Create jams_to_mirex_pattern.py

scripts/jams_to_mirex_pattern.py

@@ -0,0 +1,65 @@
+
+import csv
+import json
+
+x = open('/path/to/file.jams').read()
+x = json.loads(x)
+
+record = []
+
+for x in x["annotations"]:
+    # looping through the annotators
+    metadata = x["annotation_metadata"]
+    namespace = x["namespace"]
+
+    # annotator names
+    anno = metadata["annotator"]
+
+    # actual data of annotation
+    d = x["data"]
+
+    # open file to write to the MIREX pattern discovery format 
+    text_file = open("anno{0}.txt".format(anno), "w")
+    text_file.write("Pattern1"+"\n")
+    text_file.write("Occ1"+"\n")
+    pcount = 1 # keep pattern count
+    ocount = 1 # keep pattern occurrence count
+
+    pastpid = 1 # initiatise the pattern count to compare for a change
+    pastoid = 1 # initiatise the occurrence count to compare for a change
+
+    initpid = d[0]["value"]["pattern_id"]
+
+    for y in d:
+        # lopping through the events given an annotator
+        time = y["time"]
+        dur = y["duration"]
+
+        pid = y["value"]["pattern_id"]
+
+        # update the pattern count when there is a change in the pattern id in the jams file
+        if pid != pastpid:
+            pcount += 1
+            ocount = 0
+            text_file.write("Pattern"+str(pcount)+"\n")
+            text_file.write("Occ1"+"\n")
+
+        midip = y["value"]["midi_pitch"]
+        morphp = y["value"]["morph_pitch"]
+        staff = y["value"]["staff"]
+        occid = y["value"]["occurrence_id"]
+
+        if occid != pastoid:
+            ocount += 1
+            text_file.write("Occ"+str(ocount)+"\n")
+
+        c = y["confidence"]
+
+        # keep a record of the last pattern count to compare with for a change
+        pastpid = pid
+        pastoid = occid
+
+        # write the actual (time, pitch) pairs
+        text_file.write(str(time)+", "+str(midip)+"\n")
+
+    text_file.close()