From ee158d02d92c597a35bc2e8704a85989a547f0e4 Mon Sep 17 00:00:00 2001
From: Thibaut Horel <thibaut.horel@gmail.com>
Date: Mon, 14 Sep 2015 17:08:20 -0400
Subject: Getting read of old stuff

---
 hawkes/data.py  | 20 +++++++++++-----
 hawkes/data2.py | 71 ---------------------------------------------------------
 hawkes/main.py  | 11 ---------
 3 files changed, 14 insertions(+), 88 deletions(-)
 delete mode 100644 hawkes/data2.py

diff --git a/hawkes/data.py b/hawkes/data.py
index 4d7744e..e5c33f8 100644
--- a/hawkes/data.py
+++ b/hawkes/data.py
@@ -2,6 +2,7 @@ from csv import DictReader
 import sys
 from itertools import product
 from cPickle import dump
+from math import cos
 
 MAX_TIME = 3012
 
@@ -10,13 +11,18 @@ def parse(s):
     return None if s == "NA" else int(float(s))
 
 
+def fluctuation_int(t):
+    if t is None:
+        t = MAX_TIME
+    return t + 0.43 / 0.0172 * (cos(4.36) - cos(0.0172 * t + 4.36))
+
+
 def load_nodes(filename):
     with open(filename) as fh:
         reader = DictReader(fh)
         d = {parse(row["name"]): parse(row["fatal_day"]) for row in reader}
     for n, t in d.iteritems():
-        if t is None:
-            d[n] = MAX_TIME
+        d[n] = fluctuation_int(t)
     return d
 
 
@@ -26,10 +32,12 @@ def load_edges(filename):
     with open(filename) as fh:
         reader = DictReader(fh)
         for row in reader:
-            fro, to, t, weight = map(parse, [row["from"], row["to"],
-                                             row["t1"], row["w1"]])
+            fro, to, t, dist = map(parse, [row["from"], row["to"],
+                                           row["t1"], row["dist"]])
+            if dist >= 2:
+                continue
             d = edges.get(fro, dict())
-            d[to] = weight
+            d[to] = dist
             edges[fro] = d
             s = events.get(fro, set())
             s.add(t)
@@ -62,4 +70,4 @@ if __name__ == "__main__":
     nodes = load_nodes(sys.argv[1])
     edges, events = load_edges(sys.argv[2])
     event_edges = compute_event_edges(events, edges)
-    dump((nodes, edges, events, event_edges), open("data.pickle", "wb"))
+    dump((nodes, edges, events, event_edges), open("data-dist1.pickle", "wb"))
diff --git a/hawkes/data2.py b/hawkes/data2.py
deleted file mode 100644
index b8f83c9..0000000
--- a/hawkes/data2.py
+++ /dev/null
@@ -1,71 +0,0 @@
-from csv import DictReader
-import sys
-from itertools import product
-from cPickle import dump
-from math import cos
-
-MAX_TIME = 3012
-
-
-def parse(s):
-    return None if s == "NA" else int(float(s))
-
-
-def fluctuation_int(t):
-    if t is None:
-        t = MAX_TIME
-    return t + 0.43 / 0.0172 * (cos(4.36) - cos(0.0172 * t + 4.36))
-
-
-def load_nodes(filename):
-    with open(filename) as fh:
-        reader = DictReader(fh)
-        d = {parse(row["name"]): parse(row["fatal_day"]) for row in reader}
-    for n, t in d.iteritems():
-        d[n] = fluctuation_int(t)
-    return d
-
-
-def load_edges(filename):
-    events = {}
-    edges = {}
-    with open(filename) as fh:
-        reader = DictReader(fh)
-        for row in reader:
-            fro, to, t, dist = map(parse, [row["from"], row["to"],
-                                           row["t1"], row["dist"]])
-            d = edges.get(fro, dict())
-            d[to] = dist
-            edges[fro] = d
-            s = events.get(fro, set())
-            s.add(t)
-            events[fro] = s
-    return edges, events
-
-
-def compute_event_edges(events, edges):
-    event_edges = {}
-
-    for fro in events:
-        for t in events[fro]:
-            event_edges[(fro, t)] = set()
-
-    for fro in edges:
-        for to in edges[fro]:
-            try:
-                e1, e2 = events[fro], events[to]
-            except KeyError:
-                continue
-            for t1, t2 in product(e1, e2):
-                if t1 < t2:
-                    s = event_edges[(to, t2)]
-                    s.add((fro, t1, edges[fro][to]))
-                    event_edges[(to, t2)] = s
-    return event_edges
-
-
-if __name__ == "__main__":
-    nodes = load_nodes(sys.argv[1])
-    edges, events = load_edges(sys.argv[2])
-    event_edges = compute_event_edges(events, edges)
-    dump((nodes, edges, events, event_edges), open("data2.pickle", "wb"))
diff --git a/hawkes/main.py b/hawkes/main.py
index 40993a1..bc319cc 100644
--- a/hawkes/main.py
+++ b/hawkes/main.py
@@ -34,17 +34,6 @@ def iter_events(events):
             yield (n, t)
 
 
-def ll_old(lamb, alpha, mu):
-    r1 = sum(log(lamb + sum(alpha * w * mu * exp(-mu * (t1 - t2))
-                            for (n2, t2, w) in s))
-             for ((n1, t1), s) in event_edges.iteritems())
-    r2 = sum(sum(alpha * w * (1 - exp(-mu * (nodes[n2] - t1)))
-                 for n2, w in edges[n1].iteritems() if nodes[n2] > t1)
-             for (n1, t1) in iter_events(events))
-    r3 = lamb * sum(nodes.itervalues())
-    return -(r1 - r2 - r3)
-
-
 def ll(lamb, alpha, mu):
     r1 = sum(log(lamb * (1 + 0.43 * sin(0.0172 * t1 + 4.36))
                  + sum(alpha / d * mu * exp(-mu * (t1 - t2))
-- 
cgit v1.2.3-70-g09d2