Codex/modernize baseline

.github/workflows/pypi.yaml

@@ -27,26 +27,28 @@ jobs:
 
     - name: Install dssp
       run: |
-        sudo apt-get install libboost-all-dev
-        wget https://github.com/cmbi/dssp/archive/refs/tags/2.3.0.zip
-        mv 2.3.0.zip dssp-2.3.0.zip
-        unzip dssp-2.3.0.zip
-        cd dssp-2.3.0
-        ./autogen.sh
-        ./configure
-        make mkdssp
-        cp mkdssp ..
+        sudo apt-get update
+        sudo apt-get install -y cmake libboost-all-dev zlib1g-dev libbz2-dev
+        git clone --branch v4.5.8 --depth 1 https://github.com/PDB-REDO/dssp.git
+        cmake -S dssp -B dssp/build
+        cmake --build dssp/build --target mkdssp
+        cp dssp/build/mkdssp .
+
+    - name: Install DSSP data files
+      run: |
+        mkdir -p "$GITHUB_WORKSPACE/libcifpp-data"
+        curl -L -o "$GITHUB_WORKSPACE/libcifpp-data/components.cif" https://files.wwpdb.org/pub/pdb/data/monomers/components.cif
+        curl -L -o "$GITHUB_WORKSPACE/libcifpp-data/mmcif_pdbx.dic" https://mmcif.wwpdb.org/dictionaries/ascii/mmcif_pdbx_v50.dic
+        echo "LIBCIFPP_DATA_DIR=$GITHUB_WORKSPACE/libcifpp-data" >> "$GITHUB_ENV"
 
     - name: Install stride
       run: |
-        wget --no-check-certificate https://webclu.bio.wzw.tum.de/stride/stride.tar.gz
-        mkdir stride-src
-        mv stride.tar.gz stride-src/
+        git clone https://github.com/MDAnalysis/stride.git stride-src
         cd stride-src
-        tar zxvf stride.tar.gz
-        make
-        cp stride ..
-        cd ..
+        git checkout 867a5eb0f2479cb16615512a53ee472c54649505
+        make -C src
+        cp src/stride "$GITHUB_WORKSPACE/stride"
+        chmod +x "$GITHUB_WORKSPACE/stride"
 
     - name: Install msms
       run: | 
@@ -56,11 +58,11 @@ jobs:
     - name: Install Python dependencies
       run: |
         python3 -m pip install --upgrade pip
-        pip3 install .
+        pip3 install ".[dev]"
 
     - name: Test with pytest
       run: |
-        pytest --exitfirst --verbose --failed-first \
+        python -m pytest --exitfirst --verbose --failed-first \
         --cov=. --cov-report html
 
     - name: Install pypa/build

.github/workflows/tests.yaml

@@ -22,26 +22,28 @@ jobs:
 
     - name: Install dssp
       run: |
-        sudo apt-get install libboost-all-dev
-        wget https://github.com/cmbi/dssp/archive/refs/tags/2.3.0.zip
-        mv 2.3.0.zip dssp-2.3.0.zip
-        unzip dssp-2.3.0.zip
-        cd dssp-2.3.0
-        ./autogen.sh
-        ./configure
-        make mkdssp
-        cp mkdssp ..
+        sudo apt-get update
+        sudo apt-get install -y cmake libboost-all-dev zlib1g-dev libbz2-dev
+        git clone --branch v4.5.8 --depth 1 https://github.com/PDB-REDO/dssp.git
+        cmake -S dssp -B dssp/build
+        cmake --build dssp/build --target mkdssp
+        cp dssp/build/mkdssp .
+
+    - name: Install DSSP data files
+      run: |
+        mkdir -p "$GITHUB_WORKSPACE/libcifpp-data"
+        curl -L -o "$GITHUB_WORKSPACE/libcifpp-data/components.cif" https://files.wwpdb.org/pub/pdb/data/monomers/components.cif
+        curl -L -o "$GITHUB_WORKSPACE/libcifpp-data/mmcif_pdbx.dic" https://mmcif.wwpdb.org/dictionaries/ascii/mmcif_pdbx_v50.dic
+        echo "LIBCIFPP_DATA_DIR=$GITHUB_WORKSPACE/libcifpp-data" >> "$GITHUB_ENV"
 
     - name: Install stride
       run: |
-        wget --no-check-certificate https://webclu.bio.wzw.tum.de/stride/stride.tar.gz
-        mkdir stride-src
-        mv stride.tar.gz stride-src/
+        git clone https://github.com/MDAnalysis/stride.git stride-src
         cd stride-src
-        tar zxvf stride.tar.gz
-        make
-        cp stride ..
-        cd ..
+        git checkout 867a5eb0f2479cb16615512a53ee472c54649505
+        make -C src
+        cp src/stride "$GITHUB_WORKSPACE/stride"
+        chmod +x "$GITHUB_WORKSPACE/stride"
 
     - name: Install msms
       run: | 
@@ -51,10 +53,10 @@ jobs:
     - name: Install Python dependencies
       run: |
         python3 -m pip install --upgrade pip
-        pip3 install .
+        pip3 install ".[dev]"
 
     - name: Test with pytest
       run: |
-        pytest --exitfirst --verbose --failed-first \
+        python -m pytest --exitfirst --verbose --failed-first \
         --cov=. --cov-report html
 

sbmlcore/AminoAcidProperties.py

@@ -1,4 +1,4 @@
-import pkg_resources
+from importlib.resources import files
 
 import pandas
 #N.B. All property scales have been checked! CIL
@@ -87,7 +87,7 @@ class AminoAcidRogovChange(AminoAcidPropertyChange):
 
     def __init__(self):
 
-        filename = pkg_resources.resource_filename("sbmlcore", 'data/rogov.csv')
+        filename = files("sbmlcore").joinpath("data/rogov.csv")
         self.lookup = pandas.read_csv(filename)
 
         def split_row(row):
@@ -356,4 +356,3 @@ def __init__(self):
         self.lookup = _make_table(h_donors)
         self.lookup.rename(columns = {'value': 'h_acceptors'}, inplace=True)
         self.lookup.set_index(['amino_acid'],inplace=True)
-

sbmlcore/ResidueDepth.py

@@ -32,6 +32,7 @@ def __init__(self, pdb_file, segids=None, offsets=None):
             for chain in offsets:
                 assert chain in chain_list, "Need to specify a segid that exists in pdb!"
                 assert isinstance(offsets[chain], int), "Offsets for each segid must be an integer!"
+            segids = list(offsets)
 
         rows = {'segid': [],\
                 'resid': [],\
@@ -82,4 +83,4 @@ def split_mutation(row):
 
         other.drop(columns = ['resid'], inplace=True)
 
-        return(other)
+        return(other)

sbmlcore/StructuralDistances.py

@@ -1,9 +1,19 @@
 import pandas
 import pathlib
-import MDAnalysis
 import sbmlcore
 
 
+def _load_mdanalysis():
+    try:
+        import MDAnalysis
+    except Exception as exc:  # pragma: no cover - environment dependent
+        raise ImportError(
+            "MDAnalysis is required for StructuralDistances. Install sbmlcore with the 'md' extra."
+        ) from exc
+
+    return MDAnalysis
+
+
 class StructuralDistances(object):
     """
     Distances between a specified region (i.e. origin) and all amino acids in a protein.
@@ -41,6 +51,7 @@ def __init__(
         # check file exists
         assert pathlib.Path(pdb_file).is_file(), "File does not exist!"
 
+        MDAnalysis = _load_mdanalysis()
         u = MDAnalysis.Universe(pdb_file)
 
         assert dataset_type in [

sbmlcore/TempFactors.py

@@ -1,9 +1,19 @@
 import pandas
 import pathlib
-import MDAnalysis
 import sbmlcore
 
 
+def _load_mdanalysis():
+    try:
+        import MDAnalysis
+    except Exception as exc:  # pragma: no cover - environment dependent
+        raise ImportError(
+            "MDAnalysis is required for TempFactors. Install sbmlcore with the 'md' extra."
+        ) from exc
+
+    return MDAnalysis
+
+
 class TempFactors(object):
     """
     Distances between a specified region (i.e. origin) and all amino acids in a protein.
@@ -29,6 +39,7 @@ def __init__(self, pdb_file, offsets=None):
         # check file exists
         assert pathlib.Path(pdb_file).is_file(), "File does not exist!"
 
+        MDAnalysis = _load_mdanalysis()
         u = MDAnalysis.Universe(pdb_file)
 
         # apply any offsets to the residue numbering

sbmlcore/TrajectoryDihedrals.py

@@ -2,11 +2,21 @@
 
 import pandas
 import numpy
-import MDAnalysis
-from MDAnalysis.analysis.dihedrals import Dihedral
 import sbmlcore
 
 
+def _load_mdanalysis():
+    try:
+        import MDAnalysis
+        from MDAnalysis.analysis.dihedrals import Dihedral
+    except Exception as exc:  # pragma: no cover - environment dependent
+        raise ImportError(
+            "MDAnalysis is required for TrajectoryDihedrals. Install sbmlcore with the 'md' extra."
+        ) from exc
+
+    return MDAnalysis, Dihedral
+
+
 class TrajectoryDihedrals(object):
     """Average dihedrals between a specified region (i.e. origin) and all amino acids in a protein.
 
@@ -106,6 +116,7 @@ def __init__(
         self.dihedral = dihedral
         self.angle_type = angle_type
 
+        MDAnalysis, Dihedral = _load_mdanalysis()
         first_pass = True
 
         for trajectory in trajectory_list:
@@ -190,6 +201,7 @@ def calculate_dihedrals(self, trajectory, protein_res):
         and returns array of shape (timesteps, residues)
         """
 
+        _, Dihedral = _load_mdanalysis()
         selection_call = "res." + self.dihedral + "_selection()"
 
         # generate list of nonetype dihedral indexes (residue index)
@@ -345,6 +357,8 @@ def _filter_frames(pdb, traj, boundary, spec_time, dt):
         """Returns MDAnalysis.Universe with frames greater and
         less than the specified start and end times"""
 
+        MDAnalysis, _ = _load_mdanalysis()
+
         # Becuase a new universe is essentially being created, every coordinate in the original is needed
         coordinates = (
             MDAnalysis.analysis.base.AnalysisFromFunction(
@@ -385,6 +399,8 @@ def _filter_frames(pdb, traj, boundary, spec_time, dt):
     def _add_bonds(traj):
         """add bonds to protein (only) in a trajectory"""
 
+        MDAnalysis, _ = _load_mdanalysis()
+
         protein_res = traj.select_atoms("protein")
         # run the bond guessing algorithm
         bonds = MDAnalysis.topology.guessers.guess_bonds(

sbmlcore/TrajectoryDistances.py

@@ -2,10 +2,20 @@
 
 import pandas
 import numpy
-import MDAnalysis
 import sbmlcore
 
 
+def _load_mdanalysis():
+    try:
+        import MDAnalysis
+    except Exception as exc:  # pragma: no cover - environment dependent
+        raise ImportError(
+            "MDAnalysis is required for TrajectoryDistances. Install sbmlcore with the 'md' extra."
+        ) from exc
+
+    return MDAnalysis
+
+
 class TrajectoryDistances(object):
     """
     Average distances between a specified region (i.e. origin) and all amino acids in a protein.
@@ -92,13 +102,13 @@ def __init__(
             if start_time is not None:
                 assert start_time < end_time
 
+        MDAnalysis = _load_mdanalysis()
         first_pass = True
 
         for trajectory in trajectory_list:
 
             u = MDAnalysis.Universe(pdb_file, trajectory)
             u_static = MDAnalysis.Universe(static_pdb)
-
             reference_com = u.select_atoms(distance_selection).center_of_mass()
 
             # check atom selection exists
@@ -129,6 +139,8 @@ def __init__(
                 else:
                     distance_array = numpy.concatenate([distance_array, distances])
 
+        assert not first_pass, "No trajectory frames matched the requested time range."
+
         # inverts the array to make subseqeunt calculations more intuitive
         distance_array = distance_array.T