From 9bf96d1dd4d53214c1c2155a738a5b795cf38424 Mon Sep 17 00:00:00 2001
From: Abdulrahman Semrie <hsamireh@gmail.com>
Date: Fri, 21 Feb 2020 16:09:22 +0300
Subject: [PATCH] Add target feature option for moses runner

---
 crossval/moses_cross_val.py | 6 +++---
 crossval/moses_runner.py    | 5 +++--
 crossval/post_process.py    | 5 +++--
 3 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/crossval/moses_cross_val.py b/crossval/moses_cross_val.py
index d2f9f9e..8abfa2a 100644
--- a/crossval/moses_cross_val.py
+++ b/crossval/moses_cross_val.py
@@ -6,7 +6,7 @@
 
 import pandas as pd
 from scipy import stats
-from sklearn.model_selection import StratifiedShuffleSplit
+from sklearn.model_selection import StratifiedShuffleSplit, ShuffleSplit
 
 from crossval.model_evaluator import ModelEvaluator
 from crossval.moses_runner import MosesRunner
@@ -95,7 +95,7 @@ def split_dataset(self):
 
         x, y = df.values, df[self.session.target_feature].values
         splits, test_size = self.session.crossval_options["folds"], self.session.crossval_options["testSize"]
-        cv = StratifiedShuffleSplit(n_splits=splits, test_size=test_size)
+        cv = ShuffleSplit(n_splits=splits, test_size=test_size, random_state=42)
 
         return x, df.columns.values, cv.split(x, y)
 
@@ -111,7 +111,7 @@ def run_seeds(self, seeds, i, file):
             output_file = tempfile.NamedTemporaryFile(mode="w+")
             moses_options = " ".join([self.session.moses_options, "--random-seed " + str(seed)])
 
-            moses_runner = MosesRunner(file, output_file.name, moses_options, self.session.mnemonic)
+            moses_runner = MosesRunner(file, output_file.name, moses_options, self.session.mnemonic, self.session.target_feature)
             returncode, stdout, stderr = moses_runner.run_moses()
 
             if returncode != 0:
diff --git a/crossval/moses_runner.py b/crossval/moses_runner.py
index 24f8059..8be67f2 100644
--- a/crossval/moses_runner.py
+++ b/crossval/moses_runner.py
@@ -12,7 +12,7 @@ class MosesRunner:
     A class that handles running of the MOSES binary program
     """
 
-    def __init__(self, input_file, output_file, moses_opts, session_id):
+    def __init__(self, input_file, output_file, moses_opts, session_id, target_feature="case"):
         """
         :param input_file: The input file to run MOSES on
         :param output_file: The file to write MOSES program outputs to
@@ -24,6 +24,7 @@ def __init__(self, input_file, output_file, moses_opts, session_id):
         if not "W1" in moses_opts: moses_opts += ' -W1'
         self.output_regex = re.compile(r"(-?\d+) (.+) \[(.+)\]")
         self.logger = get_logger(session_id)
+        self.target_feature = target_feature
 
     def run_moses(self):
         """
@@ -34,7 +35,7 @@ def run_moses(self):
         :returns stdin: the error output of the process, if any
         """
 
-        cmd = ["moses", "-i", self.input, "-o", self.output]
+        cmd = ["moses", "-i", self.input, "-o", self.output, "--target-feature", self.target_feature]
 
         for opt in self.moses_options.split():
             cmd.append(opt)
diff --git a/crossval/post_process.py b/crossval/post_process.py
index 3ed4ad9..7f7d479 100644
--- a/crossval/post_process.py
+++ b/crossval/post_process.py
@@ -1,10 +1,11 @@
 __author__ = 'Abdulrahman Semrie<hsamireh@gmail.com>'
 
-import pymongo
-from config import MONGODB_URI, DB_NAME, DATASET_DIR
 import os
 import pathlib
+
 import pandas as pd
+
+from config import DATASET_DIR
 from crossval.filters import loader
 from models.objmodel import Score, MosesModel