From 7c09ec981c6e06af2e62d67a609eb53728267954 Mon Sep 17 00:00:00 2001
From: Jonathan Herman <hermanjl@cs.unc.edu>
Date: Thu, 27 Sep 2012 19:03:22 -0400
Subject: Added script to parse directory data, create CSVs for every chagned
 value.

This change also makes SchedTrace and OverheadTrace events configurable.
---
 parse/tuple_table.py | 76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 76 insertions(+)
 create mode 100644 parse/tuple_table.py

(limited to 'parse/tuple_table.py')

diff --git a/parse/tuple_table.py b/parse/tuple_table.py
new file mode 100644
index 0000000..df80b37
--- /dev/null
+++ b/parse/tuple_table.py
@@ -0,0 +1,76 @@
+from collections import defaultdict
+from point import SummaryPoint
+from dir_map import DirMap
+
+class ColMap(object):
+    def __init__(self):
+        self.rev_map = {}
+        self.col_list = []
+
+    def columns(self):
+        return self.col_list
+
+    def get_key(self, kv):
+        key = ()
+        added = 0
+        
+        for col in self.col_list:
+            if col not in kv:
+                key += (None,)
+            else:
+                added += 1
+                key += (kv[col],)
+
+        if added != len(kv):
+            raise Exception("column map '%s' missed field in map\n%s" %
+                            (self.col_list, kv))
+               
+        return key
+
+    def get_map(self, tuple):
+        map = {}
+        for i in range(0, len(tuple)):
+            map[self.col_list[i]] = tuple[i]
+        return map
+
+    def try_add(self, column):
+        if column not in self.rev_map:
+            self.rev_map[column] = len(self.col_list)
+            self.col_list += [column]
+
+    def __str__(self):
+        return "<ColMap>%s" % (self.rev_map)
+    
+class TupleTable(object):
+    def __init__(self, col_map):
+        self.col_map = col_map
+        self.table = defaultdict(lambda: [])
+        self.reduced = False
+
+    def add_exp(self, kv, point):
+        key = self.col_map.get_key(kv)
+        self.table[key] += [point]
+
+    def __reduce(self):
+        if self.reduced:
+            raise Exception("cannot reduce twice!")
+        self.reduced = True
+        for key, values in self.table.iteritems():
+            self.table[key] = SummaryPoint(key, values)
+
+    def write_result(self, out_dir):
+        dir_map = DirMap(out_dir)
+        self.__reduce()
+        for key, point in self.table.iteritems():
+            kv = self.col_map.get_map(key)
+
+            for col in self.col_map.columns():
+                val = kv[col]
+                kv.pop(col)
+
+                dir_map.add_point(col, val, kv, point)
+
+                kv[col] = val
+
+        dir_map.reduce()
+        dir_map.write()
-- 
cgit v1.2.2