import json
from os.path import join, dirname
from os import environ
from watson_developer_cloud import VisualRecognitionV3 as VisualRecognition
import picamera
import subprocess
from time import sleep
import os
import zipfile
visual_recognition = VisualRecognition(
'2016-05-20', api_key='dec041de3393d4b66120d22e628b81f2f3c1e9c0')
with open(join(dirname(__file__), 'surya.zip'), 'rb') as surya, \
open(join(dirname(__file__), 'vijay.zip'), 'rb') as vijay:
print(
json.dumps(visual_recognition.create_classifier(
'face',
surya_positive_examples=surya,
vijay_positive_examples=vijay),
indent=2))
import json
from os.path import join, dirname
from os import environ
from watson_developer_cloud import VisualRecognitionV3 as VisualRecognition
from watson_developer_cloud import VisualRecognitionV3 as VisualRecognition
#31e589a7ae4990a6261e68d8b2931acf5145b973 - alvaro - junto
#37ff1e95c9da3f5e5161d6f49b0139469c087f8d . watsonR - separados
visual_recognition = VisualRecognition('2016-05-20', api_key='31e589a7ae4990a6261e68d8b2931acf5145b973')
with open(join(dirname(__file__), 'positive_todas.zip'), 'rb') as carne_uptc, \
open(join(dirname(__file__), 'negative_arregladas.zip'), 'rb') as carne_otro:
print(json.dumps(visual_recognition.create_classifier('carne_uptc', completo_positive_examples=carne_uptc, negative_examples=carne_otro), indent=2))
#with open(join(dirname(__file__), 'positive_arregladas.zip'), 'rb') as carne_completo, \
# open(join(dirname(__file__), 'positive_arregladas.zip'), 'rb') as carne_cortado, \
# open(join(dirname(__file__), 'negative_arregladas.zip'), 'rb') as carne_otro:
# print(json.dumps(visual_recognition.create_classifier('CarnesUPTCvsCarnesOtros', completo_positive_examples=carne_completo, cortado_positive_examples=carne_cortado, negative_examples=carne_otro), indent=2))
class CustomClassifier():
def __init__(self):
self.visual_recognition = VisualRecognition('2016-05-20', api_key=config.api_key)
self.downloader = ImageDownloader(config.yoox_rest_endpoint)
"""
This method creates a new visual custom classifier starting from a file path.
The path will be a directory containing a group of zip files filled with jpg images, each file represents a class sample data
@type classifierName: string
@param classifierName: new custom classifier name
@type inputDataPath: string
@param inputDataPath: zip files directory path
"""
def create_custom_classifier(self, classifierName, inputDataPath): # it could take a negative examples data path as input
params = {}
for dirname, subdirs, files in os.walk(inputDataPath):
for file in files:
absname = os.path.abspath(os.path.join(dirname, file))
name = os.path.splitext(os.path.basename(file))[0]
params[name.replace("-", "_") + "_positive_examples"] = open(absname)
if config.print_file_params_long_object:
print params
# for dirname, subdirs, files in os.walk(negDataPath):
# for file in files:
# absname = os.path.abspath(os.path.join(dirname, file))
# name = os.path.splitext(os.path.basename(file))[0]
# params["negative_examples"] = open(absname)
print "Uploading files to Visual Recognition servers... It will take several time based on the number of images"
response = self.visual_recognition.create_classifier(classifierName, **params)
print "Watson classifier created: ", json.dumps(response)
self.classifier_id = response['classifier_id']
def sameIdCondition(self, id, inputID):
return id == inputID
"""
This method retrieves a classifier by id
@type classifier_id: string
@param classifier_id: classifier id
@return response: Watson's representation of a classifier
"""
def getClassifiers(self, classifier_id):
response = self.visual_recognition.get_classifier(classifier_id=classifier_id)
return response
"""
This method retrieves this classifier representation
@return response: Watson's representation of a classifier
"""
def getClassifierStatus(self):
return self.getClassifiers(self.classifier_id)
"""
This method recharge a Watson visual recognition classifier by id
@type classifier_id: string
@param classifier_id: classifier id
"""
def recharge_classifier(self, classifier_id):
# if the classifier does not exist it will fail
self.classifier_id = self.getClassifiers(classifier_id)['classifier_id']
"""
This method requests to the classifier for an image classification starting from an image ID of the csv file
@type imageId: string
@param imageId: an image ID contained in the csv training set file
@return: the visual recognition representation of the requested image classification
"""
def classify_image_fromTrainingSet(self, imageId, threshold):
# Read the CSV File
allImages = pre_processing.readingFromInputFile(config.csv_path)
# in a production env it should be necessary to handle the not found exception
imgEl = [el for el in allImages if imageId == el.id][0]
print config.temp_data_path
if not os.path.isdir(config.temp_data_path):
os.makedirs(config.temp_data_path)
# Download the image
imgPath = self.downloader.downloadAndSaveImage(config.temp_data_path, imgEl)
print "Sending classification request to classifier"
return self.visual_recognition.classify(open(imgPath), classifier_ids=[self.classifier_id], threshold=threshold)
"""
This method requests to the classifier for an image classification starting from its ID
@type imageId: string
@param imageId: an image ID contained in the csv training set file
@return: the visual recognition representation of the requested image classification
"""
def classify_image(self, imageId, threshold):
imgPath = self.downloader.downloadAndSaveImage(config.temp_data_path, InputImgElement(imageId))
print "Sending classification request to classifier"
return self.visual_recognition.classify(open(imgPath), classifier_ids=[self.classifier_id], threshold=threshold)
