[gpuCI] Forward-merge master to dev-v0.4.0

Dockerfile

@@ -11,22 +11,33 @@
 # Inherit from base NVIDIA CUDA 10.1 image with CUDNN
 FROM nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04
 
+RUN apt-get update && apt-get install -y \
+        software-properties-common
+    RUN add-apt-repository ppa:deadsnakes/ppa
+    RUN apt-get update && apt-get install -y \
+        python3.7
+
+
 # Set python version to 3.6
 RUN apt update && apt install -y \
-    python3.6 \
     rsync \
     git \
     python3-pip \
     libz-dev \
     vim \
-    wget
+    wget \
+    curl \
+    python3.7-dev
+
+RUN ln -nsf /usr/bin/python3.7 /usr/bin/python
 
-RUN ln -nsf /usr/bin/python3.6 /usr/bin/python
+# Use python3.7 pip. Otherwise some packages won't be compatible.
+RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.7 1
 
 # Download 3rd party binaries
 RUN rsync -aP rsync://hgdownload.soe.ucsc.edu/genome/admin/exe/linux.x86_64/bedGraphToBigWig /usr/local/bin
 RUN rsync -aP rsync://hgdownload.soe.ucsc.edu/genome/admin/exe/linux.x86_64/bigWigToBedGraph /usr/local/bin
 
 # Download AtacWorks repo
-RUN pip3 install atacworks==0.3.0
+RUN pip3 install atacworks==0.3.4
 RUN pip3 install macs2==2.2.4

README.md

@@ -10,7 +10,7 @@ AtacWorks is a deep learning toolkit for coverage track denoising and peak calli
 
 * Ubuntu 16.04+
 * CUDA 9.0+
-* Python 3.6.7+
+* Python 3.7.0+
 * GCC 5+
 * (Optional) A conda or virtualenv setup
 * Any NVIDIA GPU that supports CUDA 9.0+

atacworks/io/h5io.py

@@ -26,7 +26,7 @@ def dict_to_h5(x, h5file, create_new=True, compression='lzf'):
             Only valid with create_new=True.
 
     """
-    with h5py.File(h5file) as f:
+    with h5py.File(h5file, 'a') as f:
         if create_new:
             # Create new datasets
             for key in x.keys():

atacworks/io/peak2bw.py

@@ -52,9 +52,9 @@ def peak2bw(input_file, sizesfile, out_dir):
     # Read input files
     _logger.info('Reading input file')
     # Skip first line if the file is narrowPeak
-    skip = False
+    skip = 0
     if input_file.endswith("narrowPeak"):
-        skip = True
+        skip = 1
     peaks = read_intervals(input_file, skip=skip)
     _logger.info('Read ' + str(len(peaks)) + ' peaks.')
     sizes = read_sizes(sizesfile)

requirements.txt

@@ -8,23 +8,22 @@
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 #
 
-configargparse~=0.15.2
+configargparse~=1.2.3
 Cython~=0.29.13
-h5py~=2.10.0
-numpy~=1.17.0
-pandas~=1.0.0
+h5py~=3.1.0
+numpy~=1.19.4
+pandas~=1.2.0
 pyBigWig~=0.3.17
-pytest~=5.4.3
-pyyaml~=5.3.0
-scikit-learn~=0.21.3
-scipy~=1.4.0
+pytest~=6.2.1
+pyyaml~=5.3.1
+scipy~=1.6.0
 termcolor~=1.1.0
-torch~=1.2.0
-torchvision~=0.3.0
+torch~=1.7.1
+torchvision~=0.8.2
 scikit-learn~=0.21.3
-configargparse~=0.15.2
-setuptools~=46.1.3
-sphinx~=2.3.1
-recommonmark~=0.6.0
-sphinx-rtd-theme~=0.4.3
-jupyterlab~=2.1.4
+setuptools~=51.1.1
+sphinx~=3.4.2
+recommonmark~=0.7.1
+sphinx-rtd-theme~=0.5.0
+jupyterlab~=3.0.0
+

setup.cfg

@@ -7,5 +7,5 @@
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 [metadata]
-version = 0.3.3
+version = 0.3.4
 

tests/expected_results/end-to-end/classification_metrics_log

@@ -1,10 +1,10 @@
-INFO:2020-07-24 10:14:14,423:AtacWorks-calculate-metrics] Loading intervals
-INFO:2020-07-24 10:14:14,425:AtacWorks-calculate-metrics] Loading labels for classification
-INFO:2020-07-24 10:14:15,139:AtacWorks-calculate-metrics] Loading data for classification
-INFO:2020-07-24 10:14:15,894:AtacWorks-calculate-metrics] Calculating per-threshold classification metrics
-INFO:2020-07-24 10:14:20,130:AtacWorks-calculate-metrics] Calculating AUC metrics
-INFO:2020-07-24 10:14:22,563:AtacWorks-calculate-metrics] Done!
+INFO:2021-06-29 04:46:39,214:AtacWorks-calculate-metrics] Loading intervals
+INFO:2021-06-29 04:46:39,216:AtacWorks-calculate-metrics] Loading labels for classification
+INFO:2021-06-29 04:46:39,902:AtacWorks-calculate-metrics] Loading data for classification
+INFO:2021-06-29 04:46:40,682:AtacWorks-calculate-metrics] Calculating per-threshold classification metrics
+INFO:2021-06-29 04:46:43,972:AtacWorks-calculate-metrics] Calculating AUC metrics
+INFO:2021-06-29 04:46:46,554:AtacWorks-calculate-metrics] Done!
 Bases per class in clean data: positive: 294990 | negative: 9689010
-Bases per class at threshold 0.5: positive: 1514137 | negative: 8469863
-Classification metrics at threshold 0.5 : recall: 0.8262 | precision: 0.1610 | specificity: 0.8689 | accuracy: 0.8676 | f1: 0.2694
-AUC metrics : auroc: 0.8475 | auprc: 0.1381
+Bases per class at threshold 0.5: positive: 1245708 | negative: 8738292
+Classification metrics at threshold 0.5 : recall: 0.9232 | precision: 0.2186 | specificity: 0.8995 | accuracy: 0.9002 | f1: 0.3535
+AUC metrics : auroc: 0.9113 | auprc: 0.2041