From ca85703de808e9c5ed6f204ae2217761469ee42a Mon Sep 17 00:00:00 2001
From: Manos <manos@delkappa.com>
Date: Wed, 6 Oct 2021 13:14:28 +0200
Subject: [PATCH] code: WIP

---
 code/expt/copenhagen-sel.py | 23 ++++++-----------------
 code/expt/hue-sel.py        | 23 ++++++-----------------
 code/expt/t-drive-sel.py    | 23 ++++++-----------------
 code/lib/lmdk_sel.py        | 18 ++++++++++++++++++
 4 files changed, 36 insertions(+), 51 deletions(-)

diff --git a/code/expt/copenhagen-sel.py b/code/expt/copenhagen-sel.py
index 26597af..321b48f 100644
--- a/code/expt/copenhagen-sel.py
+++ b/code/expt/copenhagen-sel.py
@@ -8,6 +8,8 @@ from datetime import datetime
 from geopy.distance import distance
 import lmdk_bgt
 import lmdk_lib
+import lmdk_sel
+import exp_mech
 import math
 import numpy as np
 from matplotlib import pyplot as plt
@@ -66,33 +68,20 @@ def main(args):
 
     for _ in range(args.iter):
 
-      eps_sel = 0
-      if pct != 0 and pct != 100:
-        # Get landmarks timestamps in sequence
-        lmdks_seq = lmdk_lib.find_lmdks_seq(seq, lmdks)
-        # Turn landmarks to histogram
-        hist, h = lmdk_lib.get_hist(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-        # Find all possible options
-        opts = lmdk_sel.get_opts_from_top_h(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-        # Landmarks selection budget
-        eps_sel = epsilon/(len(lmdks_seq) + 1)
-        # Get private landmarks timestamps
-        lmdks_seq, _ = exp_mech.exponential_pareto(hist, opts, exp_mech.score, 1.0, eps_sel)
-        # Get actual landmarks values
-        lmdks = seq[lmdks_seq] 
+      lmdks, eps_out = lmdk_sel.find_lmdks(seq, lmdks, epsilon)
 
       # Skip
-      rls_data_s, bgts_s = lmdk_bgt.skip_cont(seq, lmdks, epsilon - eps_sel)
+      rls_data_s, bgts_s = lmdk_bgt.skip_cont(seq, lmdks, eps_out)
       # lmdk_bgt.validate_bgts(seq, lmdks, epsilon, bgts_s)
       mae_s[i] += lmdk_bgt.mae_cont(rls_data_s)/args.iter
 
       # Uniform
-      rls_data_u, bgts_u = lmdk_bgt.uniform_cont(seq, lmdks, epsilon - eps_sel)
+      rls_data_u, bgts_u = lmdk_bgt.uniform_cont(seq, lmdks, eps_out)
       # lmdk_bgt.validate_bgts(seq, lmdks, epsilon, bgts_u)
       mae_u[i] += lmdk_bgt.mae_cont(rls_data_u)/args.iter
 
       # Adaptive
-      rls_data_a, _, _ = lmdk_bgt.adaptive_cont(seq, lmdks, epsilon - eps_sel, .5, .5)
+      rls_data_a, _, _ = lmdk_bgt.adaptive_cont(seq, lmdks, eps_out, .5, .5)
       mae_a[i] += lmdk_bgt.mae_cont(rls_data_a)/args.iter
 
       # Calculate once
diff --git a/code/expt/hue-sel.py b/code/expt/hue-sel.py
index 240fe37..bb70f83 100644
--- a/code/expt/hue-sel.py
+++ b/code/expt/hue-sel.py
@@ -8,6 +8,8 @@ from datetime import datetime
 from geopy.distance import distance
 import lmdk_bgt
 import lmdk_lib
+import lmdk_sel
+import exp_mech
 import math
 import numpy as np
 from matplotlib import pyplot as plt
@@ -62,32 +64,19 @@ def main(args):
 
     for _ in range(args.iter):
 
-      eps_sel = 0
-      if pct != 0 and pct != 100:
-        # Get landmarks timestamps in sequence
-        lmdks_seq = lmdk_lib.find_lmdks_seq(seq, lmdks)
-        # Turn landmarks to histogram
-        hist, h = lmdk_lib.get_hist(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-        # Find all possible options
-        opts = lmdk_sel.get_opts_from_top_h(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-        # Landmarks selection budget
-        eps_sel = epsilon/(len(lmdks_seq) + 1)
-        # Get private landmarks timestamps
-        lmdks_seq, _ = exp_mech.exponential_pareto(hist, opts, exp_mech.score, 1.0, eps_sel)
-        # Get actual landmarks values
-        lmdks = seq[lmdks_seq] 
+      lmdks, eps_out = lmdk_sel.find_lmdks(seq, lmdks, epsilon)
 
       # Skip
-      rls_data_s, bgts_s = lmdk_bgt.skip_cons(seq, lmdks, epsilon - eps_sel)
+      rls_data_s, bgts_s = lmdk_bgt.skip_cons(seq, lmdks, eps_out)
       # lmdk_bgt.validate_bgts(seq, lmdks, epsilon, bgts_s)
       mae_s[i] += lmdk_bgt.mae_cons(seq, rls_data_s)/args.iter
 
       # Uniform
-      rls_data_u, bgts_u = lmdk_bgt.uniform_cons(seq, lmdks, epsilon - eps_sel)
+      rls_data_u, bgts_u = lmdk_bgt.uniform_cons(seq, lmdks, eps_out)
       mae_u[i] += lmdk_bgt.mae_cons(seq, rls_data_u)/args.iter
 
       # Adaptive
-      rls_data_a, _, _ = lmdk_bgt.adaptive_cons(seq, lmdks, epsilon - eps_sel, .5, .5)
+      rls_data_a, _, _ = lmdk_bgt.adaptive_cons(seq, lmdks, eps_out, .5, .5)
       mae_a[i] += lmdk_bgt.mae_cons(seq, rls_data_a)/args.iter
 
       # Calculate once
diff --git a/code/expt/t-drive-sel.py b/code/expt/t-drive-sel.py
index 77a4a1d..1f1683c 100644
--- a/code/expt/t-drive-sel.py
+++ b/code/expt/t-drive-sel.py
@@ -7,6 +7,8 @@ from datetime import datetime
 from geopy.distance import distance
 import lmdk_bgt
 import lmdk_lib
+import lmdk_sel
+import exp_mech
 import numpy as np
 from matplotlib import pyplot as plt
 import time
@@ -85,31 +87,18 @@ def main(args):
       for bgt in bgt_conf:
         for _ in range(args.iter):
 
-          eps_sel = 0
-          if lmdk != 0 and lmdk != 100:
-            # Get landmarks timestamps in sequence
-            lmdks_seq = lmdk_lib.find_lmdks_seq(seq, lmdks)
-            # Turn landmarks to histogram
-            hist, h = lmdk_lib.get_hist(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-            # Find all possible options
-            opts = lmdk_sel.get_opts_from_top_h(lmdk_lib.get_seq(1, len(seq)), lmdks_seq)
-            # Landmarks selection budget
-            eps_sel = bgt['epsilon']/(len(lmdks_seq) + 1)
-            # Get private landmarks timestamps
-            lmdks_seq, _ = exp_mech.exponential_pareto(hist, opts, exp_mech.score, 1.0, eps_sel)
-            # Get actual landmarks values
-            lmdks = seq[lmdks_seq] 
+          lmdks, eps_out = lmdk_sel.find_lmdks(seq, lmdks, bgt['epsilon'])
 
           # Skip
-          rls_data_s, _ = lmdk_bgt.skip(seq, lmdks, bgt['epsilon'] - eps_sel)
+          rls_data_s, _ = lmdk_bgt.skip(seq, lmdks, eps_out)
           mae_s[i] += lmdk_bgt.mae(seq, rls_data_s)/args.iter
 
           # Uniform
-          rls_data_u, _ = lmdk_bgt.uniform_r(seq, lmdks, bgt['epsilon'] - eps_sel)
+          rls_data_u, _ = lmdk_bgt.uniform_r(seq, lmdks, eps_out)
           mae_u[i] += lmdk_bgt.mae(seq, rls_data_u)/args.iter
 
           # Adaptive
-          rls_data_a, _, _ = lmdk_bgt.adaptive(seq, lmdks, bgt['epsilon'] - eps_sel, .5, .5)
+          rls_data_a, _, _ = lmdk_bgt.adaptive(seq, lmdks, eps_out, .5, .5)
           mae_a[i] += lmdk_bgt.mae(seq, rls_data_a)/args.iter
 
           # Event
diff --git a/code/lib/lmdk_sel.py b/code/lib/lmdk_sel.py
index 2685844..421ab7e 100644
--- a/code/lib/lmdk_sel.py
+++ b/code/lib/lmdk_sel.py
@@ -2,6 +2,7 @@
 
 import itertools
 from lmdk_lib import *
+import exp_mech
 import numpy as np
 import random
 import time
@@ -297,6 +298,23 @@ def get_non_opts_from_bottom(seq, lmdks):
   return non_opts
 
 
+def find_lmdks(seq, lmdks, epsilon):
+  eps_sel = 0
+  if len(lmdks) > 0 and len(seq) != len(lmdks):
+    # Get landmarks timestamps in sequence
+    lmdks_seq = find_lmdks_seq(seq, lmdks)
+    # Turn landmarks to histogram
+    hist, h = get_hist(get_seq(1, len(seq)), lmdks_seq)
+    # Find all possible options
+    opts = get_opts_from_top_h(get_seq(1, len(seq)), lmdks_seq)
+    # Landmarks selection budget
+    eps_sel = epsilon/(len(lmdks_seq) + 1)
+    # Get private landmarks timestamps
+    lmdks_seq, _ = exp_mech.exponential_pareto(hist, opts, exp_mech.score, 1.0, eps_sel)
+    # Get actual landmarks values
+    lmdks = seq[lmdks_seq]
+  return lmdks, epsilon - eps_sel
+
 def test():
   # Start and end points of the sequence
   # # Nonrandom