import argparse, pdb, sys
from cichlid_bower_tracking.data_preparers.project_preparer import ProjectPreparer as PP
parser = argparse.ArgumentParser(
usage=
'This script will perform inital analysis of depth data to create a smoothed, interpolated, 3D npy array'
)
parser.add_argument('ProjectID',
type=str,
help='Which projectID you want to identify')
parser.add_argument('Number',
type=int,
help='Total number of videos to label for this project')
parser.add_argument('Initials', type=str, help='Initials of annotator')
args = parser.parse_args()
if len(args.Initials) != 3:
print('Initials must be three letters long')
sys.exit()
pp_obj = PP(args.ProjectID)
pp_obj.manuallyLabelFrames(args.Initials.upper(), args.Number)
'MLType',
type=str,
choices=['3DResnet', 'RCNN'],
help=
'Choose between a 3D Resnet for classifying videos or an RCNN to identify fish'
)
parser.add_argument('ModelID', type=str, help='Name of the model created')
parser.add_argument('--ProjectIDs',
nargs='+',
type=str,
help='Restrict training data to the following ProjectIDs')
parser.add_argument('--GPU',
type=str,
default='0',
help='Specify the GPU card to use')
args = parser.parse_args()
subprocess.run([
'python3', '-m', 'cichlid_bower_tracking.unit_scripts.download_data',
'Train' + args.MLType, '--ModelID', args.ModelID
])
pp_obj = PP(projectID=None, modelID=args.ModelID)
pp_obj.createModel(args.MLType, args.ProjectIDs, args.GPU)
subprocess.run([
'python3', '-m', 'cichlid_bower_tracking.unit_scripts.upload_data',
'Train' + args.MLType, '--ModelID', args.ModelID, '--Delete'
])
import argparse, pdb
from cichlid_bower_tracking.data_preparers.project_preparer import ProjectPreparer as PP
parser = argparse.ArgumentParser(
usage=
'This script will perform inital analysis of depth data to create a smoothed, interpolated, 3D npy array'
)
parser.add_argument('ProjectID',
type=str,
help='Which projectID you want to identify')
parser.add_argument('--VideoIndex',
type=int,
help='Index of video that should be analyzed')
parser.add_argument('--Workers',
type=int,
help='Number of threads to run this analysis in parallel')
args = parser.parse_args()
pp_obj = PP(args.ProjectID, workers=args.Workers)
pp_obj.runClusterAnalysis(args.VideoIndex)
import argparse
from cichlid_bower_tracking.data_preparers.project_preparer import ProjectPreparer as PP
parser = argparse.ArgumentParser(
usage=
'This script is run after analysis to create useful figures and summary files'
)
parser.add_argument('ProjectID',
type=str,
help='Which projectID you want to identify')
parser.add_argument('--SummaryFile',
type=str,
help='Name of csv file that specifies projects to analyze',
default=None)
args = parser.parse_args()
pp_obj = PP(args.ProjectID, summaryFile=args.SummaryFile)
pp_obj.runSummaryCreation()
'ManualLabelVideos', 'ManualLabelFrames', 'Summary',
'All'
],
help='What type of analysis to perform')
parser.add_argument(
'--ProjectID',
type=str,
help=
'Manually identify the projects you want to analyze. If All is specified, all non-prepped projects will be analyzed'
)
parser.add_argument(
'--VideoIndex',
type=int,
help=
'Specify which video should be downloaded if Cluster analysis is to be performed'
)
parser.add_argument('--ModelID',
type=str,
help='Specify a ModelID to download')
parser.add_argument('--SummaryFile',
type=str,
help='Name of csv file that specifies projects to analyze',
default=None)
args = parser.parse_args()
#print('Downloading data for ' + args.ProjectID, file = sys.stderr)
pp_obj = PP(args.ProjectID, args.ModelID, summaryFile=args.SummaryFile)
pp_obj.downloadData(args.DataType, videoIndex=args.VideoIndex)
'Prep', 'Depth', 'Cluster', 'ClusterClassification',
'Train3DResnet', 'ManualLabelVideos', 'Summary', 'All'
],
help='What type of analysis to perform')
parser.add_argument(
'--ProjectID',
type=str,
help=
'Manually identify the projects you want to analyze. If All is specified, all non-prepped projects will be analyzed'
)
parser.add_argument(
'--VideoIndex',
type=int,
help=
'Specify which video should be downloaded if Cluster analysis is to be performed'
)
parser.add_argument('--ModelID',
type=str,
help='Specify a ModelID to download')
parser.add_argument('--Delete',
action='store_true',
help='Delete data once it is uploaded')
parser.add_argument('--NoUpload',
action='store_true',
help='Do not upload data (but still delete it)')
args = parser.parse_args()
pp_obj = PP(args.ProjectID, args.ModelID)
pp_obj.uploadData(args.DataType, args.VideoIndex, args.Delete, args.NoUpload)
import argparse, subprocess
from cichlid_bower_tracking.data_preparers.project_preparer import ProjectPreparer as PP
parser = argparse.ArgumentParser(
usage=
'This script will prompt user to identify tray location and registration info between depth and video data. Works on a single projectID'
)
parser.add_argument('ProjectID',
type=str,
help='Manually identify the project you want to analyze')
args = parser.parse_args()
pp_obj = PP(args.ProjectID)
pp_obj.runPrepAnalysis()
import argparse, pdb, sys
from cichlid_bower_tracking.data_preparers.project_preparer import ProjectPreparer as PP
parser = argparse.ArgumentParser(
usage=
'This script will use a previously trained 3D Resnet model to classify videos'
)
parser.add_argument('ProjectID',
type=str,
help='Which projectID you want to identify')
args = parser.parse_args()
print('Analyzing depth data for ' + args.ProjectID, file=sys.stderr)
pp_obj = PP(args.ProjectID)
pp_obj.runDepthAnalysis()