diff --git a/scripts/maestro/make_20230129_datasets.py b/scripts/maestro/make_20230129_datasets.py
new file mode 100644
index 0000000000000000000000000000000000000000..df2246cda82d7976a4f71aa327a3cf600c9ef6be
--- /dev/null
+++ b/scripts/maestro/make_20230129_datasets.py
@@ -0,0 +1,50 @@
+ntrials = 1
+freq = 10
+labels = ['back_large', 'cast_large', 'hunch_large', 'roll_large', 'run_large', 'stop_large', 'small_motion']
+
+
+import os
+pwd = os.path.abspath('..')
+print('JULIA_PROJECT=', os.environ['JULIA_PROJECT'])
+data_path = os.environ['DATA_REPOSITORY']
+
+import sys
+taskid = int(sys.argv[1])
+
+i = 0
+_break = False
+for window in (2,):
+ for subset in ('pretrain', 'train', 'test'):
+ for trial in range(1, ntrials + 1):
+ i += 1
+ if i == taskid:
+ _break = True
+ break
+ if _break: break
+ if _break: break
+print('window:', window, 'subset:', subset, 'trial:', trial, flush=True)
+
+from math import ceil
+from julia import TaggingBackends
+write_larva_dataset_hdf5 = TaggingBackends.LarvaDatasets.write_larva_dataset_hdf5
+
+def dataset_files(subset):
+ if subset == 'pretrain':
+ subset = 'train'
+ with open(os.path.join(pwd, f"TaggingBackends/data/t15-subset1/{subset}/filelist.txt"), 'r') as f:
+ return [os.path.join(data_path, line.rstrip()) for line in f.readlines()]
+
+larva_dataset_files = {}
+
+window_length = int(window * freq)
+third_window_length = ceil(window_length / 3)
+
+winlen = window_length if subset == 'pretrain' else third_window_length
+balance = subset == 'pretrain'
+_dir = os.path.join(pwd, f"MaggotUBA-adapter/data/20230129/{subset}/{window_length}/{trial}")
+os.makedirs(_dir, exist_ok=True)
+_file = write_larva_dataset_hdf5(_dir, dataset_files(subset), winlen, labels=labels, balance=balance, frameinterval=0.1)
+larva_dataset_files[(window, subset, trial)] = os.path.join(_dir, _file)
+
+print(larva_dataset_files)
+
diff --git a/scripts/maestro/make_20230129_datasets.sh b/scripts/maestro/make_20230129_datasets.sh
new file mode 100644
index 0000000000000000000000000000000000000000..6dd136e4548f7b5654cf4172474767951219073e
--- /dev/null
+++ b/scripts/maestro/make_20230129_datasets.sh
@@ -0,0 +1,27 @@
+#!/bin/bash
+#SBATCH --job-name=Dataset20230129
+#SBATCH --partition=dbc_pmo
+#SBATCH --qos=dbc
+#SBATCH --array=1-3
+#SBATCH --mem-per-cpu=60G
+
+host=$(hostname)
+case ${host:0:8} in
+ maestro-)
+echo "running on maestro"
+if ! [ "$(pwd)" = "$MYSCRATCH/MaggotUBA-adapter" ]; then
+echo "Please run from MaggotUBA-adapter directory:"
+echo " sbatch scripts/maestro/make_20230129_datasets.sh"
+exit 1
+fi
+export DATA_REPOSITORY=/pasteur/zeus/projets/p02/hecatonchire/screens
+;;
+ *)
+echo "unknown host: $host"
+exit 1
+esac
+
+export JULIA_NUM_THREADS=1
+
+echo "running JULIA_PROJECT=$(realpath ../TaggingBackends) poetry run python scripts/maestro/make_20230129_datasets.py $SLURM_ARRAY_TASK_ID"
+JULIA_PROJECT=$(realpath ../TaggingBackends) poetry run python scripts/maestro/make_20230129_datasets.py $SLURM_ARRAY_TASK_ID
diff --git a/scripts/maestro/pretrain_on_20230129_datasets.py b/scripts/maestro/pretrain_on_20230129_datasets.py
new file mode 100644
index 0000000000000000000000000000000000000000..39cd9da3c0d705e66fb39c2200c4069695863c71
--- /dev/null
+++ b/scripts/maestro/pretrain_on_20230129_datasets.py
@@ -0,0 +1,86 @@
+import os, sys
+from math import ceil
+import subprocess
+import tempfile
+from glob import glob
+
+taskid = int(sys.argv[1])
+
+ntrials = 1
+freq = 10
+windows = (2, )
+
+larva_dataset_files = {}
+
+todo = []
+
+for window in windows:
+ window_length = int(window * freq)
+ third_window_length = ceil(window_length / 3)
+ for subset, winlen in (('pretrain', window_length), ):#('train', third_window_length), ('test', third_window_length)):
+ for trial in range(1, ntrials+1):
+ _dir = f"data/20230129/{subset}/{window_length}/{trial}"
+ _files = glob(os.path.join(_dir, '*.hdf5'))
+ if bool(_files):
+ larva_dataset_files[(window, subset, trial)] = _files[0]
+
+print(len(larva_dataset_files))
+
+i = 0
+for window in windows:
+ window_length = int(window * freq)
+ for trial in range(1, ntrials+1):
+ i += 1
+ if i != taskid:
+ continue
+ maggotuba_project_name = f"20230129-{window_length}-{trial}"
+ print(i, maggotuba_project_name)
+ if todo and maggotuba_project_name not in todo:
+ continue
+ maggotuba_dir = os.path.abspath('../MaggotUBA-core')
+ try:
+ larva_dataset = os.path.abspath(larva_dataset_files[(window, 'pretrain', trial)])
+ except KeyError:
+ print('KeyError')
+ continue
+ pretrained_model_dir = os.path.abspath(f"../MaggotUBA-adapter/pretrained_models/20230129_winlength_{window_length}_dimlatent_${{dim}}_trial_{trial}")
+ pretrain = f"""
+#!/bin/bash
+mkdir -p 20230129
+cd 20230138
+
+rm -rf {maggotuba_project_name}
+echo "poetry run maggotuba setup 20230129 {maggotuba_project_name} --len_traj {window_length}"
+poetry run maggotuba setup 20230129 {maggotuba_project_name} --len_traj {window_length}
+cd {maggotuba_project_name} || exit 1
+
+ln -s {larva_dataset} .
+sed -e 's|"data_dir": ""|"data_dir": "{larva_dataset}"|' -e 's|"optim_iter": 1000|"optim_iter": 10000|' -i config.json
+sed -e 's|^ ]$| ],|' -e 's|\\}}| "spine_interpolation": "linear",\\n "frame_interval": 0.1,\\n "swap_head_tail": false\\n\\}}|' -i config.json
+
+for dim in 25; do
+
+sed -e "s/config.json/config-$dim.json/" -e "s/\\"dim_latent\\": 10/\\"dim_latent\\": $dim/" config.json > "config-$dim.json"
+
+rm -rf "training_log/latent-$dim"
+echo "poetry run maggotuba model train --name \\"latent-$dim\\" --config \\"config-$dim.json\\""
+poetry run maggotuba model train --name "latent-$dim" --config "config-$dim.json"
+[ -f "training_log/latent-$dim/best_validated_encoder.pt" ] || exit 1
+
+eval loc="{pretrained_model_dir}"
+mkdir -p "$loc"
+# TODO: make project_dir and exp_folder empty strings
+sed -e "s|{maggotuba_dir}/||" -e "s|20230129/20230129|20230129|" "training_log/latent-$dim/config.json" > "$loc/autoencoder_config.json"
+cp "training_log/latent-$dim/best_validated_encoder.pt" "$loc"
+
+done
+"""
+ with tempfile.NamedTemporaryFile('w+') as fp:
+ fp.write(pretrain)
+ fp.flush()
+ ret = subprocess.run(['/bin/bash', fp.name], cwd=maggotuba_dir, capture_output=True)
+ if ret.stdout:
+ print(ret.stdout.decode('utf8'), flush=True)
+ if ret.stderr:
+ print(ret.stderr.decode('utf8'))
+
diff --git a/scripts/maestro/pretrain_on_20230129_datasets.sh b/scripts/maestro/pretrain_on_20230129_datasets.sh
new file mode 100644
index 0000000000000000000000000000000000000000..2e1a410112be8d8003c5f262878da69cd10a50d2
--- /dev/null
+++ b/scripts/maestro/pretrain_on_20230129_datasets.sh
@@ -0,0 +1,11 @@
+#!/bin/bash
+#SBATCH --job-name=Pretrain20230129
+#SBATCH --partition=dbc_pmo
+#SBATCH --qos=dbc
+#SBATCH --array=1-1
+#SBATCH --mem-per-cpu=20G
+
+module load Python/3.8.3
+
+echo "running python3 scripts/maestro/pretrain_on_20230129_datasets.py $SLURM_ARRAY_TASK_ID"
+python3 scripts/maestro/pretrain_on_20230129_datasets.py $SLURM_ARRAY_TASK_ID
diff --git a/scripts/maestro/train_on_20230129_datasets.py b/scripts/maestro/train_on_20230129_datasets.py
new file mode 100644
index 0000000000000000000000000000000000000000..43589773cf7464b73f6a937f28cefbda88ccfdcb
--- /dev/null
+++ b/scripts/maestro/train_on_20230129_datasets.py
@@ -0,0 +1,86 @@
+from taggingbackends.explorer import BackendExplorer
+from maggotuba.data.make_dataset import make_dataset
+from maggotuba.models.train_model import train_model
+from maggotuba.models.predict_model import predict_model
+from glob import glob
+import subprocess
+import logging
+import sys
+import os
+taskid = int(sys.argv[1])
+
+ntrials = 1
+window_lengths = (20,)#(15, 20, 25, 30)
+latent_dims = (25,)#(25, 50, 100, 200)
+layer_counts = (1,)#(1, 2, 3)
+
+ntasks = len(window_lengths) * len(latent_dims) * len(layer_counts) * ntrials * ntrials
+print(f"task {taskid}/{ntasks}", flush=True)
+
+i = 0
+for window_length in window_lengths:
+ for training_trial in range(1, ntrials+1):
+ for latent_dim in latent_dims:
+ for pretraining_trial in range(1, ntrials+1):
+
+ for nlayers in layer_counts:
+ i += 1
+ if i < taskid:
+ continue
+ elif i > taskid:
+ print('exiting', flush=True)
+ sys.exit()
+
+ pretrained_model = f"20230129_winlength_{window_length}_dimlatent_{latent_dim}_trial_{pretraining_trial}"
+ model = f"20230129_winlength_{window_length}_trials_{pretraining_trial}-{training_trial}_dimlatent_{latent_dim}_layers_{nlayers}"
+ print(f"model: {model}", flush=True)
+
+ training_dataset = glob(f"data/20230129/train/{window_length}/{training_trial}/larva_dataset_*.hdf5")[0]
+ print(f"training dataset: {training_dataset}", flush=True)
+
+ # the `BackendExplorer` will find out we are running MaggotUBA-adapter, and set up the *models* and *data* directories
+ backend = BackendExplorer(model_instance=model)
+ backend.reset()
+
+ # a model instance is allocated 3 data directories: *raw*, *interim* and *processed*;
+ # input data always go in *raw*
+ backend.move_to_raw(training_dataset, copy=False)
+ # the following step simply makes the larva_dataset file available in *interim*, as expected by `train_model`;
+ # `labels_expected=True` informs `make_dataset` that the data are shipped with the ground truth
+ make_dataset(backend, labels_expected=True)
+
+ # this populates the *models* directory with a pretrained model, and then runs the training/fine-tuning
+ train_model(backend, nlayers, pretrained_model_instance=pretrained_model, subsets=(1,0,0))
+
+ for test_trial in range(1, ntrials+1):
+ break # temporarily
+
+ test_dataset = glob(f"data/20230129/test/{window_length}/{test_trial}/larva_dataset_*.hdf5")[0]
+ print(f"test dataset: {test_dataset}", flush=True)
+
+ # score using a fresh (trained) model instance
+ backend.reset_data()
+ backend.move_to_raw(test_dataset, copy=False)
+ # again, we inform `make_dataset` about the availability of the ground truth; this makes `predict_model` return both the predicted and expected labels, so that we can compare
+ make_dataset(backend, labels_expected=True)
+
+ predictions = predict_model(backend, subset='test', subsets=(0,0,1))
+
+ with open(f"data/20230129-results/winlength_{window_length}_trials_{pretraining_trial}-{training_trial}-{test_trial}.txt", 'w') as f:
+ for a, b in zip(*predictions):
+ f.write(f"{a},{b}\n")
+
+ # copy files into the directory for upcoming 20230129 tagger
+ dir_ = 'models/20230129'
+ for cmd in (
+ ['rm', '-rf', dir_],
+ ['cp', '-ra', f"models/{model}", dir_],
+ ['rm', '-f', os.path.join(dir_, 'best_validated_encoder.pt')],
+ ['sed', '-i', os.path.join(dir_, 'autoencoder_config.json'), '-e', f"s|\"log_dir\": \"models/{model}\"|\"log_dir\": \"\"|", '-e', f"s|\"config\": \".*\"|\"config\": \"{dir_}/autoencoder_config.json\"|", '-e', f"s|\"exp_name\": \"latent-{latent_dim}\"|\"exp_name\": \"20230129\"|"],
+ ['sed', '-i', os.path.join(dir_, 'clf_config.json'), '-e', f"s|{backend.model_dir()}|{dir_}|", '-e', "s|\"clf_path\": \".*\"|\"clf_path\": \"trained_classifier.pt\"|", '-e', "s|\"enc_path\": \".*\"|\"enc_path\": \"retrained_encoder.pt\"|", '-e', r's|"enc_path": "retrained_encoder.pt"|"enc_path": "retrained_encoder.pt",\n "post_filters": [\n "ABC->AAC"\n ]|'],
+ ):
+ ret = subprocess.run(cmd, capture_output=True)
+ for out in (ret.stdout, ret.stderr):
+ if out:
+ print(out.decode('utf8'))
+
diff --git a/scripts/maestro/train_on_20230129_datasets.sh b/scripts/maestro/train_on_20230129_datasets.sh
new file mode 100644
index 0000000000000000000000000000000000000000..5132435f0409deacfa83fb69a9e52bb70f4bda23
--- /dev/null
+++ b/scripts/maestro/train_on_20230129_datasets.sh
@@ -0,0 +1,28 @@
+#!/bin/bash
+#SBATCH --job-name=Train20230129
+#SBATCH --partition=dbc_pmo
+#SBATCH --qos=dbc
+#SBATCH --array=1-1
+#SBATCH --cpus-per-task=4
+
+host=$(hostname)
+case ${host:0:8} in
+ maestro-)
+# the shebangs at the top of the present file should include:
+#SBATCH --partition=dbc_pmo
+#SBATCH --qos=dbc
+echo "running on maestro"
+if ! [ "$(pwd)" = "$MYSCRATCH/MaggotUBA-adapter" ]; then
+echo "Please run from MaggotUBA-adapter directory:"
+echo " sbatch scripts/maestro/train_on_20230129_datasets.sh"
+exit 1
+fi
+export DATA_REPOSITORY=/pasteur/zeus/projets/p02/hecatonchire/screens
+;;
+ *)
+echo "unknown host: $host"
+exit 1
+esac
+
+echo "running JULIA_PROJECT=$(realpath ../TaggingBackends) poetry run python scripts/maestro/train_on_20230129_datasets.py $SLURM_ARRAY_TASK_ID"
+JULIA_PROJECT=$(realpath ../TaggingBackends) poetry run python scripts/maestro/train_on_20230129_datasets.py $SLURM_ARRAY_TASK_ID