vardb 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 73316c44cd437fe2eef6892b3e00097d28ea23ef
+  data.tar.gz: 28be909d305e54450226c784a0c990394477ef1e
+SHA512:
+  metadata.gz: 0a4a037984d78107eff470db89cf7d6007d188a85c33c65b046f6147345cd313ec668489b56fe42ada8d9203e7458037628d8769b1c285934f120a1e575f831c
+  data.tar.gz: a1520fb5d119c45368db4340e4d317369e1d3b0f5aa55eb0eddf2d919f013f304eb8d22bf00c3b8dcb36bc1013211801fa5a5d77fedf0c6dea3fa894fe467228

data/lib/vardb/database_populator.rb

@@ -0,0 +1,111 @@
+require_relative 'xls_parser'
+require 'pg'
+
+module Populator
+  include XlsParser
+
+  def populate_matrix
+
+	host = ConfigData.get_connection
+
+	conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+	
+	#Matrix File Command Preparation
+	conn.prepare('load_snps', 'INSERT INTO snps (id, locus, annotation_id) values ($1, $2, $3)')
+	conn.prepare('load_annos', 'INSERT INTO annotations (id, cds, transcript, transcript_id, info, orientation, cds_locus, codon_pos, codon, peptide, amino_a, syn ) values ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12)')
+	conn.prepare('load_samples_snps', 'INSERT INTO samples_snps (sample_id, snp_id) values ($1, $2)')
+	conn.prepare('load_samples', 'INSERT INTO samples (id, name) values ($1, $2)')
+
+	#Matrix File Load-ins
+	text=File.open(ConfigData.get_matrix).read
+
+	linenum = 1
+	sample_number = 1
+
+	snps = []
+	anno_tabs = []
+	anno_vals = []
+
+	text.each_line do |line|
+		(header, line_data) = line.split(' ', 2)
+		if (header == '#snp_pos')
+			puts "populating snps table..."
+			snps = line_data.split("\t")
+			snp_counter = 1
+			snps.each do |locus| 
+				conn.exec_prepared('load_snps', [snp_counter, locus, snp_counter])
+				snp_counter += 1
+			end	
+		elsif (header == '#annotation')
+			puts "populating annotations table..."
+			anno_tabs = line_data.split("\t")
+			anno_tabs.each { |tab| anno_vals << tab.split(',', 11) }
+			anno_counter = 1
+        	anno_vals.each do |anno|
+    	    	anno.insert(0, anno_counter)
+    	    	if anno[1].match('intergenic')
+    		    	conn.exec_prepared('load_annos', [ anno[0], 0, 0, 0,  anno[1], 0, 0, 0, 0, 0, 0, 0 ])
+            	else 
+                	conn.exec_prepared('load_annos', [ anno[0], anno[1], anno[2], anno[3], anno[4], anno[5], anno[6], anno[7], anno[8], anno[9], anno[10], anno[11] ])
+            	end
+            	anno_counter += 1  
+        	end
+		else
+			if sample_number == 1 then
+		    	puts "loading reference..."
+	    	else
+		    	puts "loading in sample #{sample_number - 1}..."
+			end  
+			conn.exec_prepared('load_samples', [sample_number, header])
+			line_data.split("\t").each_with_index do |n, i|
+				if (n == '1')
+					conn.exec_prepared('load_samples_snps', [sample_number, i])
+				end
+  	    	end
+  	    	sample_number += 1
+		end
+    	linenum += 1
+	end
+  end	
+
+  def populate_metadata
+
+  	host = ConfigData.get_connection
+
+	conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+
+	#Excel Spreadsheet Command Preparaton
+	metadata_fields = XlsParser.load_meta_fields(ConfigData.get_metadata)
+
+	metadata_fields_string = "id "
+
+	metadata_fields.each do |item|
+		metadata_fields_string << item
+	end
+
+	metadata_values_string = "$1 "
+
+	metadata_fields.length.times do |i|
+    	metadata_values_string << ", $#{i+2}"
+	end
+
+	conn.prepare('load_metadata', "INSERT INTO sample_metadata (#{metadata_fields_string}) values (#{metadata_values_string})")
+
+	#Excel Spreadsheet Load-ins
+	s = Roo::Excel.new(ConfigData.get_metadata)
+	s.default_sheet = s.sheets.first
+
+	row = 2
+
+	puts "populating sample metadata..."
+
+	until s.cell(row, 1).nil?
+		row_contents = ["#{row-1}"]
+		metadata_fields.length.times do |i|
+			row_contents << "#{s.cell(row, i)}"
+		end
+		conn.exec_prepared('load_metadata', row_contents)
+		row += 1
+	end
+  end
+end  

data/lib/vardb/snp_db_build.rb

@@ -0,0 +1,41 @@
+require_relative 'xls_parser'
+require 'pg'
+
+module Builder
+	include XlsParser
+	def format_matrix
+		host = ConfigData.get_connection
+
+		conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+
+		puts "formatting annotations table..."
+		conn.exec("CREATE TABLE annotations (id numeric(11) PRIMARY KEY, cds varchar(128), transcript varchar(128), transcript_id varchar(128), info text, orientation varchar(128), cds_locus varchar(128), codon_pos varchar(128), codon varchar(128), peptide varchar(128), amino_a varchar(128), syn varchar(128))")
+
+		puts "formatting snps table..."
+		conn.exec("CREATE TABLE snps (id numeric(11) PRIMARY KEY, locus numeric(11), annotation_id numeric(11))")
+
+		puts "formatting samples table..."
+		conn.exec("CREATE TABLE samples (id numeric(11) PRIMARY KEY, name varchar(128))")
+
+		puts "formatting samples_snps join table..."
+		conn.exec("CREATE TABLE samples_snps (sample_id numeric(11), snp_id numeric(11))")
+	end
+
+	def format_metadata
+		host = ConfigData.get_connection
+
+		conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+
+		metadata_fields = XlsParser.load_meta_fields(ConfigData.get_metadata)
+
+		metadata_field_names = ""
+
+		metadata_fields.each do |name|
+			name << " varchar(128)"
+			metadata_field_names << name
+		end
+
+		puts "formatting sample metadata table..."
+		conn.exec("CREATE TABLE sample_metadata (id numeric (11) PRIMARY KEY#{metadata_field_names})")
+	end
+end

data/lib/vardb/snpscript_configdata.rb

@@ -0,0 +1,42 @@
+module ConfigData
+
+  @@host = {}
+  @@metadata_file = ''
+  @@matrix_file = ''
+	
+	#host connection
+  def set_connection(connection_hash)
+    @@host = { 
+      :host => "#{connection_hash[:host]}", 
+  	  :port => "#{connection_hash[:port]}", 
+  	  :dbname => "#{connection_hash[:dbname]}", 
+  	  :user => "#{connection_hash[:user]}", 
+  	  :password => "#{connection_hash[:password]}",  
+  	}
+    puts "connection details: #{@@host}"
+  end
+
+  def self.get_connection
+    @@host
+  end
+
+  #metadata file
+  def set_metadata(file)
+  	@@metadata_file = file
+    puts "metadata file set to: #{@@metadata_file}"
+  end
+
+  def self.get_metadata
+    @@metadata_file
+  end
+
+  #matrix file
+  def set_matrix(file)
+  	@@matrix_file = file
+    puts "matrix file set to: #{@@matrix_file}"
+  end
+
+  def self.get_matrix
+    @@matrix_file
+  end
+end

data/lib/vardb/xls_parser.rb

@@ -0,0 +1,34 @@
+require 'roo'
+
+module XlsParser
+    ## A few rules about Excel files:
+    ## 1. .xls only, this can't accept .xlsx
+    ## 2. There should be no empty cells in the header row of a sheet (first row)
+    ## 3. For good form, there shouldn't be any subsequent rows that are 
+    ##    longer than the header row
+    ## 4. There should not be any duplicate cell names in the header row	
+	def XlsParser.load_meta_fields(file)
+		s = Roo::Excel.new(file)
+		s.default_sheet = s.sheets.first
+
+		columns = 1
+		until s.cell(1, columns).nil?
+			columns += 1
+		end
+
+		counter = 1
+
+		metadata_fields = []
+
+		columns.times do |counter|
+		    if s.cell(1, counter).nil?
+		    	metadata_fields << ", empty#{counter}"
+		    else 
+  			    metadata_fields << ", #{s.cell(1, counter).gsub(/\s+/, "").gsub("-","").gsub("(","").gsub(")","").gsub(".","").gsub("/","")}"
+  			end
+  			counter += 1
+		end
+
+		return metadata_fields
+	end
+end

data/lib/vardb.rb

@@ -0,0 +1,29 @@
+require 'vardb/snp_db_build'
+require 'vardb/database_populator'
+require 'vardb/snpscript_configdata'
+
+class Vardb
+	include Builder
+	include Populator
+	include ConfigData
+
+	#def self.set_connection(connection_hash)
+		#ConfigData.set_connection(connection_hash)
+	#end
+
+	#def self.set_metadata(file)
+		#ConfigData.set_metadata
+	#end
+
+	#def self.set_matrix(file)
+		#ConfigData.set_matrix
+	#end
+
+	#def self.format
+		#Builder.format_database
+	#end
+
+	#def self.populate
+		#Populator.populate_database
+	#end
+end

metadata

@@ -0,0 +1,78 @@
+--- !ruby/object:Gem::Specification
+name: vardb
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Peter McCaffrey
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2013-12-09 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: pg
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.17.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.17.0
+- !ruby/object:Gem::Dependency
+  name: roo
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.13.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.13.0
+description: This gem builds PostgreSQL databases from .matrix files and metadata
+  spreadsheets
+email:
+- peter@accetia.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/vardb.rb
+- lib/vardb/snp_db_build.rb
+- lib/vardb/database_populator.rb
+- lib/vardb/xls_parser.rb
+- lib/vardb/snpscript_configdata.rb
+homepage: ''
+licenses:
+- ''
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.0.3
+signing_key: 
+specification_version: 4
+summary: Variant database builder
+test_files: []