From 4ddef7c3136101ecd189f18715faca36e46f713f Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Sat, 6 Jun 2020 16:54:34 +0530
Subject: [PATCH 01/10] Implements ELO Ranking Algorithm

---
 pythia/cleaning/elo.py | 217 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 217 insertions(+)
 create mode 100644 pythia/cleaning/elo.py

diff --git a/pythia/cleaning/elo.py b/pythia/cleaning/elo.py
new file mode 100644
index 0000000..04610ec
--- /dev/null
+++ b/pythia/cleaning/elo.py
@@ -0,0 +1,217 @@
+import pandas as pd
+import numpy as np
+from collections import deque
+from sunpy.util import SunpyUserWarning
+
+
+__all__ = ['ELO']
+
+
+class ELO:
+    """
+    Recreating the ELO rating algirithm for Sunspotter.
+    """
+
+    def __init__(self, score_board: pd.DataFrame, *, k_value=32, default_score=1400,
+                 max_comparisons=50, max_score_change=32, min_score_change=16, score_memory=10,
+                 delimiter=';', column_map={"player 0": "image_id_0",
+                                            "player 1": "image_id_1",
+                                            "score for player 0": "image0_more_complex_image1"}):
+        """
+        Parameters
+        ----------
+        score_board : pandas.DataFrame
+            DataFrame holding the scores of individual matches.
+        k_value : int, optional
+            Initial K Value to be used for calculating new ratings, by default 32
+        default_score : int, optional
+            Initial rating, by default 1400
+        max_comparisons : int, optional
+            Max comparisions for any player, by default 50
+        max_score_change : int, optional
+            Upper limit on K Value updation, by default 32
+        min_score_change : int, optional
+            Lower limit on K Value updation, by default 16
+        score_memory : int, optional
+            Number of previous scores to consider while calculating
+            standard deviation and new K value, by default 10
+        column_map : dict, optional
+            Dictionary, for mapping the column names of the score_board dataframe
+            to variable names used in the ELO ranking system.
+            by default {"player 0": "image_id_0", 
+                                    "player 1": "image_id_1", 
+                                    "score for player 0": "image0_more_complex_image1"}
+        """
+        self.score_board = score_board
+        self.k_value = k_value
+        self.default_score = default_score
+        self.score_change = {'min': min_score_change, 'max': max_score_change}
+        self.max_comparisions = max_comparisons
+        self.score_memory = score_memory
+        self.column_map = column_map
+
+        if not set(self.column_map.values()).issubset(self.score_board.columns):
+            missing_columns = set(self.column_map.values()) - set(self.column_map.values()).intersection(self.score_board.columns)
+            missing_columns = ", ".join(missing_columns)
+
+            raise SunpyUserWarning("The following columns mentioned in the column map"
+                                   " are not present in the score board: " +
+                                   missing_columns)
+
+        self._create_ranking()
+
+    def _create_ranking(self):
+        """
+        Prepares the Ranking DataFrame.
+        """
+        image_ids = set(self.score_board[self.column_map['player 0']]).union(self.score_board[self.column_map['player 1']])
+        self.rankings = pd.DataFrame(image_ids, columns=['player id'])
+        self.rankings.set_axis(self.rankings['player id'], inplace=True)
+        self.rankings['score'] = self.default_score
+        self.rankings['k value'] = self.k_value
+        self.rankings['count'] = 0
+        self.rankings['std dev'] = self.score_change['max']
+        self.rankings['last scores'] = str(self.default_score)
+
+    def expected_score(self, score_image_0, score_image_1):
+        """
+        Given two AR scores, calculates expected probability of `image_0` being more complex.
+
+        Parameters
+        ----------
+        score_image_0 : int
+            Score for first image
+        score_image_1 : int
+            Score for second image
+
+        Returns
+        -------
+        expected_0_score : float
+            Expected probability of `image_0` being more complex.
+        """
+        expected_0_score = 1.0 / (1.0 + 10 ** ((score_image_1 - score_image_0) / 400.00))
+        return expected_0_score
+
+    def new_rating(self, rating_for_image, k_value, score_for_image, image_expected_score):
+        """
+        Calculates new rating based on the ELO algorithm.
+
+        Parameters
+        ----------
+        rating_for_image : float
+            Current Rating for the image
+        k_value : float
+            Current k_value for the image
+        score_for_image : int
+            Actual result of classification of the image in a pairwise match.
+            `0` denotes less complex, `1` denotes more complex
+        image_expected_score : float
+            Expected result of classification of image in a pairwise match
+            based on current rating of the image.
+
+        Returns
+        -------
+        new_image_rating : float
+            New rating of image after the classification match.
+        """
+        new_image_rating = rating_for_image + k_value * (score_for_image - image_expected_score)
+        return new_image_rating
+
+    def score_update(self, image_0, image_1, score_for_image_0):
+        """
+        Updates the ratings of the two images based on the complexity classification.
+
+        Parameters
+        ----------
+        image_0 : int
+            Image id for first image
+        image_1 : int
+            Image id for second image
+        score_for_image_0 : int
+            Actual result of classification of the image 0 in a pairwise match.
+            `0` denotes less complex, `1` denotes more complex
+
+        Notes
+        -----
+        To make updates in the original rankings DataFrame, for each classification,
+        two state dictionaries need to be maintained, corresponfing to the two AR images.
+        The changes are made to these state dictionaries and then the ranking DataFrame is updated.
+        """
+        # state dicts
+        state_dict_0 = self.rankings.loc[image_0].to_dict()
+        state_dict_0['last scores'] = deque(map(float, state_dict_0['last scores'].split(',')), maxlen=self.score_memory)
+        state_dict_1 = self.rankings.loc[image_1].to_dict()
+        state_dict_1['last scores'] = deque(map(float, state_dict_1['last scores'].split(',')), maxlen=self.score_memory)
+
+        expected_score_0 = self.expected_score(self.rankings.loc[image_0]['score'],
+                                               self.rankings.loc[image_1]['score'])
+        expected_score_1 = 1 - expected_score_0
+
+        new_rating_0 = self.new_rating(self.rankings.loc[image_0]['score'], self.rankings.loc[image_0]['k value'],
+                                       score_for_image_0, expected_score_0)
+        new_rating_1 = self.new_rating(self.rankings.loc[image_1]['score'], self.rankings.loc[image_1]['k value'],
+                                       1 - score_for_image_0, expected_score_1)
+
+        state_dict_0['last scores'].append(new_rating_0)
+        state_dict_1['last scores'].append(new_rating_1)
+
+        new_std_dev_0 = min(np.std(state_dict_0['last scores']), 1_000_000)  # prevents Infinity
+        new_k_0 = min(max(new_std_dev_0, self.score_change['min']), self.score_change['max'])
+
+        new_std_dev_1 = min(np.std(state_dict_1['last scores']), 1_000_000)  # prevents Infinity
+        new_k_1 = min(max(new_std_dev_1, self.score_change['min']), self.score_change['max'])
+
+        # Updating Data
+        state_dict_0['score'] = new_rating_0
+        state_dict_0['std dev'] = new_std_dev_0
+        state_dict_0['k value'] = new_k_0
+        state_dict_0['count'] += 1
+        state_dict_0['last scores'] = ",".join(map(str,state_dict_0['last scores']))  # Storing the list of states as a String
+
+        state_dict_1['score'] = new_rating_1
+        state_dict_1['std dev'] = new_std_dev_1
+        state_dict_1['k value'] = new_k_1
+        state_dict_1['count'] += 1
+        state_dict_1['last scores'] = ",".join(map(str,state_dict_1['last scores']))  # Storing the list of states as a String
+
+        # Making the Update DataFrames
+        update_df = pd.DataFrame([state_dict_0, state_dict_1])
+        update_df.set_index("player id", inplace=True)
+
+        # Updating the original DataFrame
+        self.rankings.update(update_df)
+
+    def run(self, save_to_disk=True, filename='run_results.csv'):
+        """
+        Runs the ELO ranking Algorithm for all score_board.
+
+        Parameters
+        ----------
+        save_to_disk : bool, optional
+            If true, saves the rankins in a CSV file on the disk, by default True
+        filename : str, optional
+            filename to store the results, by default 'run_results.csv'
+        """
+        for index, row in self.score_board.iterrows():
+
+            if row[self.column_map['player 0']] == row[self.column_map['player 1']]:
+                continue
+
+            self.score_update(image_0=row[self.column_map['player 0']], image_1=row[self.column_map['player 1']],
+                              score_for_image_0=row[self.column_map['score for player 0']])
+            print(f"Index {index} done!")
+
+        if save_to_disk:
+            self.save_as_csv(filename)
+
+    def save_as_csv(self, filename):
+        """
+        Saves the Ranking DataFrame to the disk as a CSV file.
+
+        Parameters
+        ----------
+        filename : str
+            filename to store the results.
+        """
+        self.rankings.drop(columns=["last_scores"], inplace=True)
+        self.rankings.to_csv(filename)

From bf01a172972198bb1a87dbeb312e2a3d6c7818a6 Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Sat, 13 Jun 2020 04:21:35 +0530
Subject: [PATCH 02/10] Adds ELO tests

---
 data/elo_test/test_classifications.csv    |  6 +++
 pythia/cleaning/tests/test_elo.py         | 66 +++++++++++++++++++++++
 pythia/seo/tests/test_classifications.csv |  6 +++
 pythia/seo/tests/test_rankings.csv        | 11 ++++
 4 files changed, 89 insertions(+)
 create mode 100644 data/elo_test/test_classifications.csv
 create mode 100644 pythia/cleaning/tests/test_elo.py
 create mode 100644 pythia/seo/tests/test_classifications.csv
 create mode 100644 pythia/seo/tests/test_rankings.csv

diff --git a/data/elo_test/test_classifications.csv b/data/elo_test/test_classifications.csv
new file mode 100644
index 0000000..934a11a
--- /dev/null
+++ b/data/elo_test/test_classifications.csv
@@ -0,0 +1,6 @@
+;image_id_0;image_id_1;image0_more_complex_image1
+0;1;6;1
+1;2;7;0
+2;3;8;0
+3;4;9;1
+4;5;10;1
diff --git a/pythia/cleaning/tests/test_elo.py b/pythia/cleaning/tests/test_elo.py
new file mode 100644
index 0000000..17c07ad
--- /dev/null
+++ b/pythia/cleaning/tests/test_elo.py
@@ -0,0 +1,66 @@
+import pytest
+import pandas as pd
+from pythia.cleaning import ELO
+from pathlib import Path
+from pythia.seo import Sunspotter
+from sunpy.util import SunpyUserWarning
+
+path = Path(__file__).resolve().parent.parent.parent.parent / "data/elo_test"
+
+
+@pytest.fixture
+def elo():
+    sunspotter = Sunspotter(timesfits=path / "../all_clear/lookup_timesfits.csv",
+                            properties=path / "../all_clear/lookup_properties.csv",
+                            classifications=path / "test_classifications.csv",
+                            classifications_columns=['image_id_0', 'image_id_1',
+                                                     'image0_more_complex_image1'])
+    column_map = {"player 0": "image_id_0",
+                  "player 1": "image_id_1",
+                  "score for player 0": "image0_more_complex_image1"}
+ 
+    return ELO(score_board=sunspotter.classifications, column_map=column_map)
+
+
+@pytest.mark.parametrize('rating_0,rating_1,expected_score',
+                         [(1400, 1400.0, 0.5),
+                          (1450, 1450.5, 0.49928044265518673),
+                          (1500, 1602.0, 0.3572869311673796),
+                          (1550, 1854.5, 0.14768898365874825),
+                          (1600, 2208.0, 0.029314241270450396)])
+def test_expected_score(elo, rating_0, rating_1, expected_score):
+    assert elo.expected_score(rating_0, rating_1) == expected_score
+
+
+@pytest.mark.parametrize('rating_for_image,k_value,score_for_image,image_expected_score,new_rating',
+                          [(1400.0, 32, 1, 0.5, 1416.0),
+                           (1450.0, 32, 0, 0.49928044265518673, 1434.023025835034),
+                           (1500.0, 32, 0, 0.3572869311673796, 1488.5668182026438),
+                           (1550.0, 32, 1, 0.14768898365874825, 1577.2739525229201),
+                           (1600.0, 32, 1, 0.029314241270450396, 1631.0619442793457),
+                           (1400.0, 32, 0, 0.5, 1384.0),
+                           (1450.5, 32, 1, 0.5007195573448133, 1466.476974164966),
+                           (1602.0, 32, 1, 0.6427130688326204, 1613.4331817973562),
+                           (1854.5, 32, 0, 0.8523110163412517, 1827.2260474770799),
+                           (2208.0, 32, 0, 0.9706857587295497, 2176.9380557206546)])
+def test_new_rating(elo, rating_for_image, k_value, score_for_image, image_expected_score, new_rating):
+    assert elo.new_rating(rating_for_image, k_value, score_for_image, image_expected_score) == new_rating
+
+
+def test_column_map(elo):
+    assert set(elo.column_map.values()).issubset(elo.score_board.columns)
+
+
+def test_incorrect_column_map():
+
+    sunspotter = Sunspotter(timesfits=path / "../all_clear/lookup_timesfits.csv",
+                            properties=path / "../all_clear/lookup_properties.csv",
+                            classifications=path / "test_classifications.csv",
+                            classifications_columns=['image_id_0', 'image_id_1',
+                                                     'image0_more_complex_image1'])
+    column_map = {"player 0": "This is not player 0",
+                  "player 1": "This is not player 1",
+                  "score for player 0": "Player 0 is in it for the fun"}
+
+    with pytest.raises(SunpyUserWarning):
+        ELO(score_board=sunspotter.classifications, column_map=column_map)
diff --git a/pythia/seo/tests/test_classifications.csv b/pythia/seo/tests/test_classifications.csv
new file mode 100644
index 0000000..e3e8fb4
--- /dev/null
+++ b/pythia/seo/tests/test_classifications.csv
@@ -0,0 +1,6 @@
+,image_id_0,image_id_1,image0_more_complex_image1
+0,1,6,1
+1,2,7,0
+2,3,8,0
+3,4,9,1
+4,5,10,1
diff --git a/pythia/seo/tests/test_rankings.csv b/pythia/seo/tests/test_rankings.csv
new file mode 100644
index 0000000..bf11d4d
--- /dev/null
+++ b/pythia/seo/tests/test_rankings.csv
@@ -0,0 +1,11 @@
+,image_id,score,k_value,count,std_dev
+0,1,1400.0,32,0,32
+1,2,1450.0,32,0,32
+2,3,1500.0,32,0,32
+3,4,1550.0,32,0,32
+4,5,1600.0,32,0,32
+5,6,1400.0,32,0,32
+6,7,1450.5,32,0,32
+7,8,1602.0,32,0,32
+8,9,1854.5,32,0,32
+9,10,2208.0,32,0,32

From 6b885d9556c1f5e465850094254cd5f04e99128c Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Sat, 13 Jun 2020 04:25:19 +0530
Subject: [PATCH 03/10] Updates cleaning __init__ file

---
 pythia/cleaning/__init__.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/pythia/cleaning/__init__.py b/pythia/cleaning/__init__.py
index e69de29..f782c15 100644
--- a/pythia/cleaning/__init__.py
+++ b/pythia/cleaning/__init__.py
@@ -0,0 +1 @@
+from pythia.cleaning.elo import *

From 7f7432ad8e8203a923b5fa6e4e4cc373b6d9c2ba Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Sat, 13 Jun 2020 04:28:32 +0530
Subject: [PATCH 04/10] Adds numpy to requirements

---
 requirements.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/requirements.txt b/requirements.txt
index fc6e807..a4837f4 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,2 +1,3 @@
 sunpy
 pandas
+numpy

From 2f99d8ac28303af7860c67b6420ad0ff4aba993a Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Sat, 13 Jun 2020 04:28:40 +0530
Subject: [PATCH 05/10] Adds changelog and solves pep8

---
 changelog/26.feature.rst                  |  1 +
 pythia/cleaning/elo.py                    |  4 ++--
 pythia/cleaning/tests/test_elo.py         | 22 +++++++++++-----------
 pythia/seo/tests/test_classifications.csv |  6 ------
 pythia/seo/tests/test_rankings.csv        | 11 -----------
 5 files changed, 14 insertions(+), 30 deletions(-)
 create mode 100644 changelog/26.feature.rst
 delete mode 100644 pythia/seo/tests/test_classifications.csv
 delete mode 100644 pythia/seo/tests/test_rankings.csv

diff --git a/changelog/26.feature.rst b/changelog/26.feature.rst
new file mode 100644
index 0000000..02e252d
--- /dev/null
+++ b/changelog/26.feature.rst
@@ -0,0 +1 @@
+Adds a module for ELO Ranking Algorithm.
diff --git a/pythia/cleaning/elo.py b/pythia/cleaning/elo.py
index 04610ec..5c67f2a 100644
--- a/pythia/cleaning/elo.py
+++ b/pythia/cleaning/elo.py
@@ -38,8 +38,8 @@ def __init__(self, score_board: pd.DataFrame, *, k_value=32, default_score=1400,
         column_map : dict, optional
             Dictionary, for mapping the column names of the score_board dataframe
             to variable names used in the ELO ranking system.
-            by default {"player 0": "image_id_0", 
-                                    "player 1": "image_id_1", 
+            by default {"player 0": "image_id_0",
+                                    "player 1": "image_id_1",
                                     "score for player 0": "image0_more_complex_image1"}
         """
         self.score_board = score_board
diff --git a/pythia/cleaning/tests/test_elo.py b/pythia/cleaning/tests/test_elo.py
index 17c07ad..857a171 100644
--- a/pythia/cleaning/tests/test_elo.py
+++ b/pythia/cleaning/tests/test_elo.py
@@ -18,7 +18,7 @@ def elo():
     column_map = {"player 0": "image_id_0",
                   "player 1": "image_id_1",
                   "score for player 0": "image0_more_complex_image1"}
- 
+
     return ELO(score_board=sunspotter.classifications, column_map=column_map)
 
 
@@ -33,16 +33,16 @@ def test_expected_score(elo, rating_0, rating_1, expected_score):
 
 
 @pytest.mark.parametrize('rating_for_image,k_value,score_for_image,image_expected_score,new_rating',
-                          [(1400.0, 32, 1, 0.5, 1416.0),
-                           (1450.0, 32, 0, 0.49928044265518673, 1434.023025835034),
-                           (1500.0, 32, 0, 0.3572869311673796, 1488.5668182026438),
-                           (1550.0, 32, 1, 0.14768898365874825, 1577.2739525229201),
-                           (1600.0, 32, 1, 0.029314241270450396, 1631.0619442793457),
-                           (1400.0, 32, 0, 0.5, 1384.0),
-                           (1450.5, 32, 1, 0.5007195573448133, 1466.476974164966),
-                           (1602.0, 32, 1, 0.6427130688326204, 1613.4331817973562),
-                           (1854.5, 32, 0, 0.8523110163412517, 1827.2260474770799),
-                           (2208.0, 32, 0, 0.9706857587295497, 2176.9380557206546)])
+                         [(1400.0, 32, 1, 0.5, 1416.0),
+                          (1450.0, 32, 0, 0.49928044265518673, 1434.023025835034),
+                          (1500.0, 32, 0, 0.3572869311673796, 1488.5668182026438),
+                          (1550.0, 32, 1, 0.14768898365874825, 1577.2739525229201),
+                          (1600.0, 32, 1, 0.029314241270450396, 1631.0619442793457),
+                          (1400.0, 32, 0, 0.5, 1384.0),
+                          (1450.5, 32, 1, 0.5007195573448133, 1466.476974164966),
+                          (1602.0, 32, 1, 0.6427130688326204, 1613.4331817973562),
+                          (1854.5, 32, 0, 0.8523110163412517, 1827.2260474770799),
+                          (2208.0, 32, 0, 0.9706857587295497, 2176.9380557206546)])
 def test_new_rating(elo, rating_for_image, k_value, score_for_image, image_expected_score, new_rating):
     assert elo.new_rating(rating_for_image, k_value, score_for_image, image_expected_score) == new_rating
 
diff --git a/pythia/seo/tests/test_classifications.csv b/pythia/seo/tests/test_classifications.csv
deleted file mode 100644
index e3e8fb4..0000000
--- a/pythia/seo/tests/test_classifications.csv
+++ /dev/null
@@ -1,6 +0,0 @@
-,image_id_0,image_id_1,image0_more_complex_image1
-0,1,6,1
-1,2,7,0
-2,3,8,0
-3,4,9,1
-4,5,10,1
diff --git a/pythia/seo/tests/test_rankings.csv b/pythia/seo/tests/test_rankings.csv
deleted file mode 100644
index bf11d4d..0000000
--- a/pythia/seo/tests/test_rankings.csv
+++ /dev/null
@@ -1,11 +0,0 @@
-,image_id,score,k_value,count,std_dev
-0,1,1400.0,32,0,32
-1,2,1450.0,32,0,32
-2,3,1500.0,32,0,32
-3,4,1550.0,32,0,32
-4,5,1600.0,32,0,32
-5,6,1400.0,32,0,32
-6,7,1450.5,32,0,32
-7,8,1602.0,32,0,32
-8,9,1854.5,32,0,32
-9,10,2208.0,32,0,32

From 384cd1a1aae13ee4c227a59461e12f5e7308e0b1 Mon Sep 17 00:00:00 2001
From: David Perez-Suarez <d.perez-suarez@ucl.ac.uk>
Date: Thu, 18 Jun 2020 21:07:30 +0100
Subject: [PATCH 06/10] =?UTF-8?q?=F0=9F=8C=9E=20drying=20the=20update=20of?=
 =?UTF-8?q?=20state=20dictionaries?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 pythia/cleaning/elo.py | 41 +++++++++++++++--------------------------
 1 file changed, 15 insertions(+), 26 deletions(-)

diff --git a/pythia/cleaning/elo.py b/pythia/cleaning/elo.py
index 5c67f2a..5621d6f 100644
--- a/pythia/cleaning/elo.py
+++ b/pythia/cleaning/elo.py
@@ -147,32 +147,8 @@ def score_update(self, image_0, image_1, score_for_image_0):
                                                self.rankings.loc[image_1]['score'])
         expected_score_1 = 1 - expected_score_0
 
-        new_rating_0 = self.new_rating(self.rankings.loc[image_0]['score'], self.rankings.loc[image_0]['k value'],
-                                       score_for_image_0, expected_score_0)
-        new_rating_1 = self.new_rating(self.rankings.loc[image_1]['score'], self.rankings.loc[image_1]['k value'],
-                                       1 - score_for_image_0, expected_score_1)
-
-        state_dict_0['last scores'].append(new_rating_0)
-        state_dict_1['last scores'].append(new_rating_1)
-
-        new_std_dev_0 = min(np.std(state_dict_0['last scores']), 1_000_000)  # prevents Infinity
-        new_k_0 = min(max(new_std_dev_0, self.score_change['min']), self.score_change['max'])
-
-        new_std_dev_1 = min(np.std(state_dict_1['last scores']), 1_000_000)  # prevents Infinity
-        new_k_1 = min(max(new_std_dev_1, self.score_change['min']), self.score_change['max'])
-
-        # Updating Data
-        state_dict_0['score'] = new_rating_0
-        state_dict_0['std dev'] = new_std_dev_0
-        state_dict_0['k value'] = new_k_0
-        state_dict_0['count'] += 1
-        state_dict_0['last scores'] = ",".join(map(str,state_dict_0['last scores']))  # Storing the list of states as a String
-
-        state_dict_1['score'] = new_rating_1
-        state_dict_1['std dev'] = new_std_dev_1
-        state_dict_1['k value'] = new_k_1
-        state_dict_1['count'] += 1
-        state_dict_1['last scores'] = ",".join(map(str,state_dict_1['last scores']))  # Storing the list of states as a String
+        _update_state_dict(state_dict_0, image_0, expected_score_0, score_for_image_0)
+        _update_state_dict(state_dict_1, image_1, expected_score_1, 1 - score_for_image_0)
 
         # Making the Update DataFrames
         update_df = pd.DataFrame([state_dict_0, state_dict_1])
@@ -181,6 +157,19 @@ def score_update(self, image_0, image_1, score_for_image_0):
         # Updating the original DataFrame
         self.rankings.update(update_df)
 
+    def _update_state_dict(state_dict, image, expected_score, score):
+        new_rating = self.new_rating(self.rankings.loc[image]['score'], self.rankings.loc[image]['k value'],
+                                       score, expected_score)
+        state_dict['last scores'].append(new_rating)
+        new_std_dev = min(np.std(state_dict['last scores']), 1_000_000)  # prevents Infinity
+        new_k = min(max(new_std_dev, self.score_change['min']), self.score_change['max'])
+        # Updating Data
+        state_dict['score'] = new_rating
+        state_dict['std dev'] = new_std_dev
+        state_dict['k value'] = new_k
+        state_dict['count'] += 1
+        state_dict['last scores'] = ",".join(map(str, state_dict['last scores']))  # Storing the list of states as a String
+
     def run(self, save_to_disk=True, filename='run_results.csv'):
         """
         Runs the ELO ranking Algorithm for all score_board.

From 061a6a104d9035b66afe153d2cdc42d01e02d979 Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Tue, 30 Jun 2020 15:53:55 +0530
Subject: [PATCH 07/10] codestyle

---
 pythia/cleaning/elo.py            | 6 +++---
 pythia/cleaning/tests/test_elo.py | 4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/pythia/cleaning/elo.py b/pythia/cleaning/elo.py
index 5c67f2a..c8adda4 100644
--- a/pythia/cleaning/elo.py
+++ b/pythia/cleaning/elo.py
@@ -1,8 +1,8 @@
-import pandas as pd
-import numpy as np
 from collections import deque
-from sunpy.util import SunpyUserWarning
 
+import numpy as np
+import pandas as pd
+from sunpy.util import SunpyUserWarning
 
 __all__ = ['ELO']
 
diff --git a/pythia/cleaning/tests/test_elo.py b/pythia/cleaning/tests/test_elo.py
index 857a171..24b6013 100644
--- a/pythia/cleaning/tests/test_elo.py
+++ b/pythia/cleaning/tests/test_elo.py
@@ -1,7 +1,7 @@
+from pathlib import Path
+
 import pytest
-import pandas as pd
 from pythia.cleaning import ELO
-from pathlib import Path
 from pythia.seo import Sunspotter
 from sunpy.util import SunpyUserWarning
 

From e54264138079b0850686193207402983a60507f3 Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Tue, 30 Jun 2020 16:00:56 +0530
Subject: [PATCH 08/10] Use pytest approx

---
 pythia/cleaning/tests/test_elo.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/pythia/cleaning/tests/test_elo.py b/pythia/cleaning/tests/test_elo.py
index 24b6013..f59cb0f 100644
--- a/pythia/cleaning/tests/test_elo.py
+++ b/pythia/cleaning/tests/test_elo.py
@@ -29,7 +29,7 @@ def elo():
                           (1550, 1854.5, 0.14768898365874825),
                           (1600, 2208.0, 0.029314241270450396)])
 def test_expected_score(elo, rating_0, rating_1, expected_score):
-    assert elo.expected_score(rating_0, rating_1) == expected_score
+    assert pytest.approx(elo.expected_score(rating_0, rating_1)) == expected_score
 
 
 @pytest.mark.parametrize('rating_for_image,k_value,score_for_image,image_expected_score,new_rating',
@@ -44,7 +44,7 @@ def test_expected_score(elo, rating_0, rating_1, expected_score):
                           (1854.5, 32, 0, 0.8523110163412517, 1827.2260474770799),
                           (2208.0, 32, 0, 0.9706857587295497, 2176.9380557206546)])
 def test_new_rating(elo, rating_for_image, k_value, score_for_image, image_expected_score, new_rating):
-    assert elo.new_rating(rating_for_image, k_value, score_for_image, image_expected_score) == new_rating
+    assert pytest.approx(elo.new_rating(rating_for_image, k_value, score_for_image, image_expected_score)), new_rating
 
 
 def test_column_map(elo):

From 7f827d59317eee91bb9695329284ca17ccacf151 Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahul.s18@iiits.in>
Date: Tue, 30 Jun 2020 16:11:02 +0530
Subject: [PATCH 09/10] Moves test file around

---
 .../cleaning/tests}/test_classifications.csv       |  0
 pythia/cleaning/tests/test_elo.py                  | 14 +++++++-------
 2 files changed, 7 insertions(+), 7 deletions(-)
 rename {data/elo_test => pythia/cleaning/tests}/test_classifications.csv (100%)

diff --git a/data/elo_test/test_classifications.csv b/pythia/cleaning/tests/test_classifications.csv
similarity index 100%
rename from data/elo_test/test_classifications.csv
rename to pythia/cleaning/tests/test_classifications.csv
diff --git a/pythia/cleaning/tests/test_elo.py b/pythia/cleaning/tests/test_elo.py
index f59cb0f..a3f8bc0 100644
--- a/pythia/cleaning/tests/test_elo.py
+++ b/pythia/cleaning/tests/test_elo.py
@@ -5,14 +5,14 @@
 from pythia.seo import Sunspotter
 from sunpy.util import SunpyUserWarning
 
-path = Path(__file__).resolve().parent.parent.parent.parent / "data/elo_test"
+path = Path(__file__).parent.parent.parent.parent / "data/all_clear"
 
 
 @pytest.fixture
 def elo():
-    sunspotter = Sunspotter(timesfits=path / "../all_clear/lookup_timesfits.csv",
-                            properties=path / "../all_clear/lookup_properties.csv",
-                            classifications=path / "test_classifications.csv",
+    sunspotter = Sunspotter(timesfits=path / "lookup_timesfits.csv",
+                            properties=path / "lookup_properties.csv",
+                            classifications= Path(__file__).parent / "test_classifications.csv",
                             classifications_columns=['image_id_0', 'image_id_1',
                                                      'image0_more_complex_image1'])
     column_map = {"player 0": "image_id_0",
@@ -53,9 +53,9 @@ def test_column_map(elo):
 
 def test_incorrect_column_map():
 
-    sunspotter = Sunspotter(timesfits=path / "../all_clear/lookup_timesfits.csv",
-                            properties=path / "../all_clear/lookup_properties.csv",
-                            classifications=path / "test_classifications.csv",
+    sunspotter = Sunspotter(timesfits=path / "lookup_timesfits.csv",
+                            properties=path / "lookup_properties.csv",
+                            classifications=Path(__file__).parent / "test_classifications.csv",
                             classifications_columns=['image_id_0', 'image_id_1',
                                                      'image0_more_complex_image1'])
     column_map = {"player 0": "This is not player 0",

From f7581c256d4435913e70215b97fc29a9369132cb Mon Sep 17 00:00:00 2001
From: Raahul Singh <raahulsingh002@gmail.com>
Date: Tue, 30 Jun 2020 16:17:53 +0530
Subject: [PATCH 10/10] Update pythia/cleaning/elo.py
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Co-authored-by: David Pérez-Suárez <dps.helio@gmail.com>
---
 pythia/cleaning/elo.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/pythia/cleaning/elo.py b/pythia/cleaning/elo.py
index 5621d6f..5c9f020 100644
--- a/pythia/cleaning/elo.py
+++ b/pythia/cleaning/elo.py
@@ -55,8 +55,7 @@ def __init__(self, score_board: pd.DataFrame, *, k_value=32, default_score=1400,
             missing_columns = ", ".join(missing_columns)
 
             raise SunpyUserWarning("The following columns mentioned in the column map"
-                                   " are not present in the score board: " +
-                                   missing_columns)
+                                   f" are not present in the score board: {missing_columns}")
 
         self._create_ranking()