code: Ready to find landmarks in copenhagen

code/parse_copenhagen.py

@@ -7,6 +7,7 @@ import csv
 from datetime import datetime
 from geopy.distance import distance
 import io
+import itertools
 import lmdk_lib
 import numpy as np
 import os
@@ -56,6 +57,48 @@ def main(args):
       print('[Error: %s]' %(e))
     # Save to results
     lmdk_lib.save_data(args, np.array(cont, np.float32), 'cont')
+  # Get all users
+  usrs = np.unique(np.concatenate((cont_data[:, uid_a], cont_data[:, uid_b]), 0))
+  # Check each user
+  goal = [.2, .4, .6, .8, 1]
+  # Users suitable for experiments
+  usrs_expt = []
+  for usr_i, usr in enumerate(usrs):
+    print('Checking %d (%d%%: %d/%d)... ' %(usr, (usr_i + 1)*100/len(usrs), usr_i + 1, len(usrs)), end='', flush=True)
+    # User's contacts
+    usr_cont = cont_data[(cont_data[:, uid_a] == usr) | (cont_data[:, uid_b] == usr)]
+    # For each goal
+    for_expt = True
+    for g in goal:
+      if for_expt:
+        # Possible contacts
+        pos_cont = []
+        usrs_cur = list(usrs)
+        # Remove user
+        usrs_cur.remove(usr)
+        # Check for every possible contact
+        for u in usrs_cur:
+          # Add possible contacts gradually
+          pos_cont.append(u)
+          # Remove from user contacts
+          usr_cont_cur = np.copy(usr_cont)
+          for pos_c in pos_cont:
+            usr_cont_cur = usr_cont_cur[(usr_cont_cur[:, uid_a] != pos_c) & (usr_cont_cur[:, uid_b] != pos_c)]
+          # Compare the difference
+          diff = (len(usr_cont) - len(usr_cont_cur))/len(usr_cont)
+          # Check if it's close enough to what we need
+          if abs(diff - g)/g < .025:
+            usrs_expt.append([usr, g, pos_cont])
+            if g == 1:
+              # That's a keeper
+              print('[OK]')
+            break
+          elif diff > g:
+            print('[NOK]')
+            for_expt = False
+            break
+  # Save to results
+  lmdk_lib.save_data(args, np.array(usrs_expt, str), 'usrs_expt')
 
 
 '''