Merge pull request #402 from ExcitedStates/cvxpy-reduce-main

Cvxpy reduce main

README.md

@@ -14,59 +14,29 @@ If you use this software, please cite:
 - [Keedy, D. A., Fraser, J. S. & van den Bedem, H. Exposing Hidden Alternative Backbone Conformations in X-ray Crystallography Using qFit. PLoS Comput. Biol. 11, e1004507 (2015)](https://dx.doi.org/10.1371/journal.pcbi.1004507)
 
 
-## Installation (conda recommended)
-
-We recommend using the _conda_ package manager to install _qFit_.
+## Installation
 
 You will need the following tools:
 
 * git
-* _conda_ package manager (which you can get by installing [Miniconda3](https://docs.conda.io/en/latest/miniconda.html))
+* pip
 
 Once these are installed, you can:
 
-1. Create a new conda env & activate it
-   ```bash
-   conda create --name qfit "python>=3.9"
-   conda activate qfit
-   ```
-
 1. Install dependencies
    ```bash
-   conda install -c anaconda mkl numpy=1.22
-   conda install -c anaconda -c ibmdecisionoptimization \
-                 cvxopt cplex
+   pip install -r requirements.txt
    ```
-   For some of the post analysis scripts, you will also need sklean
-   conda install -c anaconda scikit-learn
 
-1. Clone the latest release of the qFit source, and install to your conda env
+1. Clone the latest release of the qFit source and install it using pip
    ```bash
-   git clone -b main https://github.com/ExcitedStates/qfit-3.0.git
+   git clone https://github.com/ExcitedStates/qfit-3.0.git
    cd qfit-3.0
    pip install .
    ```
 
 1. You're now ready to run qFit programs! See [usage examples](#sec:usage-examples) below for some examples.
 
-### M1 Macs
-
-Unfortunately, the Anaconda repos don't contain 'osx-arm64' binaries for IBM's CPLEX and Intel's mkl.  
-We don't currently have plans to switch to a different MIQP solver (e.g. Gurobi).
-
-As a workaround, you'll have to force conda to install the 'osx-64' binaries for everything (x86_64).
-macOS's Rosetta 2 translation will handle the Intel→AppleSilicon translation.
-
-Instead of the first step in the above Installation section, use this:
-
-1. Create a new conda env & activate it
-   ```bash
-   CONDA_SUBDIR=osx-64 conda create --name qfit "python>=3.9"
-   conda activate qfit; conda env config vars set CONDA_SUBDIR=osx-64; conda deactivate
-   conda activate qfit
-   ```
-
-then follow the rest of the instructions.
 
 ### Advanced
 
@@ -75,15 +45,10 @@ If you prefer to manage your environments using other methods, qFit has the foll
 * [Python 3.6+](https://python.org)
 * [numpy](https://numpy.org)
 * [scipy](https://scipy.org)
-* [cvxopt](https://cvxopt.org)
-* [IBM ILOG CPLEX Optimization Studio (Community Edition)](https://www.ibm.com/products/ilog-cplex-optimization-studio)
-
-Installation instructions using `pip` can be found in the `docs` folder.
+* [cvxpy](https://www.cvxpy.org)
 
 Once dependencies are installed, you can clone the qFit source, and install to your env as above.
 
-(Note: `python setup.py install` will only work if numpy has _already_ been installed.)
-
 
 ## Contributing
 
@@ -167,5 +132,3 @@ Gohlke. See file header.
 
 The `Xpleo` software and `LoopTK` package have been major inspirations for the inverse kinematics
 functionality.
-
-[1]: https://www-01.ibm.com/software/websphere/products/optimization/cplex-studio-community-edition/ "IBM website"

docs/aws_deploy.sh

@@ -2,31 +2,11 @@
 
 # Tested on Amazon Linux 2, but should work on most RPM-based Linux distros
 
-# install Anaconda RPM GPG keys
-sudo rpm --import https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc
-
-# add Anaconda repository
-cat <<EOF | sudo tee /etc/yum.repos.d/conda.repo
-[conda]
-name=Conda
-baseurl=https://repo.anaconda.com/pkgs/misc/rpmrepo/conda
-enabled=1
-gpgcheck=1
-gpgkey=https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc
-EOF
-
-sudo yum -y install conda
 sudo yum -y install git gcc
 
-source /opt/conda/etc/profile.d/conda.sh
-conda create -y --name qfit
-conda activate qfit
-
-conda install -y -c anaconda mkl
-conda install -y -c anaconda -c ibmdecisionoptimization cvxopt cplex
-
 git clone https://github.com/ExcitedStates/qfit-3.0.git
 cd qfit-3.0/
+pip install -r requirements.txt
 
 # Optionally, uncomment the following line to set a specific version of qFit
 #git checkout v3.1.2

environment.yml

@@ -7,11 +7,9 @@ dependencies:
   - libblas[build=*mkl]
   - numpy>=1.20,<2
   - scipy>=1.0
+  - cvxpy
+  - pyscipopt
   - pandas>=1.2
   - pyparsing>=2.2.0
   - cvxopt
-  - ibmdecisionoptimization::cplex
-  - osqp
-  - pip:
-    - "git+https://github.com/osqp/miosqp.git@ac672338b0593d865dd15b7a76434f25e24244a9#egg=miosqp"
   - tqdm

pyproject.toml

@@ -5,4 +5,6 @@ requires = [
 	"wheel",       # PEP 508 specifications.
 	"setuptools_scm",
 	"numpy>=1.20,<2",
+	"cvxpy",
+	"pyscipopt",
 ]

requirements.txt

@@ -0,0 +1,5 @@
+numpy
+cvxpy
+pyscipopt
+scipy
+scikit-learn

scripts/post/calc_rscc.py

@@ -8,64 +8,68 @@
 from qfit.volume import XMap
 from qfit.validator import Validator
 
-'''
+"""
 This script will calculate the RSCC of a ligand (or any residue) defined by their ligand name (--ligand) or residue number and chain id (--resi_chain). 
 It will only work on mtz maps with 2FOFCWT,PH2FOFCWT. 
 
 To run: 
  calc_rscc.py PDB_FILE.pdb MTZ_FILE.mtz --ligand AR6 --pdb PDB_NAME --directory /path/for/output/csv/file
-'''
+"""
+
 
 def build_argparser():
     p = ArgumentParser(description=__doc__)
     p.add_argument("structure", type=str, help="PDB-file containing structure.")
     p.add_argument("map", type=str, help="Map.")
-    p.add_argument("--ligand", type=str, help="name of ligand for RSCC to be calculated on")
-    p.add_argument("--residue", type=str, help="Chain_ID, Residue_ID for RSCC to be calculated on")
+    p.add_argument(
+        "--ligand", type=str, help="name of ligand for RSCC to be calculated on"
+    )
+    p.add_argument(
+        "--residue", type=str, help="Chain_ID, Residue_ID for RSCC to be calculated on"
+    )
     p.add_argument("--pdb", type=str, help="name of PDB")
     p.add_argument("--directory", type=str, help="Where to save RSCC info")
     return p
 
 
 def main():
-  p = build_argparser()
-  options = p.parse_args()
-  # Load structure and prepare it
-  structure = Structure.fromfile(options.structure)
-  if options.ligand is not None:
-     ligand = structure.extract("resn", options.ligand, "==")
-  elif options.residue is not None:
-    chainid, resi = options.residue.split(",")
-    ligand = structure.extract(f"resi {resi} and chain {chainid}")
-  else:
-      print('Please provide ligand name or residue ID and chain ID')
+    p = build_argparser()
+    options = p.parse_args()
+    # Load structure and prepare it
+    structure = Structure.fromfile(options.structure)
+    if options.ligand is not None:
+        ligand = structure.extract("resn", options.ligand, "==")
+    elif options.residue is not None:
+        chainid, resi = options.residue.split(",")
+        ligand = structure.extract(f"resi {resi} and chain {chainid}")
+    else:
+        print("Please provide ligand name or residue ID and chain ID")
+
+    # Load and process the electron density map:
+    xmap = XMap.fromfile(options.map, label="2FOFCWT,PH2FOFCWT")
+    scaler = MapScaler(xmap)
+    xmap = xmap.canonical_unit_cell()
+    footprint = ligand
+    scaler.scale(footprint, radius=1.5)
 
-  # Load and process the electron density map:
-  xmap = XMap.fromfile(
-    options.map, label='2FOFCWT,PH2FOFCWT'
-  )
-  scaler = MapScaler(xmap)
-  xmap = xmap.canonical_unit_cell()
-  footprint = ligand
-  scaler.scale(footprint, radius=1.5)
+    xmap = xmap.extract(ligand.coor, padding=8)
 
-  xmap = xmap.extract(ligand.coor, padding=8)
+    # Now that the conformers have been generated, the resulting
+    # # conformations should be examined via GoodnessOfFit:
+    validator = Validator(xmap, xmap.resolution, options.directory)
+    rscc = validator.rscc(ligand)
+    print(rscc)
 
-  # Now that the conformers have been generated, the resulting
-  # # conformations should be examined via GoodnessOfFit:
-  validator = Validator(xmap, xmap.resolution, options.directory)
-  rscc = validator.rscc(ligand)
-  print(rscc)
+    csv_filename = f"{options.pdb}_rscc.csv"
 
-  csv_filename = f"{options.pdb}_rscc.csv"
+    # Write to CSV
+    with open(csv_filename, "w", newline="") as csvfile:
+        writer = csv.writer(csvfile)
+        # Write the header
+        writer.writerow(["PDB", "RSCC"])
+        # Write the data
+        writer.writerow([options.pdb, rscc])
 
-  # Write to CSV
-  with open(csv_filename, 'w', newline='') as csvfile:
-    writer = csv.writer(csvfile)
-    # Write the header
-    writer.writerow(["PDB", "RSCC"])
-    # Write the data
-    writer.writerow([options.pdb, rscc])
 
 if __name__ == "__main__":
     main()

scripts/post/find_close_residues.py

@@ -22,7 +22,6 @@
 from qfit.structure.rotamers import ROTAMERS
 
 
-
 def parse_args():
     p = argparse.ArgumentParser(description=__doc__)
     p.add_argument("structure", type=str, help="PDB-file containing structure.")
@@ -69,11 +68,11 @@ def main():
         close_res.loc[n, "res_id"] = residue_id
         close_res.loc[n, "chain"] = chain
         n += 1
-    close_res.to_csv(pdb_name + "_" + args.ligand + "_" +str(args.dist) + "_closeres.csv",
-                index=False,
-            )
+    close_res.to_csv(
+        pdb_name + "_" + args.ligand + "_" + str(args.dist) + "_closeres.csv",
+        index=False,
+    )
 
 
 if __name__ == "__main__":
     main()
-

scripts/post/lig_occ.py

@@ -42,7 +42,8 @@ def get_occ(structure, ligand, pdb):
             )
         )
     occ = pd.DataFrame(
-        lig, columns=["PDB", "ligand_name", "chain", "min_occ", "average_b", "num_altloc"]
+        lig,
+        columns=["PDB", "ligand_name", "chain", "min_occ", "average_b", "num_altloc"],
     )
     occ.to_csv(pdb + "_ligand_occupancy.csv", index=False)
 

setup.py

@@ -27,6 +27,8 @@ def main():
     install_requires = [
         "numpy>=1.20,<2",
         "scipy>=1.0",
+        "cvxpy",
+        "pyscipopt",
         "pandas>=1.2",
         "pyparsing>=2.2.0",
         "tqdm>=4.0.0",
@@ -44,10 +46,6 @@ def main():
         ext_modules=ext_modules,
         setup_requires=setup_requires,
         install_requires=install_requires,
-        extra_dependencies={
-            "cplex": ["cvxopt", "cplex"],
-            "osqp": ["osqp", "miosqp @ git+https://github.com/osqp/miosqp.git@ac672338b0593d865dd15b7a76434f25e24244a9#egg=miosqp"],
-        },
         zip_safe=False,
         python_requires=">=3.9",
         entry_points={