Chore/restructure (#102)

* restructure project
CenterForMedicalGeneticsGhent · Apr 19, 2024 · b16b310 · b16b310
1 parent ea6342c
commit b16b310
Show file tree

Hide file tree

Showing 50 changed files with 108 additions and 69 deletions.
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -0,0 +1,9 @@
+repos:
+  - repo: https://github.com/psf/black
+    rev: 21.9b0
+    hooks:
+      - id: black
+  - repo: https://gitlab.com/pycqa/flake8
+    rev: 3.9.2
+    hooks:
+      - id: flake8
diff --git a/MANIFEST.in b/MANIFEST.in
@@ -1 +1 @@
-recursive-include wisecondorX *.R
+recursive-include wisecondorx *.R
diff --git a/README.md b/README.md
@@ -73,7 +73,7 @@ WisecondorX convert input.bam/cram output.npz [--optional arguments]
 | `--binsize x`                  | Size per bin in bp; the reference bin size should be a multiple of this value. Note that this parameter does not impact the resolution, yet it can be used to optimize processing speed (default: x=5e3) |
 | `--normdup`                    | Use this flag to avoid duplicate removal                                                                                                                                                                 |
 
-&rarr; Bash recipe at `./pipeline/convert.sh`
+&rarr; Bash recipe at `docs/include/pipeline/convert.sh`
 
 ### Stage (2) Create reference
 
@@ -91,7 +91,7 @@ WisecondorX newref reference_input_dir/*.npz reference_output.npz [--optional ar
 | `--plotyfrac x`                | plots Y read fraction histogram and Gaussian mixture fit to file x, can help when setting `--yfrac` manually; software quits after plotting |
 | `--cpus x`                     | Number of threads requested (default: x=1)                                                                                                  |
 
-&rarr; Bash recipe at `./pipeline/newref.sh`
+&rarr; Bash recipe at `docs/include/pipeline/newref.sh`
 
 ### Stage (3) Predict copy number alterations
 
@@ -116,7 +116,7 @@ WisecondorX predict test_input.npz reference_input.npz output_id [--optional arg
 
 <sup>**(\*)** At least one of these output formats should be selected</sup>
 
-&rarr; Bash recipe at `./pipeline/predict.sh`
+&rarr; Bash recipe at `docs/include/pipeline/predict.sh`
 
 ### Additional functionality
 

diff --git a/example.bed/ID_aberrations.bed → docs/include/example.bed/ID_aberrations.bed b/example.bed/ID_aberrations.bed → docs/include/example.bed/ID_aberrations.bed
diff --git a/example.bed/ID_bins.bed → docs/include/example.bed/ID_bins.bed b/example.bed/ID_bins.bed → docs/include/example.bed/ID_bins.bed
diff --git a/example.bed/ID_chr_statistics.txt → ...include/example.bed/ID_chr_statistics.txt b/example.bed/ID_chr_statistics.txt → ...include/example.bed/ID_chr_statistics.txt
diff --git a/example.bed/ID_segments.bed → docs/include/example.bed/ID_segments.bed b/example.bed/ID_segments.bed → docs/include/example.bed/ID_segments.bed
diff --git a/example.blacklist/acen.and.gvar.hg38.txt → .../example.blacklist/acen.and.gvar.hg38.txt b/example.blacklist/acen.and.gvar.hg38.txt → .../example.blacklist/acen.and.gvar.hg38.txt
diff --git a/example.blacklist/acen.hg38.txt → docs/include/example.blacklist/acen.hg38.txt b/example.blacklist/acen.hg38.txt → docs/include/example.blacklist/acen.hg38.txt
diff --git a/example.plot/chr1.png → docs/include/example.plot/chr1.png b/example.plot/chr1.png → docs/include/example.plot/chr1.png
diff --git a/example.plot/chr10.png → docs/include/example.plot/chr10.png b/example.plot/chr10.png → docs/include/example.plot/chr10.png
diff --git a/example.plot/chr11.png → docs/include/example.plot/chr11.png b/example.plot/chr11.png → docs/include/example.plot/chr11.png
diff --git a/example.plot/chr12.png → docs/include/example.plot/chr12.png b/example.plot/chr12.png → docs/include/example.plot/chr12.png
diff --git a/example.plot/chr13.png → docs/include/example.plot/chr13.png b/example.plot/chr13.png → docs/include/example.plot/chr13.png
diff --git a/example.plot/chr14.png → docs/include/example.plot/chr14.png b/example.plot/chr14.png → docs/include/example.plot/chr14.png
diff --git a/example.plot/chr15.png → docs/include/example.plot/chr15.png b/example.plot/chr15.png → docs/include/example.plot/chr15.png
diff --git a/example.plot/chr16.png → docs/include/example.plot/chr16.png b/example.plot/chr16.png → docs/include/example.plot/chr16.png
diff --git a/example.plot/chr17.png → docs/include/example.plot/chr17.png b/example.plot/chr17.png → docs/include/example.plot/chr17.png
diff --git a/example.plot/chr18.png → docs/include/example.plot/chr18.png b/example.plot/chr18.png → docs/include/example.plot/chr18.png
diff --git a/example.plot/chr19.png → docs/include/example.plot/chr19.png b/example.plot/chr19.png → docs/include/example.plot/chr19.png
diff --git a/example.plot/chr2.png → docs/include/example.plot/chr2.png b/example.plot/chr2.png → docs/include/example.plot/chr2.png
diff --git a/example.plot/chr20.png → docs/include/example.plot/chr20.png b/example.plot/chr20.png → docs/include/example.plot/chr20.png
diff --git a/example.plot/chr21.png → docs/include/example.plot/chr21.png b/example.plot/chr21.png → docs/include/example.plot/chr21.png
diff --git a/example.plot/chr22.png → docs/include/example.plot/chr22.png b/example.plot/chr22.png → docs/include/example.plot/chr22.png
diff --git a/example.plot/chr3.png → docs/include/example.plot/chr3.png b/example.plot/chr3.png → docs/include/example.plot/chr3.png
diff --git a/example.plot/chr4.png → docs/include/example.plot/chr4.png b/example.plot/chr4.png → docs/include/example.plot/chr4.png
diff --git a/example.plot/chr5.png → docs/include/example.plot/chr5.png b/example.plot/chr5.png → docs/include/example.plot/chr5.png
diff --git a/example.plot/chr6.png → docs/include/example.plot/chr6.png b/example.plot/chr6.png → docs/include/example.plot/chr6.png
diff --git a/example.plot/chr7.png → docs/include/example.plot/chr7.png b/example.plot/chr7.png → docs/include/example.plot/chr7.png
diff --git a/example.plot/chr8.png → docs/include/example.plot/chr8.png b/example.plot/chr8.png → docs/include/example.plot/chr8.png
diff --git a/example.plot/chr9.png → docs/include/example.plot/chr9.png b/example.plot/chr9.png → docs/include/example.plot/chr9.png
diff --git a/example.plot/chrX.png → docs/include/example.plot/chrX.png b/example.plot/chrX.png → docs/include/example.plot/chrX.png
diff --git a/example.plot/genome_wide.png → docs/include/example.plot/genome_wide.png b/example.plot/genome_wide.png → docs/include/example.plot/genome_wide.png
diff --git a/pipeline/convert.sh → docs/include/pipeline/convert.sh b/pipeline/convert.sh → docs/include/pipeline/convert.sh
diff --git a/pipeline/newref.sh → docs/include/pipeline/newref.sh b/pipeline/newref.sh → docs/include/pipeline/newref.sh
diff --git a/pipeline/predict.sh → docs/include/pipeline/predict.sh b/pipeline/predict.sh → docs/include/pipeline/predict.sh
diff --git a/pyproject.toml b/pyproject.toml
@@ -0,0 +1,17 @@
+[tool.black]
+line-length = 79
+target_version = ['py36','py37','py38','py39']
+include = '\.pyi?$'
+exclude = '''
+/(
+    \.git
+  | \.hg
+  | \.mypy_cache
+  | \.tox
+  | \.venv
+  | _build
+  | buck-out
+  | build
+  | dist
+)/
+'''
diff --git a/setup.cfg b/setup.cfg
@@ -0,0 +1,63 @@
+[metadata]
+name = wisecondorx
+version = 1.2.6dev
+description = description="WisecondorX -- an evolved WISECONDOR"
+long_description = file: README.md
+author = "Matthias De Smet, Lennart Raman",
+author_email = "ict@cmgg.be",
+license = "by-nc-sa"
+license_file = LICENSE.md
+platforms = unix, linux, osx, win32
+keywords=
+        bioinformatics
+        biology
+        sequencing
+        NGS
+        next generation sequencing
+        CNV
+        SWGS
+        Shallow Whole Genome Sequencing
+classifiers =
+    Environment :: Console
+    Intended Audience :: Science/Research
+    Natural Language :: English
+    Operating System :: MacOS :: MacOS X
+    Operating System :: POSIX
+    Operating System :: Unix
+    Programming Language :: Python :: 3 :: Only
+    Programming Language :: Python :: 3
+    Programming Language :: Python :: 3.6
+    Topic :: Scientific/Engineering
+    Topic :: Scientific/Engineering :: Bio-Informatics
+
+
+[options]
+packages = wisecondorx
+package_dir =
+    =src
+include_package_data=True
+python_requires= >=3.6
+install_requires =
+    scipy
+    scikit-learn
+    pysam
+    numpy
+zip_safe = False
+
+[options.entry_points]
+    console_scripts =
+        wisecondorx = wisecondorx.main:main
+        WisecondorX = wisecondorx.main:main
+        wisecondorX = wisecondorx.main:main
+
+
+[options.extras_require]
+testing =
+    flake8
+    black
+    pre-commit
+
+[flake8]
+ignore = E266, E203, E501, C901, W503
+max-line-length = 79
+max-complexity = 18
diff --git a/setup.py b/setup.py
@@ -1,55 +1,5 @@
-#! /usr/bin/env python
-from setuptools import setup, find_packages
+#!/usr/bin/env python
+from setuptools import setup
 
-version = "1.2.4"
-dl_version = "master" if "dev" in version else "{}".format(version)
-
-setup(
-    name="WisecondorX",
-    version=version,
-    author="Matthias De Smet, Lennart Raman",
-    author_email="Lennart.raman@ugent.be",
-    description="WisecondorX -- an evolved WISECONDOR",
-    long_description=__doc__,
-    keywords=[
-        "bioinformatics",
-        "biology",
-        "sequencing",
-        "NGS",
-        "next generation sequencing",
-        "CNV",
-        "SWGS",
-        "Shallow Whole Genome Sequencing",
-    ],
-    download_url="https://github.com/CenterForMedicalGeneticsGhent/WisecondorX/archive/v{}.tar.gz".format(
-        dl_version
-    ),
-    license="Attribution-NonCommercial-ShareAlike CC BY-NC-SA",
-    packages=find_packages("."),
-    python_requires=">=2.7",
-    include_package_data=True,
-    zip_safe=False,
-    install_requires=[
-        'futures;python_version<"3"',
-        "scipy",
-        "scikit-learn",
-        "pysam",
-        "numpy",
-    ],
-    entry_points={"console_scripts": ["WisecondorX = wisecondorX.main:main"]},
-    classifiers=[
-        "Development Status :: 3 - Alpha",
-        "Environment :: Console",
-        "Intended Audience :: Science/Research",
-        "Natural Language :: English",
-        "Operating System :: MacOS :: MacOS X",
-        "Operating System :: POSIX",
-        "Operating System :: Unix",
-        "Programming Language :: Python :: 2",
-        "Programming Language :: Python :: 2.7",
-        "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.6",
-        "Topic :: Scientific/Engineering",
-        "Topic :: Scientific/Engineering :: Bio-Informatics",
-    ],
-)
+if __name__ == "__main__":
+    setup()
diff --git a/wisecondorX/__init__.py → src/wisecondorx/__init__.py b/wisecondorX/__init__.py → src/wisecondorx/__init__.py
diff --git a/wisecondorX/convert_tools.py → src/wisecondorx/convert_tools.py b/wisecondorX/convert_tools.py → src/wisecondorx/convert_tools.py
diff --git a/wisecondorX/include/CBS.R → src/wisecondorx/include/CBS.R b/wisecondorX/include/CBS.R → src/wisecondorx/include/CBS.R
diff --git a/wisecondorX/include/plotter.R → src/wisecondorx/include/plotter.R b/wisecondorX/include/plotter.R → src/wisecondorx/include/plotter.R
diff --git a/wisecondorX/main.py → src/wisecondorx/main.py b/wisecondorX/main.py → src/wisecondorx/main.py
@@ -8,17 +8,17 @@
 
 import numpy as np
 
-from wisecondorX.convert_tools import convert_reads
-from wisecondorX.newref_control import (
+from wisecondorx.convert_tools import convert_reads
+from wisecondorx.newref_control import (
     tool_newref_prep,
     tool_newref_main,
     tool_newref_merge,
 )
-from wisecondorX.newref_tools import train_gender_model, get_mask
-from wisecondorX.overall_tools import gender_correct, scale_sample
-from wisecondorX.predict_control import normalize, get_post_processed_result
-from wisecondorX.predict_output import generate_output_tables, exec_write_plots
-from wisecondorX.predict_tools import (
+from wisecondorx.newref_tools import train_gender_model, get_mask
+from wisecondorx.overall_tools import gender_correct, scale_sample
+from wisecondorx.predict_control import normalize, get_post_processed_result
+from wisecondorx.predict_output import generate_output_tables, exec_write_plots
+from wisecondorx.predict_tools import (
     log_trans,
     exec_cbs,
     apply_blacklist,

diff --git a/wisecondorX/newref_control.py → src/wisecondorx/newref_control.py b/wisecondorX/newref_control.py → src/wisecondorx/newref_control.py
@@ -9,7 +9,7 @@
 import numpy as np
 from concurrent import futures
 
-from wisecondorX.newref_tools import normalize_and_mask, train_pca, get_reference
+from wisecondorx.newref_tools import normalize_and_mask, train_pca, get_reference
 
 """
 Outputs preparation files of read depth normalized

diff --git a/wisecondorX/newref_tools.py → src/wisecondorx/newref_tools.py b/wisecondorX/newref_tools.py → src/wisecondorx/newref_tools.py
diff --git a/wisecondorX/overall_tools.py → src/wisecondorx/overall_tools.py b/wisecondorX/overall_tools.py → src/wisecondorx/overall_tools.py
diff --git a/wisecondorX/predict_control.py → src/wisecondorx/predict_control.py b/wisecondorX/predict_control.py → src/wisecondorx/predict_control.py
@@ -1,6 +1,6 @@
 # WisecondorX
 
-from wisecondorX.predict_tools import (
+from wisecondorx.predict_tools import (
     coverage_normalize_and_mask,
     project_pc,
     get_optimal_cutoff,

diff --git a/wisecondorX/predict_output.py → src/wisecondorx/predict_output.py b/wisecondorX/predict_output.py → src/wisecondorx/predict_output.py
@@ -4,7 +4,7 @@
 
 import numpy as np
 
-from wisecondorX.overall_tools import (
+from wisecondorx.overall_tools import (
     exec_R,
     get_z_score,
     get_median_segment_variance,

diff --git a/wisecondorX/predict_tools.py → src/wisecondorx/predict_tools.py b/wisecondorX/predict_tools.py → src/wisecondorx/predict_tools.py
@@ -6,7 +6,7 @@
 from scipy.stats import norm
 from sklearn.decomposition import PCA
 
-from wisecondorX.overall_tools import exec_R, get_z_score
+from wisecondorx.overall_tools import exec_R, get_z_score
 
 """
 Returns gender based on Gaussian mixture