Remove local.json in favor of system variables (#19)

* remove local.json

* refactor to use third party binaries as system variables

* add THIRDPARTY.md

* update dockerfile

* update tests

Dockerfile

@@ -1,56 +1,65 @@
 #==============================================================================================
-FROM python:3.11 as base
+FROM python:3.11 AS base
 
 LABEL author="Rodrigo V. Honorato <r.vargashonorato@uu.nl>"
 
+ARG SOFTWARE_PATH=/opt/software
+
+#------------------------------------------------------------------------------------------
+# System dependencies
 RUN apt-get update && \
   apt-get install -y --no-install-recommends \
   build-essential \
   libboost-all-dev \
   && \
   apt-get clean && rm -rf /var/lib/apt/lists/*
 
-# Copy Whiscy
-WORKDIR /opt/software/whiscy
-COPY . .
-
-# install BioPython
-RUN pip install biopython==1.79
-
-# Build protdist
-WORKDIR /opt/software/whiscy/bin/protdist
-RUN sh compile.sh
-
+#------------------------------------------------------------------------------------------
 # Build Muscle
-WORKDIR /opt/software/whiscy/muscle3.8.1551
+WORKDIR ${SOFTWARE_PATH}/muscle3.8.1551
 RUN curl https://drive5.com/muscle/muscle_src_3.8.1551.tar.gz | tar xzv && \
-  make && \
-  mv muscle /opt/software/whiscy/bin/muscle3.8.1551 && \
-  sed -i "s/\/Users\/bjimenez\/bin\/muscle\/muscle3.8.31_i86darwin64/\/opt\/software\/whiscy\/bin\/muscle3.8.1551/g" /opt/software/whiscy/etc/local.json
+  make
+ENV MUSCLE_BIN=${SOFTWARE_PATH}/muscle3.8.1551/muscle
 
+#------------------------------------------------------------------------------------------
 # Build hsspconv
+WORKDIR ${SOFTWARE_PATH}
 RUN wget https://github.com/cmbi/hssp/archive/3.1.5.tar.gz && \
   tar -zxvf 3.1.5.tar.gz && \
   cd hssp-3.1.5 && \
   ./autogen.sh && \
   ./configure && \
-  make hsspconv && \
-  mv hsspconv ../ && \
-  sed -i "s/\/Users\/bjimenez\/bin\/hssp\/hsspconv/\/opt\/software\/whiscy\/bin\/hsspconv/g" /opt/software/whiscy/etc/local.json
-
+  make hsspconv
+ENV HSSPCONV_BIN=${SOFTWARE_PATH}/hssp-3.1.5/hsspconv
 
+#------------------------------------------------------------------------------------------
 # Build freesasa
+WORKDIR ${SOFTWARE_PATH}
 RUN  wget https://github.com/mittinatten/freesasa/releases/download/2.0.3/freesasa-2.0.3.tar.gz  && \
   tar -zxvf freesasa-2.0.3.tar.gz && \
   cd freesasa-2.0.3 && \
-  ./configure --disable-json --prefix=/opt/software/whiscy/bin/freesasa && \
-  make && make install
-
-# WHISCY exports
-ENV WHISCY_PATH=/opt/software/whiscy
-ENV PYTHONPATH="${PYTHONPATH}:${WHISCY_PATH}"
-ENV WHISCY_BIN="${WHISCY_PATH}/whiscy.py"
-ENV PATH="${WHISCY_PATH}:${WHISCY_PATH}/bin/freesasa/bin:${PATH}"
+  ./configure --disable-json --prefix=`pwd` && \
+  make && \
+  make install
+ENV FREESASA_BIN=${SOFTWARE_PATH}/freesasa-2.0.3/bin/freesasa
+
+#------------------------------------------------------------------------------------------
+# Install Whiscy
+WORKDIR ${SOFTWARE_PATH}/whiscy
+COPY . .
+
+#------------------------------------------------------------------------------------------
+# install BioPython
+RUN pip install biopython==1.79
+
+#------------------------------------------------------------------------------------------
+# Build protdist
+WORKDIR ${SOFTWARE_PATH}/whiscy/bin/protdist
+RUN sh compile.sh
+ENV PROTDIST_BIN=${SOFTWARE_PATH}/whiscy/bin/protdist/protdist
+
+#------------------------------------------------------------------------------------------
+# Set data directory
 
 WORKDIR /data
 

THIRDPARTY.md

@@ -0,0 +1,71 @@
+# Installation of Third-Party dependencies
+
+These instructions assuming you are installing locally in a Ubuntu Linux system, different steps may be required for other systems.
+
+All will be installed in a `software` directory in your `$HOME` directory.
+
+## System dependendencies
+
+```bash
+$ sudo apt-get update && \
+  sudo apt-get install -y build-essential libboost-all-dev
+```
+
+## Muscle
+
+```bash
+$ mkdir -p $HOME/software && cd $HOME/software
+$ mkdir muscle3.8.1551 && cd muscle3.8.1551
+$ wget https://drive5.com/muscle/muscle_src_3.8.1551.tar.gz
+$ tar -zxf muscle_src_3.8.1551.tar.gz && rm muscle_src_3.8.1551.tar.gz
+$ make
+$ export MUSCLE_BIN=$HOME/software/muscle3.8.1551/muscle
+```
+
+## Freesasa
+```bash
+$ mkdir -p $HOME/software && cd $HOME/software
+$ wget https://github.com/mittinatten/freesasa/releases/download/2.0.3/freesasa-2.0.3.tar.gz
+$ tar -zxf freesasa-2.0.3.tar.gz && rm freesasa-2.0.3.tar.gz
+$ cd freesasa-2.0.3
+$ ./configure --disable-json --prefix=`pwd`
+$ make
+$ make install
+$ export FREESASA_BIN=$HOME/software/freesasa-2.0.3/bin/freesasa
+```
+
+## HSSPCONV
+```bash
+$ mkdir -p $HOME/software && cd $HOME/software
+$ wget https://github.com/cmbi/hssp/archive/3.1.5.tar.gz
+$ tar -zxf 3.1.5.tar.gz && rm 3.1.5.tar.gz
+$ cd hssp-3.1.5
+$ ./autogen.she
+$ ./configure
+$ make hsspconv
+$ export HSSPCONV_BIN=$HOME/software/hssp-3.1.5/hsspconv
+```
+
+## Protdist
+
+Protdist is distributed together with WHISCY, you can find it in the `whiscy` directory. We are working on a better way to install this dependency 🙂
+
+```bash
+$ mkdir -p $HOME/software && cd $HOME/software
+$ git clone https://github.com/haddocking/whiscy
+$ mv whiscy/bin/protdist . && rm -rf whiscy
+$ cd protdist
+$ bash compile.sh
+$ export PROTDIST_BIN=$HOME/software/protdist/protdist
+```
+
+---
+
+In the end the system variables that define the third-party dependencies should look like this:
+
+```bash
+export MUSCLE_BIN=$HOME/software/muscle3.8.1551/muscle
+export FREESASA_BIN=$HOME/software/freesasa-2.0.3/bin/freesasa
+export HSSPCONV_BIN=$HOME/software/hssp-3.1.5/hsspconv
+export PROTDIST_BIN=$HOME/software/protdist/protdist
+```

libwhiscy/__init__.py

@@ -0,0 +1,38 @@
+import os
+from pathlib import Path
+
+MUSCLE_BIN = os.environ.get("MUSCLE_BIN")
+FREESASA_BIN = os.environ.get("FREESASA_BIN")
+HSSPCONV_BIN = os.environ.get("HSSPCONV_BIN")
+PROTDIST_BIN = os.environ.get("PROTDIST_BIN")
+
+# Make sure none of them are None
+if MUSCLE_BIN is None:
+    raise ValueError("MUSCLE_BIN not found in system variables")
+
+if FREESASA_BIN is None:
+    raise ValueError("FREESASA_BIN not found in system variables")
+
+if HSSPCONV_BIN is None:
+    raise ValueError("HSSPCONV_BIN not found in system variables")
+
+if PROTDIST_BIN is None:
+    raise ValueError("PROTDIST_BIN not found in system variables")
+
+PARAM_PATH = Path(Path(__file__).parent.parent, "param")
+
+
+CUTOFF = {
+    "sa_pred_cutoff": 15.0,
+    "sa_act_cutoff": 40.0,
+    "air_cutoff": 0.18,
+    "air_dist_cutoff": 6.5,
+}
+
+AIR = {
+    "air_pro_percentage": 10.0,
+    "air_wm_pro_or": 98.52,
+    "air_wm_whis_or": 0.370515,
+    "air_wm_pro_and": 55.42,
+    "air_wm_whis_and": 0.106667,
+}

libwhiscy/access.py

@@ -1,16 +1,17 @@
 import os
 import subprocess
 
+from libwhiscy import FREESASA_BIN
+
 
 def calculate_accessibility(pdb_file_name, output_file_name):
     """Calculates the SASA using freesasa.
 
     Uses the command line interface and not the Python bindings to be able to get
     a RSA NACCESS-format like file.
     """
-    cmd = "freesasa {} -n 20 --format=rsa --radii=naccess -o {}".format(
-        pdb_file_name, output_file_name
-    )
+    cmd = f"{FREESASA_BIN} {pdb_file_name} -n 20 --format=rsa --radii=naccess -o {output_file_name}"
+
     try:
         subprocess.run(cmd, shell=True)
     except:

libwhiscy/pdbutil.py

@@ -1,4 +1,3 @@
-
 """Util functions involving structure and PDB files"""
 
 import os
@@ -13,21 +12,26 @@
 
 class NotAlternative(Select):
     """Removes alternative AAs"""
+
     def accept_residue(self, residue):
-        return (is_aa(residue) and residue.id[2] == ' ')
+        return is_aa(residue) and residue.id[2] == " "
 
 
 _hydrogen = re.compile("[123 ]*H.*")
+
+
 def is_hydrogen(atom):
     """Checks if atom is an hydrogen"""
-    name = atom.get_id() 
+    name = atom.get_id()
     return _hydrogen.match(name)
 
 
-def download_pdb_structure(pdb_code, pdb_file_name, file_path='.'):
+def download_pdb_structure(pdb_code, pdb_file_name, file_path="."):
     """Downloads a PDB structure from the Protein Data Bank"""
     pdbl = PDBList()
-    file_name = pdbl.retrieve_pdb_file(pdb_code, file_format='pdb', pdir=file_path, overwrite=True)
+    file_name = pdbl.retrieve_pdb_file(
+        pdb_code, file_format="pdb", pdir=file_path, overwrite=True
+    )
     if os.path.exists(file_name):
         os.rename(file_name, pdb_file_name)
     else:
@@ -44,7 +48,7 @@ def get_pdb_sequence(input_pdb_file, chain_id, mapping_output=False, with_gaps=F
     residues = list(chain)
     for res in residues:
         # Remove alternative location residues
-        if "CA" in res.child_dict and is_aa(res) and res.id[2] == ' ':
+        if "CA" in res.child_dict and is_aa(res) and res.id[2] == " ":
             try:
                 mapping[res.id[1]] = three_to_one(res.get_resname())
             except KeyError:
@@ -57,35 +61,41 @@ def get_pdb_sequence(input_pdb_file, chain_id, mapping_output=False, with_gaps=F
         start, end = res_numbers[0], res_numbers[-1]
         missing = sorted(set(range(start, end + 1)).difference(res_numbers))
         for m in missing:
-            mapping[m] = '-'
+            mapping[m] = "-"
 
     if mapping_output:
         return mapping
     else:
-        return ''.join([mapping[k] for k in sorted(mapping.keys())])
+        return "".join([mapping[k] for k in sorted(mapping.keys())])
 
 
-def map_protein_to_sequence_alignment(pdb_file, chain_id, sequence, phylip_file, output_file_name):
+def map_protein_to_sequence_alignment(
+    pdb_file, chain_id, sequence, phylip_file, output_file_name
+):
     """Creates a dictionary .conv file mapping protein residue numeration to aligment"""
     mapping = get_pdb_sequence(pdb_file, chain_id, mapping_output=True)
     # Check if sequence is the same
-    pdb_seq = ''.join([mapping[k] for k in sorted(mapping.keys())])
+    pdb_seq = "".join([mapping[k] for k in sorted(mapping.keys())])
     if pdb_seq != sequence:
         raise SystemExit("ERROR: PDB sequence doest not match sequence alignment")
 
     # Account for gaps in phylipseq file
-    #alignment = list(AlignIO.parse(phylip_file, format='phylip-sequential'))[0]
-    #master_phylip = alignment[0].seq
-    #if str(master_phylip.ungap('-')) != pdb_seq:
+    # alignment = list(AlignIO.parse(phylip_file, format='phylip-sequential'))[0]
+    # master_phylip = alignment[0].seq
+    # if str(master_phylip.ungap('-')) != pdb_seq:
     #    raise SystemExit("ERROR: PDB sequence doest not match sequence alignment in phylip file")
 
-    with open(output_file_name, 'w') as output_handle:
-        output_handle.write("# Conversion table from {} and chain {} to sequence{}".format(pdb_file,
-                                                                                           chain_id,
-                                                                                           os.linesep))
+    with open(output_file_name, "w") as output_handle:
+        output_handle.write(
+            "# Conversion table from {} and chain {} to sequence{}".format(
+                pdb_file, chain_id, os.linesep
+            )
+        )
         seq_res_id = 1
         for pdb_res_id in sorted(mapping.keys()):
             # Do not map gaps if any
-            if mapping[pdb_res_id] != '-':
-                output_handle.write("{0}     {1}{2}".format(pdb_res_id, seq_res_id, os.linesep))
+            if mapping[pdb_res_id] != "-":
+                output_handle.write(
+                    "{0}     {1}{2}".format(pdb_res_id, seq_res_id, os.linesep)
+                )
             seq_res_id += 1

tests/__init__.py

@@ -2,24 +2,6 @@
 
 GOLDEN_DATA_PATH = Path(Path(__file__).parent, "golden_data")
 
-WHISCY_PATH = Path(__file__).parent.parent
-
-WHISCY_BIN = Path(WHISCY_PATH, "whiscy.py")
-
-FREESASA_PATH = Path(WHISCY_PATH, "bin", "freesasa", "bin")
-FREESASA_EXEC = Path(FREESASA_PATH, "freesasa")
-
-# If `FREESASA_EXEC` is found, add it to the PATH
-if FREESASA_EXEC.exists():
-    import os
-
-    os.environ["PATH"] += os.pathsep + str(FREESASA_PATH)
-else:
-    # Fail the tests, we need it to run the tests
-    raise FileNotFoundError(f"{FREESASA_EXEC} not found")
-
-MUSCLE_BIN = Path(WHISCY_PATH, "bin", "muscle3.8.1551")
-
 
 PAM_EXPECTED_SCORES = [
     0.001239810431396115,