Source Code for Module biana.BianaParser.taxonomyParser

  1  """ 
  2      BIANA: Biologic Interactions and Network Analysis 
  3      Copyright (C) 2009  Javier Garcia-Garcia, Emre Guney, Baldo Oliva 
  4   
  5      This program is free software: you can redistribute it and/or modify 
  6      it under the terms of the GNU General Public License as published by 
  7      the Free Software Foundation, either version 3 of the License, or 
  8      (at your option) any later version. 
  9   
 10      This program is distributed in the hope that it will be useful, 
 11      but WITHOUT ANY WARRANTY; without even the implied warranty of 
 12      MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the 
 13      GNU General Public License for more details. 
 14   
 15      You should have received a copy of the GNU General Public License 
 16      along with this program.  If not, see <http://www.gnu.org/licenses/>. 
 17   
 18  """ 
 19   
 20  from bianaParser import * 
 21  import sets 
 22   
 23   
 24 -class TaxonomyParser(BianaParser): 
 25      """ 
 26      Taxonomy Parser Class 
 27      """ 
 28   
 29      name = "taxonomy" 
 30      description = "This program fills up tables in database biana related with taxonomy ontology" 
 31      external_entity_definition = "A external entity represents a taxonomy of any type" 
 32      external_entity_relations = "" 
 33   
 34 -    def __init__(self): 
 35   
 36          # Start with the default values 
 37   
 38           BianaParser.__init__(self, default_db_description = "NCBI Species information", 
 39                                default_script_name = "taxonomyParser", 
 40                                default_script_description = "This program fills up tables in database piana related to external DB 'taxonomy'") 
 41           self.default_eE_attribute = "taxid" 
 42   
 43 -    def parse_database(self): 
 44          """ 
 45          Method that implements the specific operations of taxonomy parser 
 46          """ 
 47   
 48          self.biana_access.add_valid_external_entity_attribute_type( name = "TaxID_category", 
 49                                                                      data_type = "ENUM(\"class\",\"family\",\"forma\",\"genus\",\"infraclass\",\"infraorder\",\"kingdom\",\"no rank\",\"order\",\"parvorder\",\"phylum\",\"species\",\"species group\",\"species subgroup\",\"subclass\",\"subfamily\",\"subgenus\",\"subkingdom\",\"suborder\",\"subphylum\",\"subspecies\",\"subtribe\",\"superclass\",\"superfamily\",\"superkingdom\",\"superorder\",\"superphylum\",\"tribe\",\"varietas\")", 
 50                                                                      category = "eE descriptive attribute") 
 51   
 52           
 53          self.biana_access.add_valid_external_entity_attribute_type( name = "TaxID_name", 
 54                                                                      data_type = { "fields": [("value","varchar(255)"), 
 55                                                                                               ("taxid_name_type", "ENUM(\"acronym\",\"anamorph\",\"blast name\",\"common name\",\"equivalent name\",\"genbank acronym\",\"genbank anamorph\",\"genbank common name\",\"genbank synonym\",\"includes\",\"in-part\",\"misnomer\",\"misspelling\",\"scientific name\",\"synonym\",\"teleomorph\",\"authority\",\"unpublished name\")", True)] }, 
 56                                                                      category = "eE special attribute") 
 57   
 58   
 59          # IMPORTANT: As we have added new types and attributes that are not in the default BIANA distribution, we must execute the follwing command: 
 60          self.biana_access.refresh_database_information()                                                                         
 61                                                                       
 62          # Add the possibility to transfer taxonomy name and taxonomy category using taxID as a key 
 63          self.biana_access._add_transfer_attribute( externalDatabaseID = self.database.get_id(), # A single taxonomy element can have multiple names 
 64                                                     key_attribute = "taxID", 
 65                                                     transfer_attribute="TaxID_name" ) 
 66   
 67          self.biana_access._add_transfer_attribute( externalDatabaseID = self.database.get_id(),   # Category is stored in a different attribute, as a single taxonomy can have multiple names 
 68                                                     key_attribute = "taxID", 
 69                                                     transfer_attribute = "TaxID_category" ) 
 70   
 71          nodes_dmp_file = None 
 72          names_dmp_file = None 
 73   
 74          if os.path.isdir(self.input_file): 
 75              if( not self.input_file.endswith(os.sep) ): 
 76                  self.input_file += os.sep 
 77              nodes_dmp_file = os.path.dirname(self.input_file) + os.sep + "nodes.dmp" 
 78              names_dmp_file = os.path.dirname(self.input_file) + os.sep + "names.dmp" 
 79           
 80          ontology = Ontology( source_database = self.database, linkedAttribute="taxid", name="taxonomy", descriptionAttribute="TaxID_name", levelAttribute="TaxID_category" ) 
 81           
 82          specific_identifiers_and_parent = {} 
 83   
 84          names_fd = file(names_dmp_file,"r") 
 85   
 86          names_comment_dict = {} 
 87   
 88          print "Reading names" 
 89          for line in names_fd: 
 90   
 91              line_fields = line.split("|") 
 92   
 93              if line_fields: 
 94   
 95                  tax_id = line_fields[0].strip() 
 96                  tax_name = line_fields[1].strip().replace('"','\"').replace("\\","\\\\")[0:254] 
 97                  tax_comment = line_fields[3].strip().replace('"','\"').replace("\\","\\\\") 
 98   
 99                  if tax_id != "" and tax_name != "": 
100                          names_comment_dict.setdefault(tax_id, sets.Set()).add((tax_name, tax_comment)) 
101                      
102          names_fd.close() 
103   
104          print "Reading nodes" 
105          nodes_fd = file(nodes_dmp_file,"r") 
106          for line in nodes_fd: 
107   
108              line_fields = line.split("|") 
109              tax_id = line_fields[0].strip() 
110              parent_tax_id = line_fields[1].strip() 
111              type = line_fields[2].strip() 
112               
113   
114              externalEntity = ExternalEntity( source_database = self.database, type = "taxonomyElement" ) 
115   
116              externalEntity.add_attribute( ExternalEntityAttribute( attribute_identifier = "taxID", value = tax_id ) ) 
117              externalEntity.add_attribute( ExternalEntityAttribute( attribute_identifier = "TaxID_category", value = type ) ) 
118               
119              for (tax_name, tax_comment) in names_comment_dict[tax_id]: 
120                  externalEntity.add_attribute( ExternalEntityAttribute( attribute_identifier = "TaxID_name", 
121                                                                         value = tax_name, 
122                                                                         additional_fields = {"taxid_name_type": tax_comment} ) ) 
123   
124              self.biana_access.insert_new_external_entity( externalEntity ) 
125   
126              specific_identifiers_and_parent[tax_id] = (externalEntity.get_id(),parent_tax_id) 
127   
128          nodes_fd.close() 
129   
130          print "Done!" 
131   
132          # Set the ontology hierarch and insert elements to ontology 
133          for current_tax_ID in specific_identifiers_and_parent: 
134              ontology.add_element( ontologyElementID = specific_identifiers_and_parent[current_tax_ID][0], 
135                                    isA = [specific_identifiers_and_parent[specific_identifiers_and_parent[current_tax_ID][1]][0]] ) 
136               
137          print "Inserting ontology to database" 
138          self.biana_access.insert_new_external_entity( ontology ) 
139