Merge pull request #46 from CSDLLab/2020_05_19_optimize

final stage on parameter optimization

Author: HanZhaang

config/fake_reconstruction_configs/branch_metric.json

@@ -0,0 +1,2 @@
+This folder is used to place the temporary reconstruct configs.
+Dont' delete this folder.

config/fake_reconstruction_configs/test.json

@@ -0,0 +1,2 @@
+This folder is used to place the temporary reconstruct results.
+Dont' delete this folder.

config/fake_reconstruction_configs/test_bak.json

@@ -3,13 +3,13 @@
 # @Time    : 2019/8/17
 # @Author  : github.com/guofei9987
 import copy
-
+import os
 import numpy as np
+import multiprocessing as mp
 from sko.base import SkoBase
 from sko.operators import mutation
-import multiprocessing as mp
 
-CPU_CORE_NUM = 12
+CPU_CORE_NUM = 15
 
 
 class SimulatedAnnealingBase(SkoBase):
@@ -53,7 +53,7 @@ def __init__(self, func, x0, T_max=100, T_min=1e-7, L=300, max_stay_counter=150,
         self.n_dims = len(x0)
 
         self.best_x = np.array(x0)  # initial solution
-        self.best_y = self.func(self.best_x, "test_init.swc")[1]
+        self.best_y = self.func(self.best_x, "test_init")[1]
         self.T = self.T_max
         self.iter_cycle = 0
         self.generation_best_X, self.generation_best_Y = [self.best_x], [self.best_y]
@@ -88,7 +88,7 @@ def run(self):
                         x_new[k] = max(x_new[k], 0)
                         x_new[k] = min(x_new[k], 1)
                     res_y.append(
-                        pool.apply_async(self.func, args=tuple([x_new, "test256_{}".format(j), lock]))
+                        pool.apply_async(self.func, args=tuple([x_new, os.path.join("tmp", "tmp_res_{}".format(j)), lock]))
                     )
                     res_x.append(x_new)
 
@@ -98,8 +98,6 @@ def run(self):
                 for it in range(len(res_x)):
                     i += 1
                     x_new, y_new = res_y[it].get()
-                    print(x_new)
-                    print(y_new)
                     # Metropolis
                     df = y_new - y_current
                     if df < 0 or np.exp(-df / self.T) > np.random.rand():
@@ -110,17 +108,15 @@ def run(self):
                             self.best_x = copy.deepcopy(x_new)
                             self.best_y = y_new
                         break
-                    print("[Info: ] best x = {}".format(self.best_x))
-                    print("[Info: ] best y = {}".format(self.best_y))
+            print("[Info: ] best x = {}".format(self.best_x))
+            print("[Info: ] best y = {}".format(self.best_y))
 
             print("[Info: ] iter_cycle = {} T = {} stay_counter = {}".format(
                 self.iter_cycle, self.T, stay_counter
             ))
             print("[Info: ]origin minimalScoreAuto = {}\n"
-                  "        minimalScoreManual = {}\n"
-                  "        minimalScoreSeed = {}\n"
-                  "        minimalScore2d = {}".format(
-                self.best_x[0], self.best_x[1], self.best_x[2], self.best_x[3]
+                  "        minimalScoreSeed = {}".format(
+                self.best_x[0], self.best_x[1]
             ))
             self.iter_cycle += 1
             self.cool_down()

config/fake_reconstruction_configs/tmp/readme.txt

@@ -7,10 +7,10 @@
 from scipy import stats as st
 
 import copy
+import time
 import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
-import time
 
 g_score = -1
 g_gold_tree = None
@@ -20,31 +20,36 @@
 g_metric_method = None
 g_metric_configs = None
 
-NEUTU_PATH = "../../../../../00_program_file/00_neutu/bin/neutu"
-ORIGIN_PATH = "../../../data/optimation/test1/test1_test.tif"
-GOLD_PATH = "../../../data/optimation/test1/test1_gold.swc"
-TEST_PATH = "../../../data/optimation/output/"
-CONFIG_PATH = "../../../config/fake_reconstruction_configs/"
+# Gold file and Test tif need to placed in ../../../data/optimation
+# and named after FILE_ID_test.tif, FILE_ID_gold.swc
+FILE_ID = "test1"
+NEUTU_PATH = "neutu"
+ORIGIN_PATH = "../../../data/optimation/{}/{}_test.tif".format(FILE_ID, FILE_ID)
+GOLD_PATH = "../../../data/optimation/{}/{}_gold.swc".format(FILE_ID, FILE_ID)
 METRIC_CONFIG_PATH = "../../../config/ssd_metric.json"
 LOG_PATH = "../../../output/optimization/neutu_log.txt"
+# specific test name is given in SA.py
+TEST_PATH = "../../../data/optimation/output/"
+CONFIG_PATH = "../../../config/fake_reconstruction_configs/"
+
 
 
 def SA_optimize(configs=None, test_name=None, lock=None):
     global g_metric_method
     global g_metric_configs
     global g_rcn_config
-
-    LOC_CONFIG_PATH = os.path.join(CONFIG_PATH, test_name+".json")
-    LOC_TEST_PATH = os.path.join(TEST_PATH, test_name+".swc")
+    # identify specific TEST output path and CONFIG input PATH
+    LOC_TEST_PATH = os.path.join(TEST_PATH, test_name+"_test.swc")
     rec_config = copy.deepcopy(g_rcn_config)
 
     if configs is not None:
+        LOC_CONFIG_PATH = os.path.join(CONFIG_PATH, test_name+".json")
         rec_config["trace"]["default"]["minimalScoreAuto"] = configs[0]
-        rec_config["trace"]["default"]["minimalScoreManual"] = configs[1]
-        rec_config["trace"]["default"]["minimalScoreSeed"] = configs[2]
-        rec_config["trace"]["default"]["minimalScore2d"] = configs[3]
+        rec_config["trace"]["default"]["minimalScoreSeed"] = configs[1]
 
         read_json.save_json(LOC_CONFIG_PATH, rec_config)
+    else:
+        LOC_CONFIG_PATH = os.path.join(CONFIG_PATH, "best_x_{}.json".format(test_name))
 
     REC_CMD = "{} --command --trace {} -o {} --config {} > {}".format(
         NEUTU_PATH, ORIGIN_PATH, LOC_TEST_PATH, LOC_CONFIG_PATH, LOG_PATH
@@ -56,7 +61,7 @@ def SA_optimize(configs=None, test_name=None, lock=None):
 
     res_tree = swc_node.SwcTree()
     gold_tree = swc_node.SwcTree()
-    res_tree.load(os.path.join(TEST_PATH, test_name+".swc"))
+    res_tree.load(LOC_TEST_PATH)
     gold_tree.load(GOLD_PATH)
 
     if lock is not None:
@@ -78,29 +83,29 @@ def main():
     global g_rcn_config
     g_metric_method = ssd_metric.ssd_metric
     g_metric_configs = read_json.read_json(METRIC_CONFIG_PATH)
-    g_rcn_config = read_json.read_json(os.path.join(CONFIG_PATH, "test.json"))
+    g_rcn_config = read_json.read_json(os.path.join(CONFIG_PATH, "default.json"))
 
     # optimize with SA
     # configs here is the config of the reconstruction
-    configs = (0.3, 0.3, 0.35, 0.5)
+    configs = (0.3, 0.35)
     start = time.time()
     sa_fast = SAFast(func=SA_optimize,
-                     x0=configs, T_max=0.01, T_min=1e-5, q=0.96, L=20, max_stay_counter=50, upper=1, lower=0)
+                     x0=configs, T_max=0.01, T_min=1e-5, q=0.96, L=25, max_stay_counter=15, upper=1, lower=0)
     best_configs, best_value = sa_fast.run()
     print("[Info: ]best configs:\n"
           "        origin minimalScoreAuto = {}\n"
-          "        minimalScoreManual = {}\n"
           "        minimalScoreSeed = {}\n"
-          "        minimalScore2d = {}\n"
           "        best value = {}\n"
           "        time = {}\n" .format(
-        best_configs[0], best_configs[1], best_configs[2], best_configs[3], best_value, time.time() - start
+        best_configs[0], best_configs[1], best_value, time.time() - start
     ))
+    # save best json file
     g_rcn_config["trace"]["default"]["minimalScoreAuto"] = best_configs[0]
-    g_rcn_config["trace"]["default"]["minimalScoreManual"] = best_configs[1]
-    g_rcn_config["trace"]["default"]["minimalScoreSeed"] = best_configs[2]
-    g_rcn_config["trace"]["default"]["minimalScore2d"] = best_configs[3]
-    read_json.save_json(os.path.join(CONFIG_PATH, "best_x_{}.json".format(time.time())), g_rcn_config)
+    g_rcn_config["trace"]["default"]["minimalScoreSeed"] = best_configs[1]
+    read_json.save_json(os.path.join(CONFIG_PATH, "best_x_{}.json".format(FILE_ID)), g_rcn_config)
+    # get and save best reconstruct swc
+    print("[Info: ]Exam test score with best configs: ")
+    cfg, score = SA_optimize(test_name = FILE_ID)
     # plot the result.
     plt.plot(pd.DataFrame(sa_fast.best_y_history).cummin(axis=0))
     plt.xlabel("iterations")
@@ -114,6 +119,6 @@ def main():
 
     # g_metric_method = ssd_metric.ssd_metric
     # g_metric_configs = read_json.read_json(METRIC_CONFIG_PATH)
-    # g_rcn_config = read_json.read_json(os.path.join(CONFIG_PATH, "test3best.json"))
-    #
-    # SA_optimize(test_name="test3best")
+    # g_rcn_config = read_json.read_json(os.path.join(CONFIG_PATH, "6656_2304_22016.json"))
+    
+    # SA_optimize(test_name="6656_2816_22016")

data/optimation/output/tmp/readme.txt

@@ -9,10 +9,10 @@
       url='https://github.com/bennieHan/PyNeval.git',
       packages=['pyneval', 'pyneval.cli', 'test', 'test.test_model',
                 'test.test_model.diadem_metric', 'test.test_model.length_metric',
-                'pyneval.io', 'pyneval.metric', 'pyneval.metric.utils', 'pyneval.metric.utils.klib', 'pyneval.model', 'pyneval.tools'],
+                'pyneval.io', 'pyneval.metric', 'pyneval.metric.utils', 'pyneval.metric.utils.klib', 'pyneval.model', 'pyneval.tools', 'pyneval.tools.optimize'],
       py_modules=['pyneval', 'pyneval.cli', 'test', 'test.test_model',
                 'test.test_model.diadem_metric', 'test.test_model.length_metric',
-                'pyneval.io', 'pyneval.metric', 'pyneval.metric.utils', 'pyneval.metric.utils.klib', 'pyneval.model', 'pyneval.tools'],
+                'pyneval.io', 'pyneval.metric', 'pyneval.metric.utils', 'pyneval.metric.utils.klib', 'pyneval.model', 'pyneval.tools', 'pyneval.tools.optimize'],
       data_files=[('config', glob.glob('config/*.json')),
                   ("config/schemas", glob.glob('config/schemas/*.json'))],
       install_requires=[