Update to version 1.5.0

see changelog for details

changelog.txt

@@ -1,3 +1,9 @@
+Version 1.5.0
+    - Gene report function is now functional
+    - Reorganized features into classes, making adding new features in the future easier
+        This how features are stored and retrieved from the database, thus a new major version change was necessary.
+    - Improved the installer
+
 Version 1.4.1
     - Paralellized th MicroMiner lookup pipeline
     - Continued implementation of the gene report function

install.sh

@@ -38,20 +38,23 @@ do
     esac
 done
 
+verbose_stdout=3
+verbose_stderr=4
+
 if [ "$verbose" = true ]; then
-    verbose_stdout=/dev/stdout
-    verbose_stderr=/dev/stderr
+    eval "exec $verbose_stdout>&1"
+    eval "exec $verbose_stderr>&2"
 else
-    verbose_stdout=/dev/null
-    verbose_stderr=/dev/null
+    eval "exec $verbose_stdout>/dev/null"
+    eval "exec $verbose_stderr>/dev/null"
 fi
 
 #Check if conda environment already exits, create it if not
 env_list_result=$(conda env list | grep "$env_name")
 if [ -z "$env_list_result" ]
 then
     echo "Conda environment with name $env_name not in current environment list, setting up new environment ..."
-    conda create -n "$env_name" python=$current_python_version -y >$verbose_stdout
+    conda create -n "$env_name" python=$current_python_version -y >&$verbose_stdout
 else
     echo "Conda environment with name $env_name already in environment list."
 fi
@@ -68,13 +71,13 @@ new_env_path=$(conda env list | awk -v name="$env_name" '/^[^#]/{ if ($1 == name
     source "$conda_bash_path"
     conda activate "$env_name"
     echo "$new_env_path"" activated"
-} >$verbose_stdout
+} >&$verbose_stdout
 
 mamba_version_test_output=$(mamba --version 2>/dev/null)
 
 if [ -z "$mamba_version_test_output" ]
 then
-    conda install -y -c conda-forge mamba >$verbose_stdout 2>$verbose_stderr
+    conda install -y -c conda-forge mamba >&$verbose_stdout 2>&$verbose_stderr
 fi
 
 #install dependencies
@@ -94,15 +97,18 @@ fi
     mamba install -y requests-toolbelt
     echo "Installing package pymol ..."    
     mamba install -y -c conda-forge pymol-open-source
-} >$verbose_stdout
+} >&$verbose_stdout
 
 #install the main package
 echo "Installing StructMAn source code using pip ..."
-pip install "$SCRIPTPATH" >$verbose_stdout
+pip install "$SCRIPTPATH" >&$verbose_stdout
 
 #install mmseqs2
 echo "Installing MMseqs2 ..."
-mamba install -y -c bioconda mmseqs2 >$verbose_stdout
+mamba install -y -c bioconda mmseqs2 >&$verbose_stdout
+
+#install wkhtmltopdf
+mamba install -y -c conda-forge wkhtmltopdf >&$verbose_stdout
 
 if [ -z $storage_folder ]
 then
@@ -134,10 +140,10 @@ fi
 cp "$path_to_dssp_binary" "$new_env_path"/bin/smssp
 
 #Installing specific boost
-mamba install -y libboost==1.82 >$verbose_stdout
+mamba install -y libboost==1.82 >&$verbose_stdout
 
 #Installing grpc
-mamba install -y conda-forge::grpc-cpp==1.51.1 >$verbose_stdout
+pip install grpcio >$verbose_stdout
 
 #Init config file
 structman_config_path="$new_env_path"/lib/python"$current_python_version"/site-packages/structman/structman_config.txt
@@ -148,7 +154,7 @@ echo "    $structman_config_path"
     structman config mmseqs_tmp_folder "$tmp_folder_path" -c "$structman_config_path"
     structman config dssp_path smssp -c "$structman_config_path"
     structman config mmseqs2_db_path "$storage_folder"/pdbba_search_db_mmseqs2
-} >$verbose_stdout 2>$verbose_stderr
+} >&$verbose_stdout 2>&$verbose_stderr
 
 #install modeller
 if ! [ -z "$modeller_key" ]
@@ -161,7 +167,7 @@ then
         mamba install -y modeller
         echo "Setting the given modeller key to the modeller config ..."
         structman set_m_key "$modeller_key"
-    } >$verbose_stdout 2>$verbose_stderr
+    } >&$verbose_stdout 2>&$verbose_stderr
 fi
 
 echo "Setting StructMAn config and database ..."
@@ -177,7 +183,7 @@ then
         structman config db_address "$database_server" -c "$structman_config_path"
         echo "Setting up the database ..."
         structman database create --compute_ppi
-    } >$verbose_stdout 2>$verbose_stderr
+    } >&$verbose_stdout 2>&$verbose_stderr
 else
     database_name=main_db
     local_database_path="$local_database_folder"/local_structman_database.db
@@ -187,10 +193,10 @@ else
         structman config db_name "$database_name" -c "$structman_config_path"
         structman config db_address "-"
         structman database create --compute_ppi
-    } >$verbose_stdout 2>$verbose_stderr
+    } >&$verbose_stdout 2>&$verbose_stderr
 fi
 
-structman update check_search_db >$verbose_stdout 2>$verbose_stderr
+structman update check_search_db >&$verbose_stdout 2>&$verbose_stderr
 
 
 echo "StructMAn successfully installed, please activate the right conda environment before using it:"

setup.py

@@ -44,7 +44,7 @@
         "numpy>=1.22.3",
         "psutil>=5.8.0",
         "pymysql>=1.0.2",
-        "ray==2.9.1",
+        "ray==2.10.0",
         "msgpack>=1.0.3",
         "zstd>=1.5.2.5",
         "pandas>=1.3.3",
@@ -55,7 +55,8 @@
         "pycairo>=1.23.0",
         "powerlaw>=1.5",
         "biotite>=0.38.0",
-        "markdown>=2.6.9"
+        "markdown>=2.6.9",
+        "pdfkit==1.0.0"
     ],
 
     package_data = {

structman/_version.py

@@ -1 +1 @@
-__version__ = '1.4.1'
+__version__ = '1.5.0'

structman/base_utils/base_utils.py

@@ -7,6 +7,11 @@
 
 from structman import settings
 
+try:
+    from structguy.sampleSpace import CrossValidationSlice
+except:
+    pass
+
 import structman.lib.sdsc as sdsc
 
 import zstd
@@ -330,6 +335,32 @@ def custom_encoder(obj):
             serialized_snv.append(obj.__getattribute__(attribute_name))
         return {'__snv__': True, 'as_list': serialized_snv}
 
+    if 'Microminer_features' in str(type(obj)): #isinstance just won't work, don't know why
+        serialized_object = []
+        for attribute_name in obj.__slots__:
+            serialized_object.append(obj.__getattribute__(attribute_name))
+        return {'__Microminer_features__': True, 'as_list': serialized_object}
+
+    if 'Structural_features' in str(type(obj)): #isinstance just won't work, don't know why
+        serialized_object = []
+        for attribute_name in obj.__slots__:
+            serialized_object.append(obj.__getattribute__(attribute_name))
+        return {'__Structural_features__': True, 'as_list': serialized_object}
+
+    if 'Integrated_features' in str(type(obj)): #isinstance just won't work, don't know why
+        serialized_object = []
+        for attribute_name in obj.__slots__:
+            serialized_object.append(obj.__getattribute__(attribute_name))
+        return {'__Integrated_features__': True, 'as_list': serialized_object}
+
+    if 'RIN_based_features' in str(type(obj)): #isinstance just won't work, don't know why
+        serialized_object = obj.get_raw_list()
+        return {'__RIN_based_features__': True, 'as_list': serialized_object}
+
+    if 'CrossValidationSlice' in str(type(obj)): #isinstance just won't work, don't know why
+        serialized_object = obj.get_raw_list()
+        return {'__CrossValidationSlice__': True, 'as_list': serialized_object}
+
     return obj
 
 def custom_decoder(obj):
@@ -449,6 +480,40 @@ def custom_decoder(obj):
         for i, attribute_name in enumerate(snv_obj.__slots__):
             snv_obj.__setattr__(attribute_name, serialized_snv[i])
         return snv_obj
+
+    if '__Microminer_features__' in obj:
+        serialized_object = obj['as_list']
+        rebuild_obj = sdsc.mappings.Microminer_features()
+        for i, attribute_name in enumerate(rebuild_obj.__slots__):
+            rebuild_obj.__setattr__(attribute_name, serialized_object[i])
+        return rebuild_obj
+
+    if '__Structural_features__' in obj:
+        serialized_object = obj['as_list']
+        rebuild_obj = sdsc.mappings.Structural_features()
+        for i, attribute_name in enumerate(rebuild_obj.__slots__):
+            rebuild_obj.__setattr__(attribute_name, serialized_object[i])
+        return rebuild_obj
+
+    if '__Integrated_features__' in obj:
+        serialized_object = obj['as_list']
+        rebuild_obj = sdsc.mappings.Integrated_features()
+        for i, attribute_name in enumerate(rebuild_obj.__slots__):
+            rebuild_obj.__setattr__(attribute_name, serialized_object[i])
+        return rebuild_obj
+
+    if '__RIN_based_features__' in obj:
+        raw_rin_based_features = obj['as_list']
+        rebuild_obj = sdsc.mappings.RIN_based_features()
+        rebuild_obj.set_values(raw_rin_based_features)
+        return rebuild_obj
+
+    if '__CrossValidationSlice__' in obj:
+        serialized_object = obj['as_list']
+        rebuild_obj = CrossValidationSlice(raw_init =  True)
+        for i, attribute_name in enumerate(rebuild_obj.__slots__):
+            rebuild_obj.__setattr__(attribute_name, serialized_object[i])
+        return rebuild_obj
 
     return obj
 

structman/lib/MMseqs2.py

@@ -128,7 +128,7 @@ def apply_mmseqs(mmseqs_tmp_folder, mmseqs2_path, temp_fasta, search_db, gigs_of
     temp_outfile = '%s/tmp_outfile_%s.fasta' % (mmseqs_tmp_folder, randomString())
 
     if verbosity >= 2:
-        print(mmseqs2_path, 'easy-search', temp_fasta, search_db, temp_outfile, mmseqs_tmp_folder)
+        print(f'\nApply MMseqs2: {mmseqs2_path}, easy-search, {temp_fasta}, {search_db}, {temp_outfile}, {mmseqs_tmp_folder}\n')
 
     out_format_str = 'query,target,fident,alnlen,tlen,qcov'
 
@@ -175,12 +175,26 @@ def apply_mmseqs(mmseqs_tmp_folder, mmseqs2_path, temp_fasta, search_db, gigs_of
 
     if verbosity >= 3:
         print(f'MMseqs2 results parsed: size of hit map: {len(hits)}')
+    if verbosity >= 4 and len(hits) < 50:
+        print(hits)
 
     os.remove(temp_outfile)
 
     return hits, pdb_ids, debug_store
 
 
+def wipe_folder(config, folder_path):
+    for fn in os.listdir(folder_path):
+        subfolder_path = '%s/%s' % (folder_path, fn)
+        if os.path.exists(subfolder_path):
+            if os.path.getmtime(subfolder_path) > config.prog_start_time:
+                try:
+                    shutil.rmtree(subfolder_path)
+                except:
+                    if config.verbosity >= 4:
+                        config.errorlog.add_warning('Tmp folder wipe failed for: %s' % subfolder_path)
+
+
 # called by serializePipeline
 def search(proteins, config, custom_db=False):
 
@@ -235,6 +249,8 @@ def search(proteins, config, custom_db=False):
 
     t0 = time.time()
 
+    wipe_folder(config, mmseqs_tmp_folder)
+
     temp_fasta = '%s/tmp_%s.fasta' % (mmseqs_tmp_folder, randomString())
     to_fasta_out = geneSeqMapToFasta(proteins, temp_fasta, config)
 
@@ -271,15 +287,7 @@ def search(proteins, config, custom_db=False):
                     hits[u_ac] = {(pdb_id, chain): [100.0, 1.0, [chain], len(proteins.get_sequence(u_ac)), len(proteins.get_sequence(u_ac))]}
                     pdb_ids.add(pdb_id)
 
-        for fn in os.listdir(mmseqs_tmp_folder):
-            subfolder_path = '%s/%s' % (mmseqs_tmp_folder, fn)
-            if os.path.exists(subfolder_path):
-                if os.path.getmtime(subfolder_path) > config.prog_start_time:
-                    try:
-                        shutil.rmtree(subfolder_path)
-                    except:
-                        if config.verbosity >= 4:
-                            config.errorlog.add_warning('Tmp folder wipe failed for: %s' % subfolder_path)
+        wipe_folder(config, mmseqs_tmp_folder)
 
         search_results.append((hits, pdb_ids, is_model_db))