def create_new_model(cls,
book: DatabaseBook,
id: Optional[str] = None) -> Model:
import datetime
id = id if id else str(uuid.uuid4())
time = datetime.datetime.now()
models = ModelsId.from_external(book.book, cls.type())
return Model(MetaId(models, time.strftime("%Y-%m-%dT%H:%M:%S")),
ModelMeta(id, time, style=book.get_meta().notationStyle))
def op_to_task_runner(operation: str, book: DatabaseBook, body: dict) -> TaskRunner:
from omr.steps.algorithmtypes import AlgorithmTypes
for at in AlgorithmTypes:
if at.value == operation:
from restapi.operationworker.taskrunners.taskrunnerprediction import TaskRunnerPrediction, AlgorithmPredictorParams, Settings
r = AlgorithmRequest.from_dict(body)
meta = book.get_meta()
meta.algorithmPredictorParams[at] = r.params
return TaskRunnerPrediction(at,
PageSelection.from_params(r.selection, book),
Settings(meta.algorithm_predictor_params(at), store_to_pcgts=r.store_to_pcgts)
)
# check if operation is linked to a task
if operation == 'train_symbols':
from restapi.operationworker.taskrunners.taskrunnersymboldetectiontrainer import TaskRunnerSymbolDetectionTrainer, TaskTrainerParams
return TaskRunnerSymbolDetectionTrainer(book, TaskTrainerParams.from_dict(body.get('trainParams', {})))
elif operation == 'train_staff_line_detector':
from restapi.operationworker.taskrunners.taskrunnertrainer import TaskRunnerTrainer, TaskTrainerParams
return TaskRunnerTrainer(book, TaskTrainerParams.from_dict(body.get('trainParams', {})), AlgorithmTypes.STAFF_LINES_PC)
elif operation == 'train_character_recognition':
from restapi.operationworker.taskrunners.taskrunnertrainer import TaskRunnerTrainer, TaskTrainerParams
return TaskRunnerTrainer(book, TaskTrainerParams.from_dict(body.get('trainParams', {})), AlgorithmTypes.OCR_CALAMARI)
else:
raise NotImplementedError()
def default_model_for_book(cls, book: DatabaseBook) -> Model:
models = ModelsId.from_internal(book.get_meta().notationStyle,
cls.type())
return Model(MetaId(models, cls.model_dir()))
if __name__ == '__main__':
import matplotlib.pyplot as plt
from PIL import Image
# page = Book('demo').page('page00000001')
book = DatabaseBook('test')
page = book.pages()[0]
# page = book.page('Graduel_de_leglise_de_Nevers_032_rot') # zacken in linie
# page = book.page('Graduel_de_leglise_de_Nevers_531')
# page = book.page('Graduel_de_leglise_de_Nevers_030')
# page = book.page('Graduel_de_leglise_de_Nevers_520')
# page = book.page('Graduel_de_leglise_de_Nevers_513')
pages = [page]
settings = AlgorithmPredictorSettings(
Meta.best_model_for_book(book),
book.get_meta().algorithm_predictor_params(
BasicStaffLinePredictor.meta().type()),
#["/home/wick/Documents/Projects/ommr4all-deploy/modules/ommr4all-server/internal_storage/default_models/french14/pc_staff_lines/model"],
#["/home/wick/Documents/Projects/ommr4all-deploy/modules/ommr4all-server/models_out/all/line_detection_4/best"],
# ["/home/wick/Documents/Projects/ommr4all-deploy/modules/ommr4all-server/storage/Graduel/pc_staff_lines/model"],
# ["/home/wick/Downloads/line_detection_0/best"],
)
detector = BasicStaffLinePredictor(settings)
for prediction in detector.predict(pages):
canvas = PcGtsCanvas(prediction.line.operation.page,
PageScaleReference.NORMALIZED)
def scale(p):
return prediction.line.operation.page.page_to_image_scale(
p, ref=PageScaleReference.NORMALIZED)
f, ax = plt.subplots(1, 3)