rename update script

.gitignore

@@ -6,8 +6,8 @@ tools/symop-map-generator
 test/unit-test
 test/pdb2cif-test
 test/rename-compound-test
-tools/update-dictionary-script
-data/
+tools/update-libcifpp-data
+data/components.cif*
 CMakeSettings.json
 msvc/
 Testing/

CMakeLists.txt

@@ -287,16 +287,17 @@ if (CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
     target_link_options(cifpp PRIVATE -undefined dynamic_lookup)
 endif (CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
 
-# download the components.cif file from CCD
-set(COMPONENTS_CIF ${PROJECT_SOURCE_DIR}/data/components.cif)
+option(DOWNLOAD_CCD "Download the CCD file components.cif during installation" OFF)
+if(DOWNLOAD_CCD)
+	# download the components.cif file from CCD
+	set(COMPONENTS_CIF ${PROJECT_SOURCE_DIR}/data/components.cif)
 
-if (NOT EXISTS ${COMPONENTS_CIF})
+	if (NOT EXISTS ${COMPONENTS_CIF})
 
-	if (NOT EXISTS ${PROJECT_SOURCE_DIR}/data)
-		file(MAKE_DIRECTORY ${PROJECT_SOURCE_DIR}/data/)
-	endif()
+		if (NOT EXISTS ${PROJECT_SOURCE_DIR}/data)
+			file(MAKE_DIRECTORY ${PROJECT_SOURCE_DIR}/data/)
+		endif()
 
-	# if(${CMAKE_VERSION} VERSION_LESS "3.19.0")
 		find_program(GUNZIP gunzip)
 
 		if(GUNZIP)
@@ -309,16 +310,10 @@ if (NOT EXISTS ${COMPONENTS_CIF})
 			file(DOWNLOAD ftp://ftp.wwpdb.org/pub/pdb/data/monomers/components.cif ${COMPONENTS_CIF}
 				SHOW_PROGRESS)
 		endif()
-	# else()
-	# 	file(DOWNLOAD ftp://ftp.wwpdb.org/pub/pdb/data/monomers/components.cif.gz ${COMPONENTS_CIF}.gz
-	# 		SHOW_PROGRESS)
-	# 	file(ARCHIVE_EXTRACT INPUT ${COMPONENTS_CIF}.gz
-	# 		DESTINATION ${CMAKE_SOURCE_DIR}/data/
-	# 		VERBOSE)
-	# endif()
-endif()
+	endif()
 
-add_custom_target(COMPONENTS ALL DEPENDS ${COMPONENTS_CIF})
+	add_custom_target(COMPONENTS ALL DEPENDS ${COMPONENTS_CIF})
+endif()
 
 if(UNIX)
 	option(INSTALL_UPDATE_SCRIPT "Install the script to update CCD and dictionary files" OFF)
@@ -471,9 +466,9 @@ message("Will install in ${CMAKE_INSTALL_PREFIX}")
 if(INSTALL_UPDATE_SCRIPT)
 	set(CIFPP_CRON_DIR "$ENV{DESTDIR}/etc/cron.weekly")
 
-	configure_file(${CMAKE_SOURCE_DIR}/tools/update-dictionary-script.in update-dictionary-script @ONLY)
+	configure_file(${CMAKE_SOURCE_DIR}/tools/update-libcifpp-data.in update-libcifpp-data @ONLY)
 	install(
-		FILES ${CMAKE_CURRENT_BINARY_DIR}/update-dictionary-script
+		FILES ${CMAKE_CURRENT_BINARY_DIR}/update-libcifpp-data
 		DESTINATION ${CIFPP_CRON_DIR}
 		PERMISSIONS OWNER_EXECUTE OWNER_READ GROUP_EXECUTE GROUP_READ WORLD_EXECUTE WORLD_READ
 	)
@@ -482,13 +477,14 @@ if(INSTALL_UPDATE_SCRIPT)
 	install(DIRECTORY DESTINATION "$ENV{DESTDIR}/etc/libcifpp/cache-update.d")
 
 	# a config to, to make it complete
-	file(WRITE ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf [[# Uncomment the next line to enable automatic updates
+	if(NOT EXISTS "$ENV{DESTDIR}/etc/libcifpp.conf")
+		file(WRITE ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf [[# Uncomment the next line to enable automatic updates
 # update=true
 ]])
-	install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf DESTINATION "$ENV{DESTDIR}/etc")
+		install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf DESTINATION "$ENV{DESTDIR}/etc")
+		install(CODE "message(\"A configuration file has been written to $ENV{DESTDIR}/etc/libcifpp.conf, please edit this file to enable automatic updates\")")
+	endif()
 
 	target_compile_definitions(cifpp PUBLIC CACHE_DIR="${CIFPP_CACHE_DIR}")
-
-	install(CODE "message(\"A configuration file has been written to $ENV{DESTDIR}/etc/libcifpp.conf, please edit this file to enable automatic updates\")")
 endif()
 

src/Compound.cpp

@@ -457,7 +457,7 @@ Compound *CCDCompoundFactoryImpl::create(const std::string &id)
 	{
 		ccd = cif::loadResource("components.cif");
 		if (not ccd)
-			throw std::runtime_error("Could not locate the CCD components.cif file, please make sure the software is installed properly and/or use the update-dictionary-script to fetch the data.");
+			throw std::runtime_error("Could not locate the CCD components.cif file, please make sure the software is installed properly and/or use the update-libcifpp-data to fetch the data.");
 	}
 	else
 		ccd.reset(new std::ifstream(mCompoundsFile));
@@ -484,7 +484,7 @@ Compound *CCDCompoundFactoryImpl::create(const std::string &id)
 		{
 			ccd = cif::loadResource("components.cif");
 			if (not ccd)
-				throw std::runtime_error("Could not locate the CCD components.cif file, please make sure the software is installed properly and/or use the update-dictionary-script to fetch the data.");
+				throw std::runtime_error("Could not locate the CCD components.cif file, please make sure the software is installed properly and/or use the update-libcifpp-data to fetch the data.");
 		}
 		else
 			ccd.reset(new std::ifstream(mCompoundsFile));

test/REA.cif

@@ -0,0 +1,188 @@
+data_REA
+# 
+_chem_comp.id                                    REA 
+_chem_comp.name                                  "RETINOIC ACID" 
+_chem_comp.type                                  NON-POLYMER 
+_chem_comp.pdbx_type                             HETAIN 
+_chem_comp.formula                               "C20 H28 O2" 
+_chem_comp.mon_nstd_parent_comp_id               ? 
+_chem_comp.pdbx_synonyms                         ? 
+_chem_comp.pdbx_formal_charge                    0 
+_chem_comp.pdbx_initial_date                     1999-07-08 
+_chem_comp.pdbx_modified_date                    2016-10-18 
+_chem_comp.pdbx_ambiguous_flag                   N 
+_chem_comp.pdbx_release_status                   REL 
+_chem_comp.pdbx_replaced_by                      ? 
+_chem_comp.pdbx_replaces                         3KV 
+_chem_comp.formula_weight                        300.435 
+_chem_comp.one_letter_code                       ? 
+_chem_comp.three_letter_code                     REA 
+_chem_comp.pdbx_model_coordinates_details        ? 
+_chem_comp.pdbx_model_coordinates_missing_flag   N 
+_chem_comp.pdbx_ideal_coordinates_details        Corina 
+_chem_comp.pdbx_ideal_coordinates_missing_flag   N 
+_chem_comp.pdbx_model_coordinates_db_code        1CBS 
+_chem_comp.pdbx_subcomponent_list                ? 
+_chem_comp.pdbx_processing_site                  RCSB 
+# 
+loop_
+_chem_comp_atom.comp_id 
+_chem_comp_atom.atom_id 
+_chem_comp_atom.alt_atom_id 
+_chem_comp_atom.type_symbol 
+_chem_comp_atom.charge 
+_chem_comp_atom.pdbx_align 
+_chem_comp_atom.pdbx_aromatic_flag 
+_chem_comp_atom.pdbx_leaving_atom_flag 
+_chem_comp_atom.pdbx_stereo_config 
+_chem_comp_atom.model_Cartn_x 
+_chem_comp_atom.model_Cartn_y 
+_chem_comp_atom.model_Cartn_z 
+_chem_comp_atom.pdbx_model_Cartn_x_ideal 
+_chem_comp_atom.pdbx_model_Cartn_y_ideal 
+_chem_comp_atom.pdbx_model_Cartn_z_ideal 
+_chem_comp_atom.pdbx_component_atom_id 
+_chem_comp_atom.pdbx_component_comp_id 
+_chem_comp_atom.pdbx_ordinal 
+REA C1   C1   C 0 1 N N N 21.972 29.831 16.739 -4.684 0.932  -0.497 C1   REA 1  
+REA C2   C2   C 0 1 N N N 20.921 30.524 15.841 -5.837 0.190  -1.176 C2   REA 2  
+REA C3   C3   C 0 1 N N N 20.245 29.635 14.848 -6.441 -0.798 -0.171 C3   REA 3  
+REA C4   C4   C 0 1 N N N 19.555 28.479 15.488 -5.418 -1.903 0.100  C4   REA 4  
+REA C5   C5   C 0 1 N N N 20.389 27.812 16.587 -4.082 -1.301 0.429  C5   REA 5  
+REA C6   C6   C 0 1 N N N 21.425 28.446 17.218 -3.756 -0.048 0.161  C6   REA 6  
+REA C7   C7   C 0 1 N N N 22.242 27.851 18.297 -2.457 0.396  0.516  C7   REA 7  
+REA C8   C8   C 0 1 N N N 21.868 26.977 19.240 -1.363 -0.229 0.007  C8   REA 8  
+REA C9   C9   C 0 1 N N N 22.705 26.434 20.286 -0.076 0.257  0.298  C9   REA 9  
+REA C10  C10  C 0 1 N N N 22.159 25.536 21.131 1.022  -0.370 -0.213 C10  REA 10 
+REA C11  C11  C 0 1 N N N 22.875 24.924 22.234 2.306  0.115  0.077  C11  REA 11 
+REA C12  C12  C 0 1 N N N 22.237 24.026 22.990 3.405  -0.513 -0.435 C12  REA 12 
+REA C13  C13  C 0 1 N N N 22.856 23.377 24.125 4.689  -0.028 -0.144 C13  REA 13 
+REA C14  C14  C 0 1 N N N 22.135 22.473 24.834 5.787  -0.655 -0.656 C14  REA 14 
+REA C15  C15  C 0 1 N N N 22.563 21.710 26.016 7.077  -0.265 -0.244 C15  REA 15 
+REA C16  C16  C 0 1 N N N 22.238 30.737 17.948 -5.246 1.886  0.559  C16  REA 16 
+REA C17  C17  C 0 1 N N N 23.292 29.620 15.948 -3.911 1.737  -1.544 C17  REA 17 
+REA C18  C18  C 0 1 N N N 19.791 26.449 16.947 -3.056 -2.175 1.103  C18  REA 18 
+REA C19  C19  C 0 1 N N N 24.181 26.841 20.385 0.090  1.471  1.175  C19  REA 19 
+REA C20  C20  C 0 1 N N N 24.303 23.747 24.489 4.855  1.186  0.733  C20  REA 20 
+REA O1   O1   O 0 1 N N N 23.640 21.075 25.978 7.210  0.553  0.648  O1   REA 21 
+REA O2   O2   O 0 1 N N N 21.840 21.712 27.037 8.166  -0.798 -0.840 O2   REA 22 
+REA H21  H21  H 0 1 N N N 20.147 30.955 16.494 -6.598 0.905  -1.490 H21  REA 23 
+REA H22  H22  H 0 1 N N N 21.425 31.330 15.288 -5.462 -0.353 -2.044 H22  REA 24 
+REA H31  H31  H 0 1 N N N 19.501 30.227 14.295 -6.673 -0.278 0.759  H31  REA 25 
+REA H32  H32  H 0 1 N N N 21.001 29.250 14.148 -7.349 -1.234 -0.586 H32  REA 26 
+REA H41  H41  H 0 1 N N N 18.613 28.835 15.931 -5.756 -2.511 0.938  H41  REA 27 
+REA H42  H42  H 0 1 N N N 19.335 27.730 14.713 -5.322 -2.531 -0.786 H42  REA 28 
+REA H7   H7   H 0 1 N N N 23.276 28.162 18.329 -2.337 1.230  1.191  H7   REA 29 
+REA H8   H8   H 0 1 N N N 20.840 26.645 19.217 -1.482 -1.100 -0.622 H8   REA 30 
+REA H10  H10  H 0 1 N N N 21.127 25.256 20.977 0.903  -1.241 -0.842 H10  REA 31 
+REA H11  H11  H 0 1 N N N 23.902 25.189 22.440 2.425  0.985  0.706  H11  REA 32 
+REA H12  H12  H 0 1 N N N 21.216 23.774 22.743 3.286  -1.383 -1.063 H12  REA 33 
+REA H14  H14  H 0 1 N N N 21.127 22.292 24.490 5.667  -1.451 -1.376 H14  REA 34 
+REA H161 H161 H 0 0 N N N 22.984 30.265 18.604 -5.802 1.316  1.303  H161 REA 35 
+REA H162 H162 H 0 0 N N N 22.618 31.709 17.601 -4.426 2.415  1.044  H162 REA 36 
+REA H163 H163 H 0 0 N N N 21.302 30.887 18.506 -5.911 2.605  0.081  H163 REA 37 
+REA H171 H171 H 0 0 N N N 24.033 29.127 16.595 -4.598 2.394  -2.077 H171 REA 38 
+REA H172 H172 H 0 0 N N N 23.095 28.989 15.069 -3.146 2.335  -1.050 H172 REA 39 
+REA H173 H173 H 0 0 N N N 23.683 30.595 15.620 -3.439 1.054  -2.251 H173 REA 40 
+REA H181 H181 H 0 0 N N N 20.397 25.979 17.736 -3.448 -3.187 1.201  H181 REA 41 
+REA H182 H182 H 0 0 N N N 18.761 26.584 17.308 -2.145 -2.194 0.503  H182 REA 42 
+REA H183 H183 H 0 0 N N N 19.786 25.804 16.056 -2.831 -1.775 2.092  H183 REA 43 
+REA H191 H191 H 0 0 N N N 24.647 26.327 21.238 0.171  1.159  2.216  H191 REA 44 
+REA H192 H192 H 0 0 N N N 24.702 26.559 19.458 0.993  2.008  0.885  H192 REA 45 
+REA H193 H193 H 0 0 N N N 24.252 27.929 20.529 -0.774 2.125  1.058  H193 REA 46 
+REA H201 H201 H 0 0 N N N 24.620 23.168 25.369 5.026  0.871  1.762  H201 REA 47 
+REA H202 H202 H 0 0 N N N 24.965 23.516 23.641 5.707  1.771  0.386  H202 REA 48 
+REA H203 H203 H 0 0 N N N 24.360 24.822 24.717 3.952  1.795  0.685  H203 REA 49 
+REA HO2  HO2  H 0 1 N N N 22.244 21.180 27.713 9.006  -0.469 -0.490 HO2  REA 50 
+# 
+loop_
+_chem_comp_bond.comp_id 
+_chem_comp_bond.atom_id_1 
+_chem_comp_bond.atom_id_2 
+_chem_comp_bond.value_order 
+_chem_comp_bond.pdbx_aromatic_flag 
+_chem_comp_bond.pdbx_stereo_config 
+_chem_comp_bond.pdbx_ordinal 
+REA C1  C2   SING N N 1  
+REA C1  C6   SING N N 2  
+REA C1  C16  SING N N 3  
+REA C1  C17  SING N N 4  
+REA C2  C3   SING N N 5  
+REA C2  H21  SING N N 6  
+REA C2  H22  SING N N 7  
+REA C3  C4   SING N N 8  
+REA C3  H31  SING N N 9  
+REA C3  H32  SING N N 10 
+REA C4  C5   SING N N 11 
+REA C4  H41  SING N N 12 
+REA C4  H42  SING N N 13 
+REA C5  C6   DOUB N N 14 
+REA C5  C18  SING N N 15 
+REA C6  C7   SING N N 16 
+REA C7  C8   DOUB N E 17 
+REA C7  H7   SING N N 18 
+REA C8  C9   SING N N 19 
+REA C8  H8   SING N N 20 
+REA C9  C10  DOUB N E 21 
+REA C9  C19  SING N N 22 
+REA C10 C11  SING N N 23 
+REA C10 H10  SING N N 24 
+REA C11 C12  DOUB N E 25 
+REA C11 H11  SING N N 26 
+REA C12 C13  SING N N 27 
+REA C12 H12  SING N N 28 
+REA C13 C14  DOUB N E 29 
+REA C13 C20  SING N N 30 
+REA C14 C15  SING N N 31 
+REA C14 H14  SING N N 32 
+REA C15 O1   DOUB N N 33 
+REA C15 O2   SING N N 34 
+REA C16 H161 SING N N 35 
+REA C16 H162 SING N N 36 
+REA C16 H163 SING N N 37 
+REA C17 H171 SING N N 38 
+REA C17 H172 SING N N 39 
+REA C17 H173 SING N N 40 
+REA C18 H181 SING N N 41 
+REA C18 H182 SING N N 42 
+REA C18 H183 SING N N 43 
+REA C19 H191 SING N N 44 
+REA C19 H192 SING N N 45 
+REA C19 H193 SING N N 46 
+REA C20 H201 SING N N 47 
+REA C20 H202 SING N N 48 
+REA C20 H203 SING N N 49 
+REA O2  HO2  SING N N 50 
+# 
+loop_
+_pdbx_chem_comp_descriptor.comp_id 
+_pdbx_chem_comp_descriptor.type 
+_pdbx_chem_comp_descriptor.program 
+_pdbx_chem_comp_descriptor.program_version 
+_pdbx_chem_comp_descriptor.descriptor 
+REA SMILES           ACDLabs              12.01 "C1(CCCC(=C1\C=C\C(=C\C=C\C(=C\C(=O)O)C)C)C)(C)C"                                                                                                     
+REA InChI            InChI                1.03  "InChI=1S/C20H28O2/c1-15(8-6-9-16(2)14-19(21)22)11-12-18-17(3)10-7-13-20(18,4)5/h6,8-9,11-12,14H,7,10,13H2,1-5H3,(H,21,22)/b9-6+,12-11+,15-8+,16-14+" 
+REA InChIKey         InChI                1.03  SHGAZHPCJJPHSC-YCNIQYBTSA-N                                                                                                                           
+REA SMILES_CANONICAL CACTVS               3.385 "CC1=C(\C=C\C(C)=C\C=C\C(C)=C\C(O)=O)C(C)(C)CCC1"                                                                                                     
+REA SMILES           CACTVS               3.385 "CC1=C(C=CC(C)=CC=CC(C)=CC(O)=O)C(C)(C)CCC1"                                                                                                          
+REA SMILES_CANONICAL "OpenEye OEToolkits" 1.7.6 "CC1=C(C(CCC1)(C)C)/C=C/C(=C/C=C/C(=C/C(=O)O)/C)/C"                                                                                                   
+REA SMILES           "OpenEye OEToolkits" 1.7.6 "CC1=C(C(CCC1)(C)C)C=CC(=CC=CC(=CC(=O)O)C)C"                                                                                                          
+# 
+loop_
+_pdbx_chem_comp_identifier.comp_id 
+_pdbx_chem_comp_identifier.type 
+_pdbx_chem_comp_identifier.program 
+_pdbx_chem_comp_identifier.program_version 
+_pdbx_chem_comp_identifier.identifier 
+REA "SYSTEMATIC NAME" ACDLabs              12.01 "retinoic acid"                                                                             
+REA "SYSTEMATIC NAME" "OpenEye OEToolkits" 1.7.6 "(2E,4E,6E,8E)-3,7-dimethyl-9-(2,6,6-trimethylcyclohexen-1-yl)nona-2,4,6,8-tetraenoic acid" 
+# 
+loop_
+_pdbx_chem_comp_audit.comp_id 
+_pdbx_chem_comp_audit.action_type 
+_pdbx_chem_comp_audit.date 
+_pdbx_chem_comp_audit.processing_site 
+REA "Create component"   1999-07-08 RCSB 
+REA "Modify descriptor"  2011-06-04 RCSB 
+REA "Other modification" 2016-10-18 RCSB 
+# 

test/rename-compound-test.cpp

@@ -20,9 +20,10 @@ int main(int argc, char* argv[])
 		if (argc == 3)
 			testdir = argv[2];
 
-		if (std::filesystem::exists(testdir / ".."/"data"/"components.cif"))
-			cif::addFileResource("components.cif", testdir / ".."/"data"/"components.cif");
+		if (std::filesystem::exists(testdir / ".."/"data"/"ccd-subset.cif"))
+			cif::addFileResource("components.cif", testdir / ".."/"data"/"ccd-subset.cif");
 
+		mmcif::CompoundFactory::instance().pushDictionary(testdir / "REA.cif");
 		mmcif::CompoundFactory::instance().pushDictionary(testdir / "RXA.cif");
 
 		mmcif::File f(testdir / ".."/"examples"/"1cbs.cif.gz");

tools/update-dictionary-script

@@ -0,0 +1,47 @@
+#!/bin/bash
+
+set -e
+
+if [ "$EUID" -ne 0 ]
+	then echo "Please run as root"
+	exit
+fi
+
+if [ -f /etc/libcifpp.conf ] ; then
+	. /etc/libcifpp.conf
+fi
+
+# check to see if we're supposed to run at all
+if [ "$update" != "true" ] ; then
+	exit
+fi
+
+# if cache directory doesn't exist, exit. 
+if ! [ -d /var/cache/libcifpp ]; then
+	exit
+fi
+
+fetch_dictionary () {
+	dict=$1
+	source=$2
+
+	wget -O${dict}.gz ${source}
+
+	# be careful not to nuke an existing dictionary file
+	# extract to a temporary file first
+
+	gunzip -c ${dict}.gz > ${dict}-tmp
+
+	# then move the extracted file to the final location
+
+	mv ${dict}-tmp ${dict}
+
+	# and clean up afterwards
+
+	rm ${dict}.gz
+}
+
+# fetch the dictionaries
+
+fetch_dictionary "/var/cache/libcifpp/mmcif_pdbx_v50.dic" "https://mmcif.wwpdb.org/dictionaries/ascii/mmcif_pdbx_v50.dic.gz"
+fetch_dictionary "/var/cache/libcifpp/components.cif" "ftp://ftp.wwpdb.org/pub/pdb/data/monomers/components.cif.gz"