Rama Analysis - Refactor a nuevo sistema de ficheros

Analysis/MallTimes.py

@@ -45,33 +45,26 @@ def record(f, observation, line):
   #else:
    # getData(lineS, observation, 13)
 
-
-#-----------------------------------------------
-def read_file(f, dataA, dataB, it):
-  recording = False
-  resizes = 0
-  timer = 0
-  previousNP = 0
-
-  for line in f: 
-    lineS = line.split()
-
-    if len(lineS) > 0:
-      if lineS[0] == "Config": # CONFIG LINE
-        recording = True
-        it += 1
+# Obtains the value of a given index in a splited line
+# and returns it as a float values
+def get_value(line, index):
+  return float(line[index].split('=')[1].split(',')[0])
+
+# Obtains the general parameters of an execution and
+# stores them for creating a dataframe
+def record_config_line(lineS, dataA, dataB):
   dataA.append([None]*13)
   dataB.append([None]*15)
-        #resizes = int(lineS[2].split('=')[1].split(',')[0])
-        resizes = 2
-        compute_tam = int(lineS[3].split('=')[1].split(',')[0])
-        comm_tam = int(lineS[4].split('=')[1].split(',')[0])
-        sdr = int(lineS[5].split('=')[1].split(',')[0])
-        adr = int(lineS[6].split('=')[1].split(',')[0]) #TODO Que lo tome como porcentaje
-        css = int(lineS[8].split('=')[1].split(',')[0])
-        cst = int(lineS[9].split('=')[1].split(',')[0])
-        # TODO Que obtenga Aib
-        time = float(lineS[10].split('=')[1])
+  resizes = int(get_value(lineS, 2))
+  stages = int(get_value(lineS, 3))
+  compute_tam = int(get_value(lineS, 4))
+  sdr = int(get_value(lineS, 5))
+  adr = int(get_value(lineS, 6)) #TODO Que lo tome como porcentaje
+  at  = int(get_value(lineS, 7))
+  sm = int(get_value(lineS, 8))
+  ss = int(get_value(lineS, 9))
+  latency = get_value(lineS, 10)
+  bw = get_value(lineS, 11)
 
   dataB[it][0] = sdr
   dataB[it][1] = adr 
@@ -93,6 +86,64 @@ def read_file(f, dataA, dataB, it):
   dataA[it][10] = time
   dataA[it][11] = ""
 
+def record_stage_line(lineS, dataG_it, dataM_it):
+  pt = int(get_value(lineS, 2))
+  t_stage = get_value(lineS, 3)
+  u_bytes = int(get_value(lineS, 4))
+
+  dataG_it[].append(pt)
+  dataG_it[].append(t_stage)
+  dataG_it[].append(u_bytes)
+
+  dataM_it[].append(pt)
+  dataM_it[].append(t_stage)
+  dataM_it[].append(u_bytes)
+  
+def record_resize_line(lineS, dataG_it, dataM_it):
+        iters = int(lineS[2].split('=')[1].split(',')[0])
+        npr = int(lineS[3].split('=')[1].split(',')[0])
+        dist = lineS[5].split('=')[1]
+
+        resizes = resizes - 1
+        if resizes == 0:
+          dataB[it][3] = npr
+          dataB[it][4] += dist
+          dataB[it][10] += str(iters)
+
+          dataA[it][4] = npr #FIXME No sera correcta si hay mas de una reconfig
+          dataA[it][2] = str(previousNP) + "," + str(npr)
+          dataA[it][5] += dist
+          dataA[it][11] += str(iters)
+          timer = 4
+        else:
+          dataB[it][2] = npr
+          dataB[it][4] += dist + ","
+          dataB[it][10] += str(iters) + ","
+
+          dataA[it][3] = npr
+          dataA[it][5] += dist + ","
+          dataA[it][11] += str(iters) + ","
+          previousNP = npr
+
+#-----------------------------------------------
+def read_file(f, dataA, dataB, it):
+  recording = False
+  resizes = 0
+  timer = 0
+  previousNP = 0
+
+  for line in f: 
+    lineS = line.split()
+
+    if len(lineS) > 0:
+      if lineS[0] == "Config": # CONFIG LINE
+        recording = True
+        it += 1
+        record_config(lineS, dataG, dataM)
+
+      elif lineS[0] == "Stage":
+          record_stage_line(lineS, dataG, dataM)
+      elif lineS[0] == "Resize":
       elif recording and resizes != 0: # RESIZE LINE
         iters = int(lineS[2].split('=')[1].split(',')[0])
         npr = int(lineS[3].split('=')[1].split(',')[0])
@@ -165,26 +216,27 @@ else:
 
 insideDir = "Run"
 lista = glob.glob("./" + BaseDir + insideDir + "*/" + sys.argv[1]+ "*Global.o*")
+lista += (glob.glob("./" + BaseDir + sys.argv[1]+ "*Global.o*")) # Se utiliza cuando solo hay un nivel de directorios
 print("Number of files found: "+ str(len(lista)));
 
 it = -1
-dataA = []
-dataB = []
-columnsA = ["N", "%Async", "Groups", "NP", "NS", "Dist", "Matrix", "CommTam", "Cst", "Css", "Time", "Iters", "TE"] #13
-columnsB = ["N", "%Async", "NP", "NS", "Dist", "Matrix", "CommTam", "Cst", "Css", "Time", "Iters", "TC", "TH", "TS", "TA"] #15
+dataG = []
+dataM = []
+columnsG = ["N", "%Async", "Groups", "NP", "NS", "Dist", "Matrix", "CommTam", "Cst", "Css", "Time", "Iters", "TE"] #13
+columnsM = ["N", "%Async", "NP", "NS", "Dist", "Matrix", "CommTam", "Cst", "Css", "Time", "Iters", "TC", "TH", "TS", "TA"] #15
 
 for elem in lista:
   f = open(elem, "r")
-  it = read_file(f, dataA, dataB, it)
+  it = read_file(f, dataG, dataM, it)
   f.close()
 
 #print(data)
-dfA = pd.DataFrame(dataA, columns=columnsA)
-dfA.to_csv(name + 'G.csv')
+dfG = pd.DataFrame(dataG, columns=columnsG)
+dfG.to_csv(name + 'G.csv')
 
-dfB = pd.DataFrame(dataB, columns=columnsB)
+dfM = pd.DataFrame(dataM, columns=columnsM)
 
 #Poner en TC el valor real y en TH el necesario para la app
-cond = dfB.TH != 0
-dfB.loc[cond, ['TC', 'TH']] = dfB.loc[cond, ['TH', 'TC']].values
-dfB.to_csv(name + 'M.csv')
+cond = dfM.TH != 0
+dfM.loc[cond, ['TC', 'TH']] = dfB.loc[cond, ['TH', 'TC']].values
+dfM.to_csv(name + 'M.csv')