bioNMF - Webservices API

#!/usr/bin/env ruby # ###################################################### # # Example of client for bioNMF Webservices. # # Usage: # $0 # # Output files are saved in directory 'output_'. # # NOTE: # Consider downloading and using the following files for testing: # http://bionmf.dacya.ucm.es/data/withlabels.txt # http://bionmf.dacya.ucm.es/data/without_labels.txt # # This test only works with ASCII-text files with non-numeric labels # (or with no labels at all). # ###################################################### require 'soap/wsdlDriver' require 'fileutils' require 'base64' require 'yaml' ############################# # HTML headers for API file. def html_headers return < bioNMF - Webservices API END end ############################# # Help message. puts <> This programs uploads an input matrix and performs a Biclustering by using a Non-negative Matrix Factorization (NMF). Usage: #{ $0 } Output files are saved in directory 'output_'. Please consider downloading and using these matrices for testing: http://bionmf.dacya.ucm.es/data/withlabels.txt http://bionmf.dacya.ucm.es/data/without_labels.txt WARNING: This test works *ONLY* with *ASCII-text files with non-numeric labels (or with no labels at all)* like the ones above. To use it with binary files, please edit this file and set to 'true' the second argument of the function call to 'upload_matrix()'. To use numeric row or column labels, please set to true the third and fourth arguments (respectively) in the same function call. -------- END ############# # Input matrix if ARGV[0] == nil puts "Please specify an input filename." exit 1 end matrix_file = File.expand_path( ARGV[0] ) matrix_bfilename = File.basename( ARGV[0] ) # Output directory OUTPUT_DIR = File.expand_path("output_#{ matrix_bfilename }") ############# begin ######################## # Read input matrixs ######################## puts "Reading #{ matrix_file }" input_matrix = nil File.open(matrix_file, 'r') { |f| input_matrix = Base64.encode64( f.read ) # Encodes to base-64 before uploading. } ######################## # Connect to server ######################## WSDL_URL="http://bionmf.dacya.ucm.es/WebService/BioNMFWS.wsdl" #WSDL_URL="file://#{ENV['PWD']}/../wsdl/BioNMFWS.wsdl" # For internal testing. puts "Connecting to #{ WSDL_URL }..." driver = SOAP::WSDLDriverFactory.new(WSDL_URL).create_rpc_driver ######################## # Downloads API documentation ######################## # puts "Writting API documentation to 'API.html'" # File.open( 'API.html', 'w' ) do |f| # # First, writes headers. # f.write(html_headers) # # # API (HTML Table) # f.write driver.documentation # # # Finally, ending headers. # f.write "\n\n" #end ######################## # Upload input matrix. # # In this example, we force to process a text file with no numeric row/column labels. # Preprocessing is set to "Exponential scaling". # # NOTE: PLEASE REMEMBER TO ENCODE YOUR INPUT DATA TO BASE-64 BEFORE UPLOADING. # You can use the Base64::encode64 method. ######################## print "Uploading input matrix " $stdout.flush matrix_id = driver.upload_matrix( # Returns a matrix ID input_matrix, # Base64-encoded matrix false, # Forces to process ASCII-text data. false, # Matrix does NOT have numeric column labels. false, # Matrix does NOT have numeric row labels. false, # Do NOT transpose before preprocessing. "No", # No normalization method. "ExpScal", # Use exponential scaling to make data positive. matrix_bfilename ) # Internal suggested filename (optional). print "(returned matrix ID: #{ matrix_id }).\nWaiting ..." $stdout.flush # Waits until it is done. while !driver.done(matrix_id) sleep(5) print '.' $stdout.flush end puts " done.\n\n" # Exits if error. if driver.error(matrix_id) raise "\nError preprocessing input matrix.\nJob's output log:\n\n" \ + driver.messages(matrix_id).join("\n") end ######################## # Starts the analysis. # # Executes a "Biclustering" analysis. # This method groups genes and samples based on local features generating sets of samples and genes # that are locally related. The result is a set of K biclusters (sub-matrices) encoding modular # patterns, where K is the best factorization rank within a given input range. Each bicluster matrix # contains the set of genes that are highly associated to a local pattern, and samples sorted by its # importance in this pattern. # # It this example, it will search the best factorization rank within the range K=[2...5], by testing # 40 times each candidate value. # ######################## puts "\n======\n\nStarting analysis..." job_id = driver.biclustering( # Returns a job ID. matrix_id, # Matrix ID returned by the upload_matrix() method "nsNMF", # NMF algorithm 2, # Starting factorization rank 5, # Ending factorization rank 40, # Number of runs per factorization rank 2000, # Number of iterations per run 40, # Stopping threshold 0.5, # Smoothness for the nsNMF algorithm (ignored by other NMF algorithms) true, # Generate extra info files "biclustering_#{ matrix_bfilename }") # Suggested job name (optional). # Waits until finished. print "Job submited (returned job ID: #{ job_id }).\nWaiting ..." $stdout.flush while !driver.done(job_id) sleep(5) print '.' $stdout.flush end puts " done.\n\n" ######################## # Show output messages ######################## puts "Analysis parameters:\n#{ driver.info(job_id) }\n\n============ Job's output ============\n" + \ driver.messages(job_id).join("\n") if driver.error(job_id): raise "\nError in analysis step." end ######################## # Downloads results. ######################## puts "\n\n=======\n\nDownloading output files to '#{ OUTPUT_DIR }'" # Creates directory to save output files. FileUtils.remove_dir(OUTPUT_DIR, true) FileUtils.mkdir( OUTPUT_DIR ) # Fetches array with results' ID. results = driver.results(job_id) # Prefix for filenames. out = File.join(OUTPUT_DIR, job_id) ################################ # Downloads output files listed # in the results[] array: ################################ # Matrices W and H: File.open(out + '_W.txt','w') do |f,m| m = driver.result( results[0] ) # Downloads output matrix m = Base64.decode64( m ) # Decodes matrix from base 64. f.write m # Writes decoded matrix to file. end File.open(out + '_H.txt','w') { |f| f.write Base64.decode64( driver.result(results[1]) ) } # Next files in results[] are all the biclusters found. # # Downloads results[2..-1] (excludes the last position) r = 2 len = results.length - 1 b = 1 # Counter for biclusters. while r < len filename = out + '_bicluster_' + b.to_s File.open(filename + '.txt','w') { |f| f.write Base64.decode64( driver.result(results[r]) ) } r+=1 File.open(filename + '_column_indexes.txt','w') { |f| f.write Base64.decode64( driver.result(results[r]) ) } r+=1 File.open(filename + '_row_indexes.txt','w') { |f| f.write Base64.decode64( driver.result(results[r]) ) } r+=1 b+=1 # Next bicluster end # Finally, downloads the vector of Cophenetic Correlation Coefficients (CCC) for # the given range of factorization ranks. # # NOTE: This file only exists if a range of factorization ranks was provided, AND # the number of runs (per factorization rank) is greater than 1. if ( r == len ) File.open(out + '_coph.txt','w') { |f| f.write Base64.decode64( driver.result(results[-1]) ) } end ########################## # Saves additional output files. ########################## out_tgz = out + '.bundle.tgz' File.open(out_tgz,'wb') {|f| f.write Base64.decode64( driver.bundle(job_id) )} # Expands and deletes the tgz file. cmd = "tar -vxzf #{ out_tgz } -C #{ OUTPUT_DIR }" ret = `#{ cmd }` if ($?.exitstatus != 0): raise "#{ cmd }\n#{ ret.strip }" end FileUtils.rm_f out_tgz # Clean-up files driver.clean_job_files( job_id ) driver.clean_matrix( matrix_id ) # Capture exceptions rescue Exception => e puts "\nError: #{ e.message } (#{ e.class })" exit 1 end