bio-blastxmlparser 2.0.1 → 2.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7222df89b2f60ef4b027ea7ca766a30c04de567b
-  data.tar.gz: b8d7c84c85dd58e7794a62b83a73b17a04b60ce1
+  metadata.gz: 76df7cd1f6e1bc4f2b04fdf3f0fc830110f2e376
+  data.tar.gz: 68f44a797aa5357690e6c4a10a1fe241a7b8fe37
 SHA512:
-  metadata.gz: e9feee95e3063b0c6c9e9ac28c0f7389e4036130e51c107314216fe8e30f98342d2fbc5f1af0ef16f9c5a11be95aa97d86d16f5d9e2169eda2a54d2594c0dc84
-  data.tar.gz: 63971bd220b178e7ff0dbd7c50a4df6277b9dc8035f610f173603e2e304655610e65894fb3bde7e67c17963d13557b22c3c1d4a3e6dbadbf65c7f170ddbd12f5
+  metadata.gz: f424b8cfedf921840dbf2fee412c191f1c9951d289b80db047744a56883b171b50d4b1161e86415f0e873be4bd43935031f966d32d394efea9208e65b2529903
+  data.tar.gz: 6214606ac08afa7306503c78e969046ab2a54d9c45e9dc42c597d8c033db2095fc5a25f15f56090a0337bd1049081699a81ceefa90d2acaf17cb782340074fc6

data/README.md

@@ -107,7 +107,7 @@ Print result fields of iterations containing 'lcl', using a regex
   blastxmlparser --filter 'iter.query_id=~/lcl/' test/data/nt_example_blastn.m7
 ```
 
-prints a tab delimited
+prints a (default) tab delimited to stdout
 
 ```sh
   1       1       lcl|1_0 lcl|I_74685     1       5.82208e-34
@@ -228,6 +228,59 @@ Likewise, using the RDF template
   :evalue      8.1089e-12 .
 ```
 
+### Metadata
+
+Templates can also print data as a header of the JSON/YAML/RDF output. For this
+use the '=' prefix with HEADER, BODY, FOOTER keywords in the template. A small example
+can be
+
+```Javascript
+=HEADER
+<% require 'json' %>
+[
+  { "HEADER": {
+    "options":  <%= options.to_h.to_json %>,
+    "files":    <%= ARGV %>,
+    "version":  "<%= BLASTXML_VERSION %>"
+  },
+=BODY
+  { "<%= hit.parent.query_def %>": {
+    "num":      <%= hit.hit_num %>,
+    "id":       "<%= hit.hit_id %>",
+    "len":      <%= hit.len %>,
+    "E-value":  <%= hsp.evalue %>,
+  },
+=FOOTER
+]
+```
+
+may generate something like
+
+```Javascript
+[
+  { "HEADER": {
+    "options":  {"template":"template/blast2json2.erb","filter":"hsp.evalue>0.01"},
+    "files":    ["test/data/nt_example_blastn.m7"],
+    "version":  "2.0.2-pre1"
+  },
+  { "I_1 [477 - 884] ": {
+    "num":       41,
+    "id":        "lcl|X_42251",
+    "len":       153,
+    "E-value":   0.0247015,
+  },
+  { "I_1 [477 - 884] ": {
+    "num":       43,
+    "id":        "lcl|V_105720",
+    "len":       180,
+    "E-value":   0.0247015,
+  },
+]
+```
+
+Note that the template is not smart enough to remove the final comma
+from the last BODY element. To make it valid JSON that needs to be removed.
+
 ## Additional options
 
 To use the high-mem version of the parser (slightly faster on single core) use

data/VERSION

@@ -1 +1 @@
-2.0.1
+2.0.2

data/bin/blastxmlparser

@@ -1,6 +1,6 @@
 #!/usr/bin/env ruby
 #
-# BioRuby bio-blastxmlparser Plugin
+# blastxmlparser
 # Author:: Pjotr Prins
 # License:: MIT License
 #
@@ -75,7 +75,6 @@ opts = OptionParser.new do |o|
   end
   
   o.on("-t erb","--template erb",String, "Use ERB template for output") do |s|
-    require 'erb'
     options.template = s
   end
   
@@ -91,6 +90,7 @@ opts = OptionParser.new do |o|
   
   o.on("-q", "--quiet", "Run quietly") do |q|
     Bio::Log::CLI.trace('error')
+    options.quiet = true
   end
   
   o.on("-v", "--verbose", "Run verbosely") do |v|
@@ -118,13 +118,18 @@ begin
     end
   end
 
+  # Prepare the ERB template
   if options.template
     include BioRdf
-    fn = options.template
-    raise "No template #{fn}!" if not File.exist?(fn)
-    template = ERB.new(File.read(fn))
+    require 'bio/writers/template'
+    template = Bio::Template.new(options.template)
   end
 
+  $stderr.print options,"\n" if not options.quiet
+
+  raise "No input file(s) defined" if ARGV.size == 0
+
+  
   ARGV.each do | fn |
     logger.info("XML parsing #{fn}")
     parser_type = options.parser
@@ -145,6 +150,7 @@ begin
     chunks = []
     chunks_count = 0
     NUM_CHUNKS=10_000
+    print template.header(binding) if template
 
     process = lambda { |iter2,i|  # Process one BLAST iter block
       if parser_type == :nosplit
@@ -167,7 +173,7 @@ begin
           if do_print
             line_count += 1
             if template
-              res << template.result(binding)
+              res << template.body(binding)
             elsif options.output_fasta
               res << ">"+hit.accession+' '+iter.iter_num.to_s+'|'+iter.query_id+' '+hit.hit_id+' '+hit.hit_def+"\n"
               res << hsp.qseq+"\n"
@@ -221,6 +227,7 @@ begin
             process.call(iter,i)
           }
     end
+    print template.footer(binding) if template
   end
 rescue OptionParser::InvalidOption => e
   $stderr.print e.message

data/bio-blastxmlparser.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "bio-blastxmlparser"
-  s.version = "2.0.1"
+  s.version = "2.0.2"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Pjotr Prins"]
-  s.date = "2014-09-07"
+  s.date = "2014-11-07"
   s.description = "Fast big data BLAST XML parser and library; this libxml2 based version is 50x faster than BioRuby and comes with a nice CLI"
   s.email = "pjotr.public01@thebird.nl"
   s.executables = ["blastxmlparser"]
@@ -34,6 +34,7 @@ Gem::Specification.new do |s|
     "lib/bio/db/blast/xmliterator.rb",
     "lib/bio/db/blast/xmlsplitter.rb",
     "lib/bio/writers/rdf.rb",
+    "lib/bio/writers/template.rb",
     "sample/bioruby.rb",
     "sample/blastxmlparserdemo.rb",
     "sample/libxml_sax.rb",
@@ -43,6 +44,7 @@ Gem::Specification.new do |s|
     "spec/bio-blastxmlparser_spec.rb",
     "spec/spec_helper.rb",
     "template/blast2json.erb",
+    "template/blast2json2.erb",
     "template/blast2rdf-minimal.erb",
     "template/blast2rdf.erb",
     "test/data/aa_example.fasta",

data/lib/bio/writers/template.rb

@@ -0,0 +1,75 @@
+require 'erb'
+
+module Bio
+
+  class Template
+    
+    def initialize fn
+      raise "Can not find template #{fn}!" if not File.exist?(fn)
+      parse(File.read(fn))
+    end
+
+    def parse buf
+      header = []
+      body = []
+      footer = []
+      where = :header
+      buf.split("\n").each do | line |
+        case where
+          when :header
+            next if line =~ /=HEADER/
+            if line =~ /=BODY/
+              body = []
+              where = :body
+              next
+            end
+            header << line
+          when :body
+            if line =~ /=FOOTER/
+              footer = []
+              where = :footer
+              next
+            end
+            body << line
+          else
+            footer << line
+        end
+      end
+      if body == []
+        body = header
+        header = []
+      end
+      @erb_header = ERB.new(header.join("\n")) if header.size
+      @erb_body   = ERB.new(body.join("\n")) if body.size
+      @erb_footer = ERB.new(footer.join("\n")) if footer.size
+    end
+    
+    def result env
+      @erb.result(env)
+    end
+
+    def header env
+      if @erb_header
+        @erb_header.result(env)
+      else
+        ""
+      end
+    end
+
+    def body env
+      if @erb_body
+        @erb_body.result(env)
+      else
+        ""
+      end
+    end
+
+    def footer env
+      if @erb_footer
+        @erb_footer.result(env)
+      else
+        ""
+      end
+    end
+  end
+end

data/template/blast2json2.erb

@@ -0,0 +1,28 @@
+=HEADER
+<% require 'json' %>
+[
+  { "HEADER": {
+    "options":  <%= options.to_h.to_json %>,
+    "files":    <%= ARGV %>,
+    "version":  "<%= BLASTXML_VERSION %>"
+  },
+
+=BODY
+  { "<%= hit.parent.query_def %>": {
+    "query_id":  "<%= hit.parent.query_id %>",
+    "num":       <%= hit.hit_num %>,
+    "accession": "<%= hit.accession %>",
+    "id":        "<%= hit.hit_id %>",
+    "len":       <%= hit.len %>,
+    "E-value":   <%= hsp.evalue %>,
+    "identity":  <%= hsp.identity %>,
+    "align_len": <%= hsp.align_len %>,
+    "bitscore":  <%= hsp.bit_score %>,
+    "qseq":      "<%= hsp.qseq %>",
+    "midline":   "<%= hsp.midline %>", 
+    "hseq":      "<%= hsp.hseq %>",
+  },
+
+=FOOTER
+]
+

data/template/blast2rdf-minimal.erb

@@ -4,11 +4,11 @@
 %>
 :<%= blastid %>  :query  :<%= id %>
 :<%= id %>
-  :query_def   "<%= hit.parent.query_def %>",
-  :num         <%= hit.hit_num %>,
-  :accession   "<%= hit.accession %>",
-  :len         <%= hit.len %>,
-  :identity    <%= hsp.identity %>,
-  :align_len   <%= hsp.align_len %>,
-  :bitscore    <%= hsp.bit_score %>,
-  :evalue   <%= hsp.evalue %> .
+  :query_def   "<%= hit.parent.query_def %>";
+  :num         <%= hit.hit_num %>;
+  :accession   "<%= hit.accession %>";
+  :len         <%= hit.len %>;
+  :identity    <%= hsp.identity %>;
+  :align_len   <%= hsp.align_len %>;
+  :bitscore    <%= hsp.bit_score %>;
+  :evalue      <%= hsp.evalue %> .

data/template/blast2rdf.erb

@@ -4,17 +4,17 @@
 %>
 :<%= blastid %>  :query  :<%= id %>
 :<%= id %>
-  :query_id    "<%= hit.parent.query_id %>",
-  :query_def   "<%= hit.parent.query_def %>",
-  :num         <%= hit.hit_num %>,
-  :accession   "<%= hit.accession %>",
-  :id          "<%= hit.hit_id %>",
-  :len         <%= hit.len %>,
-  :identity    <%= hsp.identity %>,
-  :align_len   <%= hsp.align_len %>,
-  :bitscore    <%= hsp.bit_score %>,
-  :qseq        "<%= hsp.qseq %>",
+  :query_id    "<%= hit.parent.query_id %>";
+  :query_def   "<%= hit.parent.query_def %>";
+  :num         <%= hit.hit_num %>;
+  :accession   "<%= hit.accession %>";
+  :id          "<%= hit.hit_id %>";
+  :len         <%= hit.len %>;
+  :identity    <%= hsp.identity %>;
+  :align_len   <%= hsp.align_len %>;
+  :bitscore    <%= hsp.bit_score %>;
+  :qseq        "<%= hsp.qseq %>";
   :midline     "<%= hsp.midline %>", 
-  :hseq        "<%= hsp.hseq %>",
+  :hseq        "<%= hsp.hseq %>";
   :evalue      <%= hsp.evalue %> .
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bio-blastxmlparser
 version: !ruby/object:Gem::Version
-  version: 2.0.1
+  version: 2.0.2
 platform: ruby
 authors:
 - Pjotr Prins
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-09-07 00:00:00.000000000 Z
+date: 2014-11-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bio-logger
@@ -134,6 +134,7 @@ files:
 - lib/bio/db/blast/xmliterator.rb
 - lib/bio/db/blast/xmlsplitter.rb
 - lib/bio/writers/rdf.rb
+- lib/bio/writers/template.rb
 - sample/bioruby.rb
 - sample/blastxmlparserdemo.rb
 - sample/libxml_sax.rb
@@ -143,6 +144,7 @@ files:
 - spec/bio-blastxmlparser_spec.rb
 - spec/spec_helper.rb
 - template/blast2json.erb
+- template/blast2json2.erb
 - template/blast2rdf-minimal.erb
 - template/blast2rdf.erb
 - test/data/aa_example.fasta