vardb 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 73316c44cd437fe2eef6892b3e00097d28ea23ef
-  data.tar.gz: 28be909d305e54450226c784a0c990394477ef1e
+  metadata.gz: a6564e63ebd0a500e3440e7ca7137f2c48f16423
+  data.tar.gz: 01681c3020ba950bbe8d132062a5c2902efa4822
 SHA512:
-  metadata.gz: 0a4a037984d78107eff470db89cf7d6007d188a85c33c65b046f6147345cd313ec668489b56fe42ada8d9203e7458037628d8769b1c285934f120a1e575f831c
-  data.tar.gz: a1520fb5d119c45368db4340e4d317369e1d3b0f5aa55eb0eddf2d919f013f304eb8d22bf00c3b8dcb36bc1013211801fa5a5d77fedf0c6dea3fa894fe467228
+  metadata.gz: 2a6244867da744515526f34ddf2151688fbde1443b70c56547cb6abd06965db6d05f97ccb010d2b0f1006804b7367b1ad35f14ee7c3af3e1c0ed043373e7683d
+  data.tar.gz: 8e0e88c954358326f2f8cfda36d5c509c043170b003d492b49c42161521b40773e2eff217b4c32df05a89735a98d22f643e5e05eac4a1111334582468f5013dd

data/lib/vardb.rb

@@ -6,24 +6,4 @@ 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

data/lib/vardb/database_populator.rb

@@ -4,17 +4,27 @@ require 'pg'
 module Populator
   include XlsParser
 
-  def populate_matrix
+  def populate_matrix(type)
 
 	host = ConfigData.get_connection
 
-	conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+	if type == 'pg'
+		#PG 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 PG 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)')
+
+	elsif type == 'sqlite'
+		#SQLite Connection
+		db = SQLite3::Database.new "sqlite_db/#{host[:dbname]}.db"
+
+	end
+
+
 
 	#Matrix File Load-ins
 	text=File.open(ConfigData.get_matrix).read
@@ -32,8 +42,12 @@ module Populator
 			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])
+			snps.each do |locus|
+				if type == 'pg' 
+					conn.exec_prepared('load_snps', [snp_counter, locus, snp_counter])
+				elsif type == 'sqlite'
+					db.execute("INSERT INTO snps (id, locus, annotation_id) VALUES (?,?,?)", [snp_counter, locus, snp_counter])
+				end
 				snp_counter += 1
 			end	
 		elsif (header == '#annotation')
@@ -44,9 +58,17 @@ module Populator
         	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 ])
+    	    		if type == 'pg'
+    		    		conn.exec_prepared('load_annos', [ anno[0], 0, 0, 0,  anno[1], 0, 0, 0, 0, 0, 0, 0 ])
+    		    	elsif type == 'sqlite'
+    		    		db.execute("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)", [ anno[0], 0, 0, 0,  anno[1], 0, 0, 0, 0, 0, 0, 0 ])
+    		    	end
             	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] ])
+            		if type == 'pg'
+                		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] ])
+                	elsif type == 'sqlite'
+                		db.execute("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)", [ 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	
             	end
             	anno_counter += 1  
         	end
@@ -55,11 +77,19 @@ module Populator
 		    	puts "loading reference..."
 	    	else
 		    	puts "loading in sample #{sample_number - 1}..."
-			end  
-			conn.exec_prepared('load_samples', [sample_number, header])
+			end
+			if type == 'pg'
+				conn.exec_prepared('load_samples', [sample_number, header])
+			elsif type == 'sqlite'
+				db.execute("INSERT INTO samples (id, name) VALUES (?,?)", [sample_number, header])
+			end
 			line_data.split("\t").each_with_index do |n, i|
 				if (n == '1')
-					conn.exec_prepared('load_samples_snps', [sample_number, i])
+					if type == 'pg'
+						conn.exec_prepared('load_samples_snps', [sample_number, i])
+					elsif type == 'sqlite'
+						db.execute("INSERT INTO samples_snps (sample_id, snp_id) VALUES (?,?)", [sample_number, i])
+					end
 				end
   	    	end
   	    	sample_number += 1
@@ -68,12 +98,10 @@ module Populator
 	end
   end	
 
-  def populate_metadata
+  def populate_metadata(type)
 
   	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)
 
@@ -89,7 +117,14 @@ module Populator
     	metadata_values_string << ", $#{i+2}"
 	end
 
-	conn.prepare('load_metadata', "INSERT INTO sample_metadata (#{metadata_fields_string}) values (#{metadata_values_string})")
+	if type == 'pg'
+		#PG Connection
+		conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+		conn.prepare('load_metadata', "INSERT INTO sample_metadata (#{metadata_fields_string}) values (#{metadata_values_string})")
+	elsif type == 'sqlite'
+		#SQLite Connection
+		db = SQLite3::Database.new "sqlite_db/#{host[:dbname]}.db"
+	end
 
 	#Excel Spreadsheet Load-ins
 	s = Roo::Excel.new(ConfigData.get_metadata)
@@ -104,7 +139,11 @@ module Populator
 		metadata_fields.length.times do |i|
 			row_contents << "#{s.cell(row, i)}"
 		end
-		conn.exec_prepared('load_metadata', row_contents)
+		if type == 'pg'
+			conn.exec_prepared('load_metadata', row_contents)
+		elsif type == 'sqlite'
+			db.execute("INSERT INTO sample_metadata (#{metadata_fields_string}) VALUES (#{metadata_values_string})", row_contents)
+		end
 		row += 1
 	end
   end

data/lib/vardb/snp_db_build.rb

@@ -1,31 +1,81 @@
 require_relative 'xls_parser'
 require 'pg'
+require 'sqlite3'
 
 module Builder
 	include XlsParser
-	def format_matrix
+	def format_matrix(type)
+		
 		host = ConfigData.get_connection
 
-		conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+		if type == 'pg'
+			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 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 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 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))")
+			puts "formatting samples_snps join table..."
+			conn.exec("CREATE TABLE samples_snps (sample_id numeric(11), snp_id numeric(11))")
+
+		elsif type == 'sqlite'
+			db = SQLite3::Database.new "sqlite_db/#{host[:dbname]}.db"
+
+			puts "formatting annotations table..."
+			db.execute <<-SQL
+				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)
+				);
+				SQL
+
+			puts "formatting snps table..."
+			db.execute <<-SQL
+				create table snps (
+				id numeric(11) PRIMARY KEY,
+				locus numeric(11),
+				annotation_id numeric(11)
+				);
+				SQL
+
+			puts "formatting samples table..."
+			db.execute <<-SQL
+				create table samples (
+				id numeric(11) PRIMARY KEY,
+				name varchar(128)
+				);
+				SQL
+
+			puts "formatting sample_snps join table..."
+			db.execute <<-SQL
+				create table samples_snps (
+				sample_id numeric(11),
+				snp_id numeric(11)
+				);
+				SQL
+
+		end 
 	end
 
-	def format_metadata
+	def format_metadata(type)
 		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 = ""
@@ -35,7 +85,21 @@ module Builder
 			metadata_field_names << name
 		end
 
-		puts "formatting sample metadata table..."
-		conn.exec("CREATE TABLE sample_metadata (id numeric (11) PRIMARY KEY#{metadata_field_names})")
+		if type == 'pg'
+			conn = PGconn.connect(:host => host[:host], :port => host[:port], :dbname => host[:dbname], :user => host[:user], :password => host[:password])
+
+			puts "formatting sample metadata table..."
+			conn.exec("CREATE TABLE sample_metadata (id numeric (11) PRIMARY KEY#{metadata_field_names})")
+		
+		elsif type == 'sqlite'
+			db = SQLite3::Database.new "sqlite_db/#{host[:dbname]}.db"
+
+			puts "formatting sample metadata table..."
+			db.execute <<-SQL
+				create table sample_metadata (
+					id numeric(11) PRIMARY KEY#{metadata_field_names}
+				);
+				SQL
+		end
 	end
 end

data/lib/vardb/xls_parser.rb

@@ -24,7 +24,7 @@ module XlsParser
 		    if s.cell(1, counter).nil?
 		    	metadata_fields << ", empty#{counter}"
 		    else 
-  			    metadata_fields << ", #{s.cell(1, counter).gsub(/\s+/, "").gsub("-","").gsub("(","").gsub(")","").gsub(".","").gsub("/","")}"
+  			    metadata_fields << ", #{s.cell(1, counter).to_s.gsub(/\s+/, "").gsub("-","").gsub("(","").gsub(")","").gsub(".","").gsub("/","")}"
   			end
   			counter += 1
 		end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: vardb
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Peter McCaffrey
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-12-09 00:00:00.000000000 Z
+date: 2014-02-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pg
@@ -38,6 +38,20 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: 1.13.0
+- !ruby/object:Gem::Dependency
+  name: sqlite3
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.3.7
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.3.7
 description: This gem builds PostgreSQL databases from .matrix files and metadata
   spreadsheets
 email:
@@ -47,10 +61,10 @@ 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/snp_db_build.rb
 - lib/vardb/snpscript_configdata.rb
+- lib/vardb/xls_parser.rb
 homepage: ''
 licenses:
 - ''
@@ -71,7 +85,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.3
+rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
 summary: Variant database builder