a quick and dirty python script to extract the features from the FMA dataset

Author: lefaivre

calljAudio.py

@@ -0,0 +1,57 @@
+import os
+import glob
+import copy
+import subprocess
+
+def getFilesInDirWithExt(dir, ext):
+    toreturn = []
+    os.chdir(dir)
+    for file in glob.glob(ext):
+        toreturn.append(dir + file)
+    return toreturn
+
+batchTemplate = []
+with open('D:\\jAudio\\batchfiles\\template.xml') as f:
+    batchTemplate = f.readlines()
+
+mainpath = "F:\\fma_large\\" 
+
+allAudioFolders = os.listdir(mainpath)
+
+# Need to create all batch files first.
+for num in allAudioFolders:
+
+    folder = mainpath + num + "\\"
+    wavFilesInFolder = getFilesInDirWithExt(folder, "*.wav")
+
+    output = batchTemplate[0:27]
+    output.append("\t<batch ID=\"" + num + "\">\n")
+
+    fileSetStrings = []
+    fileSetStrings.append("\t\t<fileSet>\n")
+    for wavFile in wavFilesInFolder:
+        fileSetStrings.append("\t\t\t<file>" + wavFile + "</file>\n")
+    fileSetStrings.append("\t\t</fileSet>\n")
+    output.extend(fileSetStrings)
+    output.extend(batchTemplate[31:770])
+    output.append("\t\t<destination>" + "D:\\jAudio\\fma_large\\" + num + ".arff</destination>\n")
+    output.append("\t\t<destination>" + "D:\\jAudio\\fma_large\\" + num + ".arff</destination>\n")
+    output.extend(batchTemplate[772:775])
+
+    with open('D:\\jAudio\\batchfiles\\' + num + ".xml", 'w') as f:
+        f.writelines(output)
+
+
+# run batch files!
+batchFileDir = "D:\\jAudio\\batchfiles\\"
+allBatchFiles = os.listdir(batchFileDir)
+
+for batchFile in allBatchFiles:
+    if not (batchFile == "template"):
+        thisBatchFile = batchFileDir + batchFile
+        os.chdir("D:\\jAudio")
+        FNULL = open(os.devnull, 'w')
+        subprocess.call('java -mx500M -jar jAudio-1.0.4.jar -b ' + thisBatchFile, shell=True,  stdout=FNULL, stderr=subprocess.STDOUT)
+
+
+