def migrate_images(src_trainer, dest_trainer, project_id, dest_project_id,
created_tags):
# Migrate any tagged images that may exist and preserve their tags and regions.
count = src_trainer.get_tagged_image_count(project_id)
print("Found:", count, "tagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_tagged_images(project_id,
take=count_to_migrate,
skip=migrated)
for i in images:
print("Migrating", i.id, i.image_uri)
regions = []
for r in i.regions:
print("Found region:", r.region_id, r.tag_id, r.left, r.top,
r.width, r.height)
regions.append(
Region(tag_id=created_tags[r.tag_id],
left=r.left,
top=r.top,
width=r.width,
height=r.height))
entry = ImageUrlCreateEntry(url=i.image_uri, regions=regions)
dest_trainer.create_images_from_urls(dest_project_id,
images=[entry])
migrated += count_to_migrate
count -= count_to_migrate
# Migrate any untagged images that may exist.
count = src_trainer.get_untagged_image_count(project_id)
print("Found:", count, "untagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_untagged_images(project_id,
take=count_to_migrate,
skip=migrated)
for i in images:
print("Migrating", i.id, i.image_uri)
regions = []
for r in i.regions:
print("Found region:", r.region_id, r.tag_id, r.left, r.top,
r.width, r.height)
regions.append(
Region(tag_id=created_tags[r.tag_id],
left=r.left,
top=r.top,
width=r.width,
height=r.height))
entry = ImageUrlCreateEntry(url=i.image_uri, regions=regions)
dest_trainer.create_images_from_urls(dest_project_id,
images=[entry])
migrated += count_to_migrate
count -= count_to_migrate
return images
def coco_to_custom_vision(stream, project_id, trainer, data_dir):
stream = stream | mp.as_list
tags = stream | mp.select_field('class_id') | mp.dedup() | mp.as_list
cv_tags = {tag: trainer.create_tag(project_id, tag) for tag in tags}
stream = (
stream
| mp.apply(['width', 'height', 'ground_truth'], 'ground_truth',
lambda x: x[2]
| mp.where(lambda box: (box['width'] >= x[0] * 0.1) and
(box['height'] >= x[1] * 0.1))
| mp.as_list)
| mp.filter('ground_truth', lambda x: len(x) > 0)
| mp.apply(
['width', 'height', 'ground_truth'], 'regions', lambda x: x[2]
| mp.select(lambda box: Region(tag_id=cv_tags[box['tag']].id,
left=box['x1'] / x[0],
top=box['y1'] / x[1],
width=box['width'] / x[0],
height=box['height'] / x[1]))
| mp.as_list)
| mp.apply(['filename', 'regions'], 'tagged_img',
lambda x: ImageFileCreateEntry(
name=x[0],
contents=open(join(data_dir, x[0]), mode="rb").read(),
regions=x[1]))
| mp.as_list)
tagged_images_with_regions = stream | mp.select_field(
'tagged_img') | mp.as_list
for i in range(0, len(tagged_images_with_regions), 50):
trainer.create_images_from_files(
project_id, images=tagged_images_with_regions[i:i + 50])
def convert_row_to_region(tag_map, row):
tag = tag_map[row[TAG_NAME_LOCATION]]
x = float(row[X_MIN_LOCATION])
y = float(row[Y_MIN_LOCATION])
width = float(row[X_MAX_LOCATION]) - x
height = float(row[Y_MAX_LOCATION]) - y
return Region(tag_id=tag.id, left=x, top=y, width=width, height=height)
def add_superfluous_region_to_regions(regions, map_tag_name_to_id):
tag_id = map_tag_name_to_id.get(superfluous_tag_name)
if tag_id == None:
raise Exception(
f"No Custom Vision tag ID found for superfluous region with tag name {superfluous_tag_name}."
)
regions.append(Region(tag_id=tag_id, left=0, top=0, width=1, height=1))
def image_data_region_to_custom_vision_region(region, map_tag_name_to_id):
tag_id = map_tag_name_to_id.get(region.tag_name)
if tag_id == None:
raise Exception(
f"No Custom Vision tag ID found for locally-defined region with tag name \"{region.tag_name}\"."
)
return Region(tag_id=tag_id,
left=region.left,
top=region.top,
width=region.width,
height=region.height)
def add_detection_images(image, coordinates, project_id):
ret, buf = cv2.imencode('.jpg', image)
tag_id = trainer.get_tags(project_id)[0].id
region = Region(tag_id=tag_id,
left=coordinates[0],
top=coordinates[1],
width=coordinates[2],
height=coordinates[3])
img = ImageFileCreateEntry(name="test",
contents=buf.tobytes(),
regions=[region],
tag_ids=[tag_id])
return trainer.create_images_from_files(project_id, images=[img])
def upload_training_images(self, training_labeled_images):
"""
Upload training images to Azure Custom Vision Object Detection.
Args:
----
training_lableded_images: list of labeledImage
"""
assert (self.project_id is not None)
print("Adding images...")
tagged_images_with_regions = []
batch = 0
for i in range(len(training_labeled_images)):
if i > 0 and (i % 64) == 0:
batch += 1
print("Adding images: batch ", batch)
self._upload_one_batch_training_images(
tagged_images_with_regions)
tagged_images_with_regions = []
# accumulating labels within one batch
labeled_img = training_labeled_images[i]
for t, labels in labeled_img.labels.items():
if t not in self.tags.keys(): self.create_tag(t)
tag_id = self.tags[t]
regions = []
for m in labels:
x, y, w, h = normalize_coordinates(m, labeled_img.shape)
regions.append(
Region(tag_id=tag_id, left=x, top=y, width=w,
height=h))
with open(labeled_img.path, mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=labeled_img.name,
contents=image_contents.read(),
regions=regions))
batch += 1
if len(tagged_images_with_regions) > 0:
print("Adding images: batch ", batch)
self._upload_one_batch_training_images(tagged_images_with_regions)
return
def Upload_Images(folder):
print("Uploading images...")
# Get the tags defined in the project
tags = training_client.get_tags(custom_vision_project.id)
# Create a list of images with tagged regions
tagged_images_with_regions = []
# Get the images and tagged regions from the JSON file
with open('tagged-images.json', 'r') as json_file:
tagged_images = json.load(json_file)
for image in tagged_images['files']:
# Get the filename
file = image['filename']
# Get the tagged regions
regions = []
for tag in image['tags']:
tag_name = tag['tag']
# Look up the tag ID for this tag name
tag_id = next(t for t in tags if t.name == tag_name).id
# Add a region for this tag using the coordinates and dimensions in the JSON
regions.append(
Region(tag_id=tag_id,
left=tag['left'],
top=tag['top'],
width=tag['width'],
height=tag['height']))
# Add the image and its regions to the list
with open(os.path.join(folder, file), mode="rb") as image_data:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file,
contents=image_data.read(),
regions=regions))
# Upload the list of images as a batch
upload_result = training_client.create_images_from_files(
custom_vision_project.id,
ImageFileCreateBatch(images=tagged_images_with_regions))
# Check for failure
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
print("Image status: ", image.status)
else:
print("Images uploaded.")
def upload_images(training_key):
trainer = CustomVisionTrainingClient(training_key, endpoint=ENDPOINT)
# Find the object detection domain
obj_detection_domain = next(domain for domain in trainer.get_domains() if domain.type == "ObjectDetection")
print("Creating project...")
try:
project = trainer.create_project("LEGO Vision", domain_id=obj_detection_domain.id)
except HttpOperationError:
print("Project already exists. Using this one.")
project = trainer.get_project(project_id="71548120-925d-4e59-ba7e-32f99de50240")
classes = os.path.join(BASE_DIRECTORY, "class_map.txt")
tags = dict()
# Make two tags in the new project
for _class in list(map(lambda line: line.split('\t')[0], open(classes).readlines())):
try:
tags[_class] = trainer.create_tag(project.id, _class)
except HttpOperationError:
print("Tag already created, continuing...")
for tag in trainer.get_tags(project_id="71548120-925d-4e59-ba7e-32f99de50240"):
tags[tag.name] = tag
# Go through the data table above and create the images
print("Adding images...")
tagged_images_with_regions = []
for image_path in glob.glob(os.path.join(IMAGES_FOLDER, "*.jpg")):
file_id, extension = image_path.split(".", 1)
image = cv2.imread(image_path)
bboxes = read_bboxes(os.path.join(IMAGES_FOLDER, file_id + ".bboxes.tsv"), scale=1, padding=(0, 0, 0, 0))
labels = read_labels(os.path.join(IMAGES_FOLDER, file_id + ".bboxes.labels.tsv"))
regions = [Region(tag_id=tags[_class].id, left=bbox[0] / image.shape[1], top=bbox[1] / image.shape[0],
width=abs(bbox[0] - bbox[2]) / image.shape[1], height=abs(bbox[1] - bbox[3]) / image.shape[0])
for _class, bbox in zip(labels,
bboxes)]
with open(image_path, mode="rb") as image_contents:
tagged_images_with_regions.append(ImageFileCreateEntry(name=file_id, contents=image_contents.read(),
regions=regions))
print("Azure Custom Vision can only accept images in batches of max 64 per batch. Cutting list up in batches..")
for batch in chunks(tagged_images_with_regions, 64):
trainer.create_images_from_files(project.id, images=batch, tag_ids=[tag.id for tag in tags.values()])
print("Finished adding images. Visit customvision.ai to start training via the GUI.")
def add_image(trainer, label, project_id, annotation, image_folder):
"""
Add images with labels and bounding box
"""
tagged_images_with_regions = []
tag = trainer.create_tag(project_id, label)
for file_name in annotation.keys():
left, top, width, height = annotation[file_name]
regions = [
Region(tag_id=tag.id, left=left, top=top, width=width, height=height)
]
file_path = os.path.join(image_folder, label, file_name + ".jpg")
with open(file_path, "rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(
name=file_name, contents=image_contents.read(), regions=regions
)
)
image_contents.close()
return tagged_images_with_regions
def _read_annotation_file(self, annotation_path: str) -> list:
annotations = []
with open(annotation_path, "r") as f:
for line in f:
line = line.strip()
parameter_list = line.split(" ")
label = int(parameter_list[0])
x, y, w, h = list(map(float, parameter_list[1:]))
left = x - w / 2
if left < 0: # Accounting for previous rounding error
left = 0
top = y - h / 2
if top < 0: # Accounting for previous rounding error
top = 0
if left + w > 1: # Accounting for previous rounding error
w = 1 - left
if top + h > 1: # Accounting for previous rounding error
h = 1 - top
try:
tag_id = self.label_to_tag_id[label]
except:
raise ValueError(
f"Wrong label {label} at {annotation_path}")
annotations.append(
Region(
tag_id=tag_id,
left=left,
top=top,
width=w,
height=h,
))
return annotations
def train_project(training_key):
trainer = CustomVisionTrainingClient(training_key, endpoint=ENDPOINT)
# Find the object detection domain
obj_detection_domain = next(domain for domain in trainer.get_domains()
if domain.type == "ObjectDetection")
# Create a new project
print("Creating project...")
project = trainer.create_project("My Detection Project",
domain_id=obj_detection_domain.id)
# Make two tags in the new project
fork_tag = trainer.create_tag(project.id, "fork")
scissors_tag = trainer.create_tag(project.id, "scissors")
fork_image_regions = {
"fork_1": [0.145833328, 0.3509314, 0.5894608, 0.238562092],
"fork_2": [0.294117659, 0.216944471, 0.534313738, 0.5980392],
"fork_3": [0.09191177, 0.0682516545, 0.757352948, 0.6143791],
"fork_4": [0.254901975, 0.185898721, 0.5232843, 0.594771266],
"fork_5": [0.2365196, 0.128709182, 0.5845588, 0.71405226],
"fork_6": [0.115196079, 0.133611143, 0.676470637, 0.6993464],
"fork_7": [0.164215669, 0.31008172, 0.767156839, 0.410130739],
"fork_8": [0.118872553, 0.318251669, 0.817401946, 0.225490168],
"fork_9": [0.18259804, 0.2136765, 0.6335784, 0.643790841],
"fork_10": [0.05269608, 0.282303959, 0.8088235, 0.452614367],
"fork_11": [0.05759804, 0.0894935, 0.9007353, 0.3251634],
"fork_12": [0.3345588, 0.07315363, 0.375, 0.9150327],
"fork_13": [0.269607842, 0.194068655, 0.4093137, 0.6732026],
"fork_14": [0.143382356, 0.218578458, 0.7977941, 0.295751631],
"fork_15": [0.19240196, 0.0633497, 0.5710784, 0.8398692],
"fork_16": [0.140931368, 0.480016381, 0.6838235, 0.240196079],
"fork_17": [0.305147052, 0.2512582, 0.4791667, 0.5408496],
"fork_18": [0.234068632, 0.445702642, 0.6127451, 0.344771236],
"fork_19": [0.219362751, 0.141781077, 0.5919118, 0.6683006],
"fork_20": [0.180147052, 0.239820287, 0.6887255, 0.235294119]
}
scissors_image_regions = {
"scissors_1": [0.4007353, 0.194068655, 0.259803921, 0.6617647],
"scissors_2": [0.426470578, 0.185898721, 0.172794119, 0.5539216],
"scissors_3": [0.289215684, 0.259428144, 0.403186262, 0.421568632],
"scissors_4": [0.343137264, 0.105833367, 0.332107842, 0.8055556],
"scissors_5": [0.3125, 0.09766343, 0.435049027, 0.71405226],
"scissors_6": [0.379901975, 0.24308826, 0.32107842, 0.5718954],
"scissors_7": [0.341911763, 0.20714055, 0.3137255, 0.6356209],
"scissors_8": [0.231617644, 0.08459154, 0.504901946, 0.8480392],
"scissors_9": [0.170343131, 0.332957536, 0.767156839, 0.403594762],
"scissors_10": [0.204656869, 0.120539248, 0.5245098, 0.743464053],
"scissors_11": [0.05514706, 0.159754932, 0.799019635, 0.730392158],
"scissors_12": [0.265931368, 0.169558853, 0.5061275, 0.606209159],
"scissors_13": [0.241421565, 0.184264734, 0.448529422, 0.6830065],
"scissors_14": [0.05759804, 0.05027781, 0.75, 0.882352948],
"scissors_15": [0.191176474, 0.169558853, 0.6936275, 0.6748366],
"scissors_16": [0.1004902, 0.279036, 0.6911765, 0.477124184],
"scissors_17": [0.2720588, 0.131977156, 0.4987745, 0.6911765],
"scissors_18": [0.180147052, 0.112369314, 0.6262255, 0.6666667],
"scissors_19": [0.333333343, 0.0274019931, 0.443627447, 0.852941155],
"scissors_20": [0.158088237, 0.04047389, 0.6691176, 0.843137264]
}
# Go through the data table above and create the images
print("Adding images...")
tagged_images_with_regions = []
for file_name in fork_image_regions.keys():
x, y, w, h = fork_image_regions[file_name]
regions = [
Region(tag_id=fork_tag.id, left=x, top=y, width=w, height=h)
]
with open(os.path.join(IMAGES_FOLDER, "fork", file_name + ".jpg"),
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name,
contents=image_contents.read(),
regions=regions))
for file_name in scissors_image_regions.keys():
x, y, w, h = scissors_image_regions[file_name]
regions = [
Region(tag_id=scissors_tag.id, left=x, top=y, width=w, height=h)
]
with open(os.path.join(IMAGES_FOLDER, "scissors", file_name + ".jpg"),
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name,
contents=image_contents.read(),
regions=regions))
trainer.create_images_from_files(project.id,
images=tagged_images_with_regions)
print("Training...")
iteration = trainer.train_project(project.id)
while (iteration.status != "Completed"):
iteration = trainer.get_iteration(project.id, iteration.id)
print("Training status: " + iteration.status)
time.sleep(1)
# The iteration is now trained. Make it the default project endpoint
trainer.update_iteration(project.id, iteration.id, is_default=True)
print("Done!")
return project, iteration
def migrate_images(src_trainer, dest_trainer, project_id, dest_project_id,
created_tags):
# Migrate any tagged images that may exist and preserve their tags and regions.
count = src_trainer.get_tagged_image_count(project_id)
print("Found:", count, "tagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_tagged_images(project_id,
take=count_to_migrate,
skip=migrated)
images_to_upload = []
for i in images:
print("Migrating", i.id, i.original_image_uri)
if i.regions:
regions = []
for r in i.regions:
print("Found region:", r.region_id, r.tag_id, r.left,
r.top, r.width, r.height)
regions.append(
Region(tag_id=created_tags[r.tag_id],
left=r.left,
top=r.top,
width=r.width,
height=r.height))
entry = ImageUrlCreateEntry(url=i.original_image_uri,
regions=regions)
else:
tag_ids = []
for t in i.tags:
print("Found tag:", t.tag_name, t.tag_id)
tag_ids.append(created_tags[t.tag_id])
entry = ImageUrlCreateEntry(url=i.original_image_uri,
tag_ids=tag_ids)
images_to_upload.append(entry)
upload_result = dest_trainer.create_images_from_urls(
dest_project_id, images=images_to_upload)
if not upload_result.is_batch_successful:
print("ERROR: Failed to upload image batch")
for i in upload_result.images:
print("\tImage status:", i.id, i.status)
exit(-1)
migrated += count_to_migrate
count -= count_to_migrate
# Migrate any untagged images that may exist.
count = src_trainer.get_untagged_image_count(project_id)
print("Found:", count, "untagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_untagged_images(project_id,
take=count_to_migrate,
skip=migrated)
images_to_upload = []
for i in images:
print("Migrating", i.id, i.original_image_uri)
images_to_upload.append(
ImageUrlCreateEntry(url=i.original_image_uri))
upload_result = dest_trainer.create_images_from_urls(
dest_project_id, images=images_to_upload)
if not upload_result.is_batch_successful:
print("ERROR: Failed to upload image batch")
for i in upload_result.images:
print("\tImage status:", i.id, i.status)
exit(-1)
migrated += count_to_migrate
count -= count_to_migrate
return images
balloon_image_regions[jpgName] = regions
print(balloon_image_regions)
# Update this with the path to where you downloaded the images.
base_image_url = "./images/"
# Go through the data table above and create the images
print("Adding images...")
tagged_images_with_regions = []
for filename in balloon_image_regions.keys():
regions = []
for region in balloon_image_regions[filename]:
x, y, w, h = region
regions.append(
Region(tag_id=balloon_tag.id, left=x, top=y, width=w, height=h))
with open(base_image_url + filename, mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=filename,
contents=image_contents.read(),
regions=regions))
upload_result = trainer.create_images_from_files(
project.id, images=tagged_images_with_regions)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
print("Image status: ", image.status)
exit(-1)
def upload_images_to_customvision_helper(project_id,
part_id,
batch_size: int = 10) -> bool:
"""upload_images_to_customvision_helper.
Helper function for uploading images to Custom Vision.
Make sure part already upload to Custom Vision (
customvision_id not null or blank).
Args:
project_id:
part_id:
batch_size (int): batch_size
Returns:
bool:
"""
logger.info("Uploading images with part_id %s", part_id)
has_new_images = False
project_obj = Project.objects.get(pk=project_id)
trainer = project_obj.setting.get_trainer_obj()
part_obj: Part = Part.objects.get(pk=part_id)
tag_id = part_obj.customvision_id
images = Image.objects.filter(part_id=part_id,
is_relabel=False,
uploaded=False)
logger.info("Tag id %s", tag_id)
count = 0
img_entries = []
img_objs = []
logger.info("Image length: %s", len(images))
for index, image_obj in enumerate(images):
logger.info("*** image %s, %s", index + 1, image_obj)
has_new_images = True
img_name = "img-" + datetime.datetime.utcnow().isoformat()
regions = []
width = image_obj.image.width
height = image_obj.image.height
try:
labels = json.loads(image_obj.labels)
if len(labels) == 0:
continue
for label in labels:
label_x = label["x1"] / width
label_y = label["y1"] / height
label_w = (label["x2"] - label["x1"]) / width
label_h = (label["y2"] - label["y1"]) / height
region = Region(tag_id=tag_id,
left=label_x,
top=label_y,
width=label_w,
height=label_h)
regions.append(region)
image = image_obj.image
image.open()
img_entry = ImageFileCreateEntry(name=img_name,
contents=image.read(),
regions=regions)
img_objs.append(image_obj)
img_entries.append(img_entry)
count += 1
except:
logger.exception("unexpected error")
if len(img_entries) >= batch_size:
logger.info("Uploading %s images", len(img_entries))
upload_result = trainer.create_images_from_files(
project_id=project_obj.customvision_project_id,
images=img_entries)
logger.info(
"Uploading images... Is batch success: %s",
upload_result.is_batch_successful,
)
img_entries = []
for i, img_obj in enumerate(img_objs):
img_obj.customvision_id = upload_result.images[i].image.id
img_obj.uploaded = True
img_obj.save()
img_objs = []
if len(img_entries) >= 1:
logger.info("Uploading %s images", len(img_entries))
upload_result = trainer.create_images_from_files(
project_id=project_obj.customvision_project_id, images=img_entries)
logger.info(
"Uploading images... Is batch success: %s",
upload_result.is_batch_successful,
)
for i, img_obj in enumerate(img_objs):
img_obj.customvision_id = upload_result.images[i].image.id
img_obj.remote_url = (
upload_result.images[i].image.original_image_uri)
img_obj.uploaded = True
img_obj.save()
logger.info("Uploading images... Done")
logger.info("Has new images: %s", has_new_images)
return has_new_images
def main(req: func.HttpRequest) -> func.HttpResponse:
logging.info(
'MLP Pbject Detection HTTP trigger function AddLabeledData processed a request.'
)
try:
image_url = req.form.get('ImageUrl')
image_labeling_json = req.form.get('DataLabels')
except:
return func.HttpResponse(
"Please pass JSON containing the labeled regions associated with this image on the query string or in the request body.",
status_code=400)
if image_url and image_labeling_json:
labels = []
labeled_images_with_regions = []
count_of_regions_applied_to_image = 0
count_of_labels_applied_to_region = 0
endpoint = os.environ['ClientEndpoint']
# Get Cognitive Services Environment Variables
project_id = os.environ["ProjectID"]
training_key = os.environ['TrainingKey']
# load labeled image regions passed in request into dictionary
image_labeling_data = json.loads(image_labeling_json)
# instanciate custom vision client
trainer = CustomVisionTrainingClient(training_key, endpoint=endpoint)
# get list of valid tags for this model
tags = trainer.get_tags(project_id)
# get the height and width of the image
image_width = image_labeling_data['asset']['size']['width']
image_height = image_labeling_data['asset']['size']['height']
# for each labeled region in this asset get the map lables to tag ids and map boundaries formats
# from vott to azure cognitive services then upload to the cognitive services project.
for labeled_region in image_labeling_data['regions']:
for label in labeled_region['tags']:
for tag in tags:
if tag.name == label:
labels.append(tag.id)
count_of_labels_applied_to_region = count_of_labels_applied_to_region + 1
break
if count_of_labels_applied_to_region > 0:
count_of_regions_applied_to_image = count_of_regions_applied_to_image + 1
else:
return func.HttpResponse(
"This Azure Cognitive Services Object Detection project does not contain any labeling tags. Please add tags to the project before attempting to add labeled data into the project.",
status_code=400)
top_left_x = labeled_region['points'][0]['x']
top_left_y = labeled_region['points'][0]['y']
top_right_x = labeled_region['points'][1]['x']
top_right_y = labeled_region['points'][1]['y']
bottom_right_x = labeled_region['points'][2]['x']
bottom_right_y = labeled_region['points'][2]['y']
bottom_left_x = labeled_region['points'][3]['x']
bottom_left_y = labeled_region['points'][3]['y']
# calculate normalized coordinates.
normalized_left = top_left_x / image_width
normalized_top = top_left_y / image_height
normalized_width = (top_right_x - top_left_x) / image_width
normalized_height = (bottom_left_y - top_left_y) / image_height
regions = [
Region(tag_id=labels[0],
left=normalized_left,
top=normalized_top,
width=normalized_width,
height=normalized_height)
]
labeled_images_with_regions.append(
ImageUrlCreateEntry(url=image_url, regions=regions))
upload_result = trainer.create_images_from_urls(
project_id, images=labeled_images_with_regions)
result = ""
if upload_result.is_batch_successful:
for image in upload_result.images:
result = result + "Image " + image.source_url + " Status: " + image.status + ", "
return func.HttpResponse(
"Images successfully uploaded with " +
str(count_of_regions_applied_to_image) + " regions and " +
str(count_of_labels_applied_to_region) +
" label(s) to project " + project_id + "with result: " +
result,
status_code=200)
else:
success = True
for image in upload_result.images:
result = result + "Image " + image.source_url + " Status: " + image.status + ", "
if not "ok" in image.status.lower():
success = False
if success:
return func.HttpResponse(
"Image batch upload succeeded with result: " + result,
status_code=200)
else:
return func.HttpResponse(
"Image batch upload failed with result: " + result,
status_code=400)
else:
return func.HttpResponse(
"Please pass valid a valid image url and labels in the request body using the parameter names: ImageUrl and DataLabels.",
status_code=400)
def upload_directory(self,
data_dir,
img_ext="*.jpg",
img_dir="images",
lbl_file="labels.csv",
default_tag_name="important"):
"""
upload_directory - Upload images from a given directory into the CV workspace
:param str data_dir: Source folder of the files.
:param str img_ext: image extension.
:param str img_dir: image folder.
:param str lbl_file: labels file.
:param str default_tag_name: default tag name.
:returns: None
"""
label_fn = os.path.join(data_dir, lbl_file)
img_folder = os.path.join(data_dir, img_dir)
# Check if required folders exist.
if not (os.path.isdir(img_folder) and os.path.exists(label_fn)):
print("Input data not found")
return
# Read labels and image list.
labels_df = pd.read_csv(os.path.join(label_fn))
image_list = glob.glob(os.path.join(img_folder, img_ext))
# Upload each image with regions
for _, path in enumerate(image_list):
tagged_images_with_regions = []
regions = []
file_name = path.split("\\")[-1]
img = Image.open(path)
img_width, img_height = img.size
for _, row in labels_df[labels_df.FileName ==
file_name].iterrows():
x, y, w, h = row.XMin, row.YMin, row.XMax - row.XMin, row.YMax - row.YMin
x = x / img_width
w = w / img_width
y = y / img_height
h = h / img_height
if "DefectType" in row:
default_tag_name = row.DefectType
tag = None
for a_tag in self.tags:
if a_tag.name == default_tag_name:
tag = a_tag
if not tag:
tag = self.client.create_tag(self.project_id,
default_tag_name)
self.tags = self.client.get_tags(self.project_id)
regions.append(
Region(tag_id=tag.id, left=x, top=y, width=w, height=h))
with open(path, mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name,
contents=image_contents.read(),
regions=regions))
upload_result = self.client.create_images_from_files(
self.project.id, images=tagged_images_with_regions)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
print("Image status: ", image.status)
def detect(
cfg,
data,
weights,
images='data/samples', # input folder
output='output', # output folder
fourcc='mp4v', # video codec
img_size=416,
conf_thres=0.8,
nms_thres=0.8,
save_txt=False,
save_images=True):
# Initialize
device = torch_utils.select_device(force_cpu=ONNX_EXPORT)
torch.backends.cudnn.benchmark = False # set False for reproducible results
if os.path.exists(output):
shutil.rmtree(output) # delete output folder
os.makedirs(output) # make new output folder
# Initialize model
if ONNX_EXPORT:
s = (
320, 192
) # (320, 192) or (416, 256) or (608, 352) onnx model image size (height, width)
model = Darknet(cfg, s)
else:
model = Darknet(cfg, img_size)
# Load weights
if weights.endswith('.pt'): # pytorch format
model.load_state_dict(
torch.load(weights, map_location=device)['model'])
else: # darknet format
_ = load_darknet_weights(model, weights)
# Eval mode
model.to(device).eval()
# Export mode
if ONNX_EXPORT:
img = torch.zeros((1, 3, s[0], s[1]))
torch.onnx.export(model, img, 'weights/export.onnx', verbose=True)
return
# Half precision
opt.half = opt.half and device.type != 'cpu' # half precision only supported on CUDA
if opt.half:
model.half()
# Set Dataloader
vid_path, vid_writer = None, None
if opt.webcam:
save_images = False
dataloader = LoadWebcam(img_size=img_size, half=opt.half)
else:
save_img = True
dataloader = LoadImages(images, img_size=img_size, half=opt.half)
# Get classes and colors
classes = load_classes(parse_data_cfg(data)['names'])
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(classes))]
#opvragen van de tag id van het custom vision project
auto_tag = next(
filter(lambda t: t.name == "auto", trainer.get_tags(project.id)), None)
# Run inference
t0 = time.time()
for i, (path, img, im0, vid_cap) in enumerate(dataloader):
t = time.time()
save_path = str(Path(output) / Path(path).name)
foto = cv2.imread(path)
h, w = foto.shape[:2]
# Get detections
img = torch.from_numpy(img).unsqueeze(0).to(device)
pred, _ = model(img)
det = non_max_suppression(pred.float(), conf_thres, nms_thres)[0]
#
if det is not None and len(det) > 0:
# Rescale boxes from 416 to true image size
det[:, :4] = scale_coords(img.shape[2:], det[:, :4],
im0.shape).round()
# Print results to screen
print('%gx%g ' % img.shape[2:], end='') # print image size
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum()
print('%g %ss' % (n, classes[int(c)]), end=', ')
lijst = []
# Draw bounding boxes and labels of detections
for *xyxy, conf, cls_conf, cls in det:
if save_img: # Add bbox to image
label = '%s %.2f' % (classes[int(cls)], conf)
plot_one_box(xyxy,
im0,
label=label,
color=colors[int(cls)])
#if save_txt: # Write to file
if int(cls) == 2:
x1 = float(('{0}').format(*xyxy, )) / w
y1 = float(('{1}').format(*xyxy, )) / h
x2 = float(('{2}').format(*xyxy, )) / w
y2 = float(('{3}').format(*xyxy, )) / h
breedte = (x2 - x1)
hoogte = (y2 - y1)
coordinaten = [x1, y1, breedte, hoogte]
kleinelijst = [os.path.basename(path), coordinaten]
lijst.append(kleinelijst)
base_image_url = Path("data/samples/")
# Go through the data table above and create the images
print("Adding images...")
tagged_images_with_regions = []
auto_image_regions = {}
dictitem = {auto_tag.id: lijst}
auto_image_regions.update(dictitem)
for tag_id in auto_image_regions:
for filename, [x, y, w, h] in auto_image_regions[tag_id]:
print("filename = " + str(filename))
#x,y,w,h = auto_image_regions[filename]
regions = [
Region(tag_id=auto_tag.id,
left=x,
top=y,
width=w,
height=h)
]
with open(str(base_image_url) + "/" + filename,
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(
name=filename,
contents=image_contents.read(),
regions=regions))
upload_result = trainer.create_images_from_files(
project.id, images=tagged_images_with_regions)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
print("Image status: ", image.status)
print('Done. (%.3fs)' % (time.time() - t))
if opt.webcam: # Show live webcam
cv2.imshow(weights, im0)
if save_images: # Save image with detections
if dataloader.mode == 'images':
cv2.imwrite(save_path, im0)
else:
if vid_path != save_path: # new video
vid_path = save_path
if isinstance(vid_writer, cv2.VideoWriter):
vid_writer.release() # release previous video writer
fps = vid_cap.get(cv2.CAP_PROP_FPS)
width = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
vid_writer = cv2.VideoWriter(
save_path, cv2.VideoWriter_fourcc(*fourcc), fps,
(width, height))
vid_writer.write(im0)
if save_images:
print('Results saved to %s' % os.getcwd() + os.sep + output)
if platform == 'darwin': # macos
os.system('open ' + output + ' ' + save_path)
print('Done. (%.3fs)' % (time.time() - t0))
path = "data/samples/"
path2 = "data/finished/"
print("we kopiëren alle foto's nu")
src_files = os.listdir(path)
for file_name in src_files:
full_file_name = os.path.join(path, file_name)
if os.path.isfile(full_file_name):
shutil.copy(full_file_name, path2)
print("we verwijderen alle foto's nu")
for the_file in os.listdir(path):
file_path = os.path.join(path, the_file)
if os.path.isfile(file_path):
os.unlink(file_path)
def _train(project_id):
project_obj = Project.objects.get(pk=project_id)
customvision_project_id = project_obj.customvision_project_id
# @FIXME (Hugh): wrap it up
obj, created = Train.objects.update_or_create(
project=project_obj,
defaults={
'status': 'Status: preparing data (images and annotations)',
'log': '',
'project': project_obj
})
Trainer.dequeue_iterations(
trainer=trainer, custom_vision_project_id=customvision_project_id)
try:
count = 10
while count > 0:
part_ids = [part.id for part in project_obj.parts.all()]
if len(part_ids) > 0: break
print('waiting parts...')
time.sleep(1)
count -= 1
# @FIXME (Hugh): wrap it up
obj, created = Train.objects.update_or_create(
project=project_obj,
defaults={
'status': 'sending',
'log': 'Status : sending data (images and annotations)',
'project': project_obj
})
print(project_obj.id)
print('Part ids:', part_ids, flush=True)
images = Image.objects.filter(part_id__in=part_ids,
is_relabel=False,
uploaded=False).all()
img_entries = []
img_objs = []
tags = trainer.get_tags(customvision_project_id)
tag_dict = {}
for tag in tags:
tag_dict[tag.name] = tag.id
print('[INFO] Submit images and do the Training...')
count = 0
for index, image_obj in enumerate(images):
print('*** image', index + 1, image_obj, flush=True)
part = image_obj.part
part_name = part.name
if part_name not in tag_dict:
print('new part name', part_name, flush=True)
print('creating new tag (part name)', flush=True)
tag = trainer.create_tag(customvision_project_id, part_name)
tag_dict[tag.name] = tag.id
tag_id = tag_dict[part_name]
name = 'img-' + datetime.datetime.utcnow().isoformat()
regions = []
width = image_obj.image.width
height = image_obj.image.height
try:
labels = json.loads(image_obj.labels)
if len(labels) == 0: continue
for label in labels:
x = label['x1'] / width
y = label['y1'] / height
w = (label['x2'] - label['x1']) / width
h = (label['y2'] - label['y1']) / height
region = Region(tag_id=tag_id,
left=x,
top=y,
width=w,
height=h)
regions.append(region)
image = image_obj.image
image.open()
img_entry = ImageFileCreateEntry(name=name,
contents=image.read(),
regions=regions)
img_objs.append(image_obj)
img_entries.append(img_entry)
count += 1
except:
print("[ERROR] Unexpected error:",
sys.exc_info()[0],
flush=True)
if len(img_entries) >= 5:
print('uploading...', flush=True)
upload_result = trainer.create_images_from_files(
customvision_project_id, images=img_entries)
print('batch success:',
upload_result.is_batch_successful,
flush=True)
img_entries = []
for img_obj in img_objs:
img_obj.uploaded = True
img_obj.save()
img_objs = []
if len(img_entries) >= 1:
print('uploading...', flush=True)
upload_result = trainer.create_images_from_files(
customvision_project_id, images=img_entries)
print('batch success:',
upload_result.is_batch_successful,
flush=True)
for img_obj in img_objs:
img_obj.uploaded = True
img_obj.save()
if count == 0:
print('Nothing changed, no training', flush=True)
# @FIXME (Hugh): wrap it up
obj, created = Train.objects.update_or_create(
project=project_obj,
defaults={
'status': 'ok',
'log': 'Status: Nothing changed, no training',
'project': project_obj
})
else:
print('training...', flush=True)
try:
trainer.train_project(customvision_project_id)
project_obj.deployed = False
project_obj.save(update_fields=['deployed'])
update_train_status(project_id)
print('[INFO] set deployed = False')
except CustomVisionErrorException:
print(
'[ERROR] From Custom Vision: Nothing changed since last training',
flush=True)
raise
return JsonResponse({'status': 'ok'})
except Exception as e:
err_msg = traceback.format_exc()
print(f'Exception: {err_msg}')
# @FIXME (Hugh): wrap it up
obj, created = Train.objects.update_or_create(
project=project_obj,
defaults={
'status': 'failed',
'log': f'Status : failed {str(err_msg)}',
'project': project_obj
})
return JsonResponse({
'status': 'failed',
'log': f'Status : failed {str(err_msg)}'
})
fullpath = os.path.join(basePath, "cvjson.json")
print("Loading Json: %s " % (fullpath))
with open(fullpath) as f:
imgList = json.load(f)
print("Adding Images")
for data in imgList:
filePath = os.path.join(basePath, "img", data.get("Name"))
print("Processing %s" % (data.get("Name")))
regions = [
Region(tag_id=tag.id,
left=data.get("x"),
top=data.get("y"),
width=data.get("width"),
height=data.get("height"))
]
#regions = [ Region(tag_id=fork_tag.id, left=x,top=y,width=w,height=h) ]
with open(filePath, mode="rb") as image_contents:
#There is a limit of 64 images and 20 tags.
tagged_images_with_regions.append(
ImageFileCreateEntry(name=data.get("Name"),
contents=image_contents.read(),
regions=regions))
trainer.create_images_from_files(project.id,
images=tagged_images_with_regions,
def migrate_images(src_trainer, dest_trainer, project_id, dest_project,
created_tags):
# Migrate any tagged images that may exist and preserve their tags and regions.
count = src_trainer.get_tagged_image_count(project_id)
print("Found:", count, "tagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_tagged_images(project_id,
take=count_to_migrate,
skip=migrated)
images_to_upload = []
for i in images:
print("Migrating", i.id, i.original_image_uri)
if i.regions:
regions = []
tag_ids = []
for r in i.regions:
print("Found region:", r.region_id, r.tag_id, r.left,
r.top, r.width, r.height)
regions.append(
Region(tag_id=created_tags[r.tag_id],
left=r.left,
top=r.top,
width=r.width,
height=r.height))
tag_ids.append(created_tags[r.tag_id])
entry = ImageUrlCreateEntry(url=i.original_image_uri,
regions=regions)
else:
tag_ids = []
for t in i.tags:
print("Found tag:", t.tag_name, t.tag_id)
tag_ids.append(created_tags[t.tag_id])
entry = ImageUrlCreateEntry(url=i.original_image_uri,
tag_ids=tag_ids)
images_to_upload.append(entry)
image_file = '{}-{}.jpg'.format(i.id, "_".join(tag_ids))
xml_file = '{}-{}.xml'.format(i.id, "_".join(tag_ids))
image = None
if not os.path.exists('export/' + image_file):
r = requests.get(i.original_image_uri)
with open('export/' + image_file, 'wb') as f:
f.write(r.content)
image = Image.open(BytesIO(r.content))
else:
image = Image.open('export/' + image_file)
w, h = image.size
annotation = ET.Element('annotation')
folder = ET.SubElement(annotation, 'folder')
filename = ET.SubElement(annotation, 'filename')
filename.text = image_file
path = ET.SubElement(annotation, 'path')
path.text = image_file
source = ET.SubElement(annotation, 'source')
database = ET.SubElement(source, 'database')
database.text = 'Egge'
size = ET.SubElement(annotation, 'size')
ET.SubElement(size, 'width').text = str(w)
ET.SubElement(size, 'height').text = str(h)
ET.SubElement(size, 'depth').text = '3'
ET.SubElement(annotation, 'segmented').text = '0'
for r in i.regions:
_object = ET.SubElement(annotation, 'object')
ET.SubElement(_object, 'name').text = created_tags[r.tag_id]
ET.SubElement(_object, 'pose').text = 'Unspecified'
ET.SubElement(_object, 'truncated').text = '0'
ET.SubElement(_object, 'difficult').text = '0'
ET.SubElement(_object, 'occluded').text = '0'
bndbox = ET.SubElement(_object, 'bndbox')
ET.SubElement(bndbox, 'xmin').text = str(int(r.left * w))
ET.SubElement(bndbox,
'xmax').text = str(int((r.left + r.width) * w))
ET.SubElement(bndbox, 'ymin').text = str(int(r.top * h))
ET.SubElement(bndbox,
'ymax').text = str(int((r.top + r.height) * h))
xmlstr = minidom.parseString(
ET.tostring(annotation)).toprettyxml(indent=" ")
with open('export/' + xml_file, "wb") as f:
f.write(xmlstr.encode('utf-8'))
if dest_trainer:
upload_result = dest_trainer.create_images_from_urls(
dest_project.id, images=images_to_upload)
if not upload_result.is_batch_successful:
print("ERROR: Failed to upload image batch")
for i in upload_result.images:
print("\tImage status:", i.id, i.status)
exit(-1)
migrated += count_to_migrate
count -= count_to_migrate
# Migrate any untagged images that may exist.
count = src_trainer.get_untagged_image_count(project_id)
print("Found:", count, "untagged images.")
migrated = 0
while (count > 0):
count_to_migrate = min(count, 50)
print("Getting", count_to_migrate, "images")
images = src_trainer.get_untagged_images(project_id,
take=count_to_migrate,
skip=migrated)
images_to_upload = []
for i in images:
print("Migrating", i.id, i.original_image_uri)
images_to_upload.append(
ImageUrlCreateEntry(url=i.original_image_uri))
upload_result = dest_trainer.create_images_from_urls(
dest_project.id, images=images_to_upload)
if not upload_result.is_batch_successful:
print("ERROR: Failed to upload image batch")
for i in upload_result.images:
print("\tImage status:", i.id, i.status)
exit(-1)
migrated += count_to_migrate
count -= count_to_migrate
return images
start = 0
for stop in range(50, len(file_list), 50):
#11312
#2771
filelist = file_list[start:stop]
tagged_images_with_regions = []
for file in filelist:
regions_ = []
for _, row in df[df.filename == file].iterrows():
#print(row)
regions_.append(
Region(tag_id=label_dict[row['class']],
left=row.left,
top=row.top,
width=row.width,
height=row.height))
with open(base_image_url + row.filename, mode='rb') as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=row.filename,
contents=image_contents.read(),
regions=regions_))
## Upload results
upload_result = trainer.create_images_from_files(
project.id, images=tagged_images_with_regions)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
regions_lis.append((ind, x, y, w, h))
images.append((k, regions_lis))
print("Adding images...")
print("%d images in total" % len(images))
SEND_BY_BATCH = 64
print("Send by batch =", SEND_BY_BATCH)
tagged_images_with_regions = []
BATCH_NUM = math.ceil(len(images) / SEND_BY_BATCH)
for i in range(BATCH_NUM):
for file_name, regions in images[i * SEND_BY_BATCH:(i + 1) *
SEND_BY_BATCH]:
regions_saved = []
for r in regions:
x, y, w, h = r[1:]
tag_id = tag_index_to_id[r[0]]
regions_saved.append(
Region(tag_id=tag_id, left=x, top=y, width=w, height=h))
with open(IMAGES_PATH + "/" + file_name + "." + FILE_EXT,
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name,
contents=image_contents.read(),
regions=regions_saved))
print("Batch no.", i)
trainer.create_images_from_files(project_id,
images=tagged_images_with_regions)
tagged_images_with_regions = []
# try to upload generated images
tag_name_to_id = {}
for t in all_tags:
tag_name_to_id[t.name] = t.id
# each element is (image file name, list of regions)
# each region is a tuple where the first element is the tag and the remaining are x, y, w, h.
images = [
("apple",[("Apple", 0.15492957746478872, 0.030985915492957747, 0.6633802816901408, 0.8802816901408451)]),
("banana",[("Banana", 0.034739454094292806, 0.04084507042253521, 0.9255583126550869, 0.919718309859155)]),
("fruits",[("Apple", 0.200845665961945, 0.09859154929577464, 0.2706131078224101, 0.347887323943662),
("Apple", 0.5190274841437632, 0.11549295774647887, 0.2420718816067653, 0.3225352112676056),
("Banana", 0.19978858350951373, 0.476056338028169, 0.595137420718816, 0.4169014084507043),
]),
]
print ("Adding images...")
tagged_images_with_regions = []
for file_name, regions in images:
regions_saved = []
for r in regions:
x,y,w,h = r[1:]
tag_id = tag_name_to_id[r[0]]
regions_saved.append(Region(tag_id=tag_id, left=x,top=y,width=w,height=h))
with open("test_upload_images/" + file_name + ".jpg", mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name, contents=image_contents.read(), regions=regions_saved)
)
trainer.create_images_from_files(project.id, images=tagged_images_with_regions)
region_count =0
for region in vehicle_regions[image]:
if region_count >= 200:
break
img = Image.open(BytesIO(response.content))
width = img.size[0]
height = img.size[1]
#normalize region coordinates
x = float((region['coordinates']['x']/width))
y = float((region['coordinates']['y']/height))
w = float((region['coordinates']['w']/width))
h = float((region['coordinates']['h']/height))
regions.append( Region(tag_id=tags[0].id, left=x, top=y, width=w, height=h) )
region_count+=1
tagged_images_with_regions.append(ImageFileCreateEntry(name=image, contents=response.content, regions=regions))
#uploading labeled images to custom vision portal
print('Uploading images to trainer')
for i in range(0,len(tagged_images_with_regions),63):
batch = tagged_images_with_regions[i:i+63]
upload_result = trainer.create_images_from_files(project.id, images=batch)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
print("Image status: ", image.status)
print(tagtoidmapper)
tagged_images_with_regions = []
z = 1
for image in images:
if z % 50 == 0:
upload_result = trainAPI.create_images_from_files(
project.id, images=tagged_images_with_regions)
tagged_images_with_regions = []
regions = []
for region in image['regions']:
regions.append(
Region(tag_id=tagtoidmapper[region['tagName']],
left=region['left'],
top=region['top'],
width=region['width'],
height=region['height']))
image_file = open(os.path.join(imagesPath, image['id'] + '.png'), 'rb')
tagged_images_with_regions.append(
ImageFileCreateEntry(name=image['id'],
contents=image_file.read(),
regions=regions))
image_file.close()
z = z + 1
upload_result = trainAPI.create_images_from_files(
project.id, images=tagged_images_with_regions)
def _train(project_id):
"""Actually do uplaod, train and deploy"""
project_obj = Project.objects.get(pk=project_id)
trainer = project_obj.setting.revalidate_and_get_trainer_obj()
customvision_project_id = project_obj.customvision_project_id
# Invalid Endpoint + Training Key
if not trainer:
project_obj.upcreate_training_status(
status="failed", log=error_messages.CUSTOM_VISION_ACCESS_ERROR)
return JsonResponse(
{
"status": "failed",
"log": error_messages.CUSTOM_VISION_ACCESS_ERROR
},
status=503,
)
project_obj.upcreate_training_status(
status="Status: preparing data (images and annotations)", log="")
project_obj.dequeue_iterations()
try:
count = 10
while count > 0:
part_ids = [part.id for part in project_obj.parts.all()]
if len(part_ids) > 0:
break
logging.info("waiting parts...")
time.sleep(1)
count -= 1
logger.info("Project id: %s", project_obj.id)
logger.info("Part ids: %s", part_ids)
try:
trainer.get_project(customvision_project_id)
project_obj.upcreate_training_status(
status="preparing",
log=(
f"Status : Project {project_obj.customvision_project_name}"
+ "found on Custom Vision"),
)
except:
project_obj.create_project()
project_obj.upcreate_training_status(
status="preparing",
log=("Status : Project created on CustomVision. " +
f"Name: {project_obj.customvision_project_name}"),
)
logger.info("Project created on CustomVision.")
logger.info("Project Id: %s", project_obj.customvision_project_id)
logger.info("Project Name: %s",
project_obj.customvision_project_name)
customvision_project_id = project_obj.customvision_project_id
project_obj.upcreate_training_status(
status="sending",
log="Status : sending data (images and annotations)")
tags = trainer.get_tags(customvision_project_id)
tag_dict = {}
project_partnames = {}
project_changed = False
has_new_parts = False
has_new_images = False
for tag in tags:
tag_dict[tag.name] = tag.id
parts_last_train = len(tags)
images_last_train = trainer.get_tagged_image_count(
project_obj.customvision_project_id)
# Update existing tags
# TODO: update tags
# trainer.update_tags(project_id, tag_id, new_tag)
# Create tags on CustomVisioin Project
# Maybe move to Project Model?
logger.info("Creating tags before training...")
counter = 0
for part_id in part_ids:
part_name = Part.objects.get(id=part_id).name
part_description = Part.objects.get(id=part_id).description
project_partnames[part_name] = "foo"
if part_name not in tag_dict:
logger.info("Creating tag: %s. Description: %s", part_name,
part_description)
tag = trainer.create_tag(
project_id=customvision_project_id,
name=part_name,
description=part_description,
)
has_new_parts = True
tag_dict[tag.name] = tag.id
counter += 1
project_changed = project_changed or (counter > 0)
logger.info("Created %s tags", counter)
logger.info("Creating tags... Done")
# Upload images to CustomVisioin Project
images = Image.objects.filter(part_id__in=part_ids,
is_relabel=False,
uploaded=False).all()
logger.info("Uploading images before training...")
count = 0
img_entries = []
img_objs = []
logger.info("Image length: %s", len(images))
for index, image_obj in enumerate(images):
logger.info("*** image %s, %s", index + 1, image_obj)
has_new_images = True
part = image_obj.part
part_name = part.name
tag_id = tag_dict[part_name]
img_name = "img-" + datetime.datetime.utcnow().isoformat()
regions = []
width = image_obj.image.width
height = image_obj.image.height
try:
labels = json.loads(image_obj.labels)
if len(labels) == 0:
continue
for label in labels:
x = label["x1"] / width
y = label["y1"] / height
w = (label["x2"] - label["x1"]) / width
h = (label["y2"] - label["y1"]) / height
region = Region(tag_id=tag_id,
left=x,
top=y,
width=w,
height=h)
regions.append(region)
image = image_obj.image
image.open()
img_entry = ImageFileCreateEntry(name=img_name,
contents=image.read(),
regions=regions)
img_objs.append(image_obj)
img_entries.append(img_entry)
project_changed = project_changed or (not image_obj.uploaded)
if project_changed:
logger.info("project_changed: %s", project_changed)
count += 1
except:
logger.exception("unexpected error")
if len(img_entries) >= 5:
logger.info("Uploading %s images", len(img_entries))
upload_result = trainer.create_images_from_files(
customvision_project_id, images=img_entries)
logger.info(
"Uploading images... Is batch success: %s",
upload_result.is_batch_successful,
)
img_entries = []
for img_obj in img_objs:
img_obj.uploaded = True
img_obj.save()
img_objs = []
if len(img_entries) >= 1:
logger.info("Uploading %s images", len(img_entries))
upload_result = trainer.create_images_from_files(
customvision_project_id, images=img_entries)
logger.info(
"Uploading images... Is batch success: %s",
upload_result.is_batch_successful,
)
for img_obj in img_objs:
img_obj.uploaded = True
img_obj.save()
logger.info("Uploading images... Done")
# Submit training task to Custom Vision
if not project_changed:
project_obj.upcreate_training_status(
status="ok", log="Status: Nothing changed. Not training")
else:
project_obj.upcreate_training_status(
status="training",
log="Status: Project changed. Submitting training task...")
training_task_submit_success = project_obj.train_project()
if training_task_submit_success:
project_obj.update_app_insight_counter(
has_new_parts=has_new_parts,
has_new_images=has_new_images,
parts_last_train=parts_last_train,
images_last_train=images_last_train,
)
# A Thread/Task to keep updating the status
update_train_status(project_id)
return JsonResponse({"status": "ok"})
except CustomVisionErrorException as customvision_err:
logger.error("CustomVisionErrorException: %s", customvision_err)
if customvision_err.message == \
"Operation returned an invalid status code 'Access Denied'":
project_obj.upcreate_training_status(
status="failed",
log=
"Training key or Endpoint is invalid. Please change the settings",
)
return JsonResponse(
{
"status":
"failed",
"log":
"Training key or Endpoint is invalid. Please change the settings",
},
status=status.HTTP_503_SERVICE_UNAVAILABLE,
)
project_obj.upcreate_training_status(status="failed",
log=customvision_err.message)
return JsonResponse(
{
"status": "failed",
"log": customvision_err.message
},
status=status.HTTP_400_BAD_REQUEST,
)
except Exception as e:
# TODO: Remove in production
err_msg = traceback.format_exc()
logger.exception("Exception: %s", err_msg)
project_obj.upcreate_training_status(
status="failed", log=f"Status : failed {str(err_msg)}")
return JsonResponse({
"status": "failed",
"log": f"Status : failed {str(err_msg)}"
})
"scissors_17": [0.2720588, 0.131977156, 0.4987745, 0.6911765],
"scissors_18": [0.180147052, 0.112369314, 0.6262255, 0.6666667],
"scissors_19": [0.333333343, 0.0274019931, 0.443627447, 0.852941155],
"scissors_20": [0.158088237, 0.04047389, 0.6691176, 0.843137264]
}
# Update this with the path to where you downloaded the images.
base_image_location = "<path to directory"
# Go through the data table above and create the images
print("Adding images...")
tagged_images_with_regions = []
for file_name in fork_image_regions.keys():
x, y, w, h = fork_image_regions[file_name]
regions = [Region(tag_id=fork_tag.id, left=x, top=y, width=w, height=h)]
with open(base_image_location + "images/fork/" + file_name + ".jpg",
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=file_name,
contents=image_contents.read(),
regions=regions))
for file_name in scissors_image_regions.keys():
x, y, w, h = scissors_image_regions[file_name]
regions = [
Region(tag_id=scissors_tag.id, left=x, top=y, width=w, height=h)
]
with open(base_image_location + "images/scissors/" + file_name + ".jpg",
tagged_images_with_regions = []
for index, row in group.iterrows():
width = row['x2'] - row['x1']
height = row['y2'] - row['y1']
x_scaled = row['x1'] / image_width
y_scaled = row['y1'] / image_height
width_scaled = width / image_width
height_scaled = height / image_height
regions.append(
Region(tag_id=tags[row['defect']].id,
left=x_scaled,
top=y_scaled,
width=width_scaled,
height=height_scaled))
with open(os.path.join('./data', "{0}.png".format(name)),
mode="rb") as image_contents:
tagged_images_with_regions.append(
ImageFileCreateEntry(name=name,
contents=image_contents.read(),
regions=regions))
upload_result = trainer.create_images_from_files(
project.id, images=tagged_images_with_regions)
if not upload_result.is_batch_successful:
print("Image batch upload failed.")
for image in upload_result.images:
