDrivers" in str(type(gesture)):
if instanceGP.brailleKeyboardLocked and (
(hasattr(script, "__func__") and
script.__func__.__name__ != "script_toggleLockBrailleKeyboard")
or not hasattr(script, "__func__")):
return
if not config.conf["brailleExtender"][
'stopSpeechUnknown'] and gesture.script == None:
stopSpeech = False
elif hasattr(script, "__func__") and (script.__func__.__name__ in [
"script_braille_dots", "script_braille_enter",
"script_volumePlus", "script_volumeMinus",
"script_toggleVolume", "script_hourDate", "script_ctrl",
"script_alt", "script_nvda", "script_win", "script_ctrlAlt",
"script_ctrlAltWin", "script_ctrlAltWinShift",
"script_ctrlAltShift", "script_ctrlWin", "script_ctrlWinShift",
"script_ctrlShift", "script_altWin", "script_altWinShift",
"script_altShift", "script_winShift"
] or (not config.conf["brailleExtender"]['stopSpeechScroll']
and script.__func__.__name__ in
["script_braille_scrollBack", "script_braille_scrollForward"])):
stopSpeech = False
else:
stopSpeech = True
else:
stopSpeech = True
focus = api.getFocusObject()
if focus.sleepMode is focus.SLEEP_FULL or (focus.sleepMode and not getattr(
script, 'allowInSleepMode', False)):
raise NoInputGestureAction
wasInSayAll = False
if gesture.isModifier:
if not self.lastModifierWasInSayAll:
wasInSayAll = self.lastModifierWasInSayAll = sayAllHandler.isRunning(
)
elif self.lastModifierWasInSayAll:
wasInSayAll = True
self.lastModifierWasInSayAll = False
else:
wasInSayAll = sayAllHandler.isRunning()
if wasInSayAll:
gesture.wasInSayAll = True
speechEffect = gesture.speechEffectWhenExecuted
if not stopSpeech: pass
elif speechEffect == gesture.SPEECHEFFECT_CANCEL:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.cancelSpeech)
elif speechEffect in (gesture.SPEECHEFFECT_PAUSE,
gesture.SPEECHEFFECT_RESUME):
queueHandler.queueFunction(queueHandler.eventQueue, speech.pauseSpeech,
speechEffect == gesture.SPEECHEFFECT_PAUSE)
if log.isEnabledFor(log.IO) and not gesture.isModifier:
self._lastInputTime = time.time()
log.io("Input: %s" % gesture.identifiers[0])
if self._captureFunc:
try:
if self._captureFunc(gesture) is False:
return
except BaseException:
log.error("Error in capture function, disabling", exc_info=True)
self._captureFunc = None
if gesture.isModifier:
raise NoInputGestureAction
if config.conf["keyboard"][
"speakCommandKeys"] and gesture.shouldReportAsCommand:
queueHandler.queueFunction(queueHandler.eventQueue,
speech.speakMessage, gesture.displayName)
gesture.reportExtra()
# #2953: if an intercepted command Script (script that sends a gesture) is queued
# then queue all following gestures (that don't have a script) with a fake script so that they remain in order.
if not script and scriptHandler._numIncompleteInterceptedCommandScripts:
script = lambda gesture: gesture.send()
if script:
scriptHandler.queueScript(script, gesture)
return
raise NoInputGestureAction
def extract_text(apiResult): words = [] for items in apiResult[u"result"]: words.append(items[u"content"]) log.io(items[u"content"]) return u" ".join(words)
