UFCData 0.7.0__tar.gz → 0.7.1__tar.gz

{ufcdata-0.7.0 → ufcdata-0.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: UFCData
-Version: 0.7.0
+Version: 0.7.1
 Summary: Obtain UFC data and functions to manipulate it
 Requires-Python: >=3.10
 License-File: LICENSE.md

{ufcdata-0.7.0 → ufcdata-0.7.1}/README.md

@@ -1,11 +1,14 @@
 # UFCData
 
+
 An open-source Python package for accessing, manipulating, and analyzing UFC data.
 
 https://pypi.org/project/UFCData/
 
 ## Table of Contents
 
+
+
 - [Installation](#installation)
 - [Updating the Package](#updating-the-package)
 
@@ -30,7 +33,7 @@ Model Evaluation
 - [Train Test Split](#train-test-split)
 - [Get Expanding Window](#get-expanding-window)
 - [Plotting Accuracy](#plotting-accuracy)
-- [Expanding Window Baseline Example](#expanding-window-cv-and-final-test-example)
+- [Expanding Window Baseline Example](#expanding-window-baseline-example)
 
 
 Data Sources
@@ -38,6 +41,12 @@ Data Sources
   - [Discrepancies in Data](#discrepancies-in-data)
   - [Update Frequency](#update-frequency)
 
+<br>
+
+---
+
+<br>
+
 ## Installation
 
 ```bash
@@ -50,18 +59,23 @@ pip install ufcdata
 pip install --no-cache-dir --upgrade ufcdata
 ```
 
+---
+
+<br>
+
 ##  Loading Data
 
 ```python
 import ufcdata as ufc
 data = ufc.get_data()
 
-## Obtain fighter_bio dataframe
 fighter_bio = data["fighter_bio"].copy()
 ```
 
 Note that data in here and other functions below reference data that is available prior to the data.
 
+<br>
+
 ## Search
 
 Due to various event naming conventions and fighters sharing the same name, the primary keys for the dataframes are links, which can be difficult for humans to interpret. The `search()` function uses fuzzy string matching to make it easier to find fighters, events, and other records.
@@ -77,28 +91,19 @@ search(data, column, query, matches=1)
 * `query` — The name or text you are searching for.
 * `matches` — The number of closest matches to return. Defaults to `1`.
 
+### Returns
+The results are returned as a dataframe, with the closest match appearing first.
+
 ### Example
 
 ```python
 john_jones = ufc.search(fighter_bio, "Name", "Jon Jones", 3)
 ```
 
-This searches the `"Name"` column of the `fighter_bio` dataframe for the three closest matches to `"Jon Jones"`.
 
-The results are returned as a dataframe, with the closest match appearing first.
-
-### Returns
-
-
-```text
-Name             Nickname   Height   Weight   Reach   Stance      Total W   Total L   Fighter Link
-Jon Jones        Bones      76.0     248.0    84.0    Orthodox    28        1         ufcstats.com/...
-Roshaun Jones    NaN        68.0     135.0    NaN     NaN         2         6         ufcstats.com/...
-Mason Jones      The Dragon 70.0     155.0    74.0    Orthodox    18        2         ufcstats.com/...
-```
-
-This is useful when the exact value stored in the dataset is unknown or when there are multiple similar names. The same function can be used with any dataframe and column containing searchable text.
+---
 
+<br>
 
 ## Elo
 
@@ -123,8 +128,8 @@ ufc.get_elo(fight_df, fighter_df, r=1500, k=30, s=400)
 
 The function returns a tuple containing:
 
-1. `fighter_elo` — A dictionary mapping each fighter's UFCStats link to their final Elo rating.
-2. `elo` — The original `fights_df` dataframe with `"Fighter 1 Elo"` and `"Fighter 2 Elo"` columns added. These columns contain each fighter's Elo rating immediately prior to the corresponding fight.
+1. `current_elo` — A dictionary mapping each fighter's UFCStats link to their final Elo rating.
+2. `past_elo` — The original `fights_df` dataframe with `"Fighter 1 Elo"` and `"Fighter 2 Elo"` columns added. These columns contain each fighter's Elo rating immediately prior to the corresponding fight.
 
 
 ### Example
@@ -143,13 +148,17 @@ current_elo, past_elo = ufc.get_elo(
 )
 ```
 
-The resulting `fight_elo` dataframe can then be used to examine or incorporate pre-fight Elo ratings into analysis and predictive models.
-
 
 ### Important
 
 `fight_df` should be sorted from most recent to least recent before being passed to `get_elo()`. The function reverses the dataframe internally to process fights chronologically and returns the resulting data in the original order.
 
+<br>
+
+---
+
+<br>
+
 ## Fighter History
 
 The function `get_fighter_history()` retrieves all UFC fights for a specific fighter and formats the results from the fighter's perspective. The requested fighter is always represented as `"Fighter 1"`, regardless of which side of the original fight dataframe they appeared on.
@@ -198,6 +207,8 @@ jon_jones_history = ufc.get_fighter_history(
 
 This function is useful for analyzing an individual fighter's career, constructing fighter-level features, or preparing historical data for predictive modeling.
 
+<br>
+
 ## Fighter Statistic
 
 
@@ -235,11 +246,7 @@ get_fighter_statistic(
 
 ### Returns
 
-The function returns two DataFrames:
-
-```python
-current_statistic, past_statistic = get_fighter_statistic(...)
-```
+The function returns a tuple containing:
 
 #### `current_statistic`
 
@@ -303,9 +310,13 @@ jon_jones_current, jon_jones_past = ufc.get_fighter_statistic(
     fights_df,
     rounds_df
 )
-
-jon_jones_current
 ```
+
+
+---
+
+<br>
+
 ## Odds Convert
 
 The function `convert_odds()` converts American betting odds from strings into integers or floating point numbers.
@@ -344,10 +355,18 @@ data = ufc.get_data()
 
 fights_df = data["past_fights"].copy()
 
-fights_df["Fighter 1 Odds"] = fights_df["Fighter 1 Odds"].apply(ufc.convert_odds)
-fights_df["Fighter 1 Probability"] = fights_df["Fighter 1 Odds"].apply(ufc.convert_odds, probability=True)
-fights_df
+fights_df["Fighter 1 Odds"] = fights_df["Fighter 1 Odds"].apply(
+    ufc.convert_odds
+)
+
+fights_df["Fighter 1 Probability"] = fights_df["Fighter 1 Odds"].apply(
+    ufc.convert_odds,
+    probability=True
+)
 ```
+
+<br>
+
 ## Gender Convert
 
 The function `convert_gender()` converts gender labels into binary numeric values, with `"Male"` represented as `1` and `"Female"` represented as `0`.
@@ -375,9 +394,10 @@ data = ufc.get_data()
 
 fights_df = data["past_fights"].copy()
 fights_df["Gender"] = fights_df["Gender"].apply(ufc.convert_gender)
-fights_df
 ```
 
+<br>
+
 ## Weight Convert
 
 The function `convert_weight()` converts UFC weight class labels into their corresponding weight limits in pounds.
@@ -418,10 +438,10 @@ data = ufc.get_data()
 
 past_fights = data["past_fights"].copy()
 past_fights["Weight Class"] = past_fights["Weight Class"].apply(ufc.convert_weight)
-
-past_fights
 ```
 
+<br>
+
 ## Mirror
 
 The function `mirror()` creates mirrored versions of a DataFrame by swapping corresponding pairs of columns. This is useful when analyzing UFC fights from both fighters' perspectives.
@@ -453,7 +473,16 @@ import ufcdata as ufc
 data = ufc.get_data()
 
 fights_df = data["past_fights"].copy()
-fights_df = fights_df[["Date", "Event Link", "Fighter 1", "Fighter 1 Odds","Fighter 2", "Fighter 2 Odds"]]
+fights_df = fights_df[
+    [
+        "Date",
+        "Event Link",
+        "Fighter 1",
+        "Fighter 1 Odds",
+        "Fighter 2",
+        "Fighter 2 Odds",
+    ]
+]
 
 cols_1 = [
     "Fighter 1",
@@ -470,8 +499,6 @@ mirrored_fights = ufc.mirror(
     cols_1,
     cols_2
 )
-
-mirrored_fights
 ```
 
 Using `in_order=True` produces rows in the following order:
@@ -483,7 +510,7 @@ Original Fight 2
 Mirrored Fight 2
 ...
 ```
-Using `in_order=True` produces rows in the following order:
+Using `in_order=False` produces rows in the following order:
 ```text
 Original Fight 1
 Original Fight 2
@@ -494,6 +521,12 @@ Mirrored Fight 2
 
 This is useful when creating fighter-level datasets where each fight should be represented from both fighters' perspectives.
 
+<br>
+
+---
+
+<br>
+
 ## Train Test Split
 The function `train_test_split()` splits a UFC fight DataFrame into training and test sets while preserving the chronological order of the data. Unlike `sklearn.model_selection.train_test_split()`, this function does not randomly shuffle the data.
 
@@ -518,9 +551,10 @@ data = ufc.get_data()
 
 past_fights = data["past_fights"].copy()
 train, test = ufc.train_test_split(past_fights, 0.2)
-train
 ```
 
+<br>
+
 ## Get Expanding Window
 The function `get_expanding_window()` creates multiple training and test sets for time-series cross-validation using an expanding training window.
 
@@ -567,6 +601,7 @@ train, test = ufc.train_test_split(past_fights, 0.2)
 folds = ufc.get_expanding_window(train, 10, 500)
 folds['Fold 2']['test']
 ```
+<br>
 
 ## Plotting Accuracy
 
@@ -593,8 +628,127 @@ plot_test(accuracy, title="Test Accuracy")
 * `accuracy` — The accuracy.
 * `title` — The title of the plot.
 
+<br>
+
 ## Expanding Window Baseline Example
 
+In this example, we split the data into 80% for cross-validation and 20% for final testing. The cross-validation set is evaluated using a 15-fold expanding window approach. Using only betting odds, we convert the odds into implied probabilities and predict the fighter with the higher probability as the winner. The example also demonstrates how the model can be deployed.
+
+<br>
+
+```python
+import ufcdata as ufc
+import numpy as np
+
+data = ufc.get_data()
+
+## Select rows where outcome is W or L and both fighters have odds associated with the fights.
+past_fights = data["past_fights"].copy()
+past_fights = past_fights[~past_fights["Fighter 1 Outcome"].isin(["NC", "D"])]
+past_fights = past_fights[past_fights[["Fighter 1 Odds", "Fighter 2 Odds"]].notna().all(axis=1)]
+
+## Convert odds to probabilities
+past_fights["Fighter 1 Probability"] = (
+    past_fights["Fighter 1 Odds"]
+    .apply(ufc.convert_odds, probability=True)
+)
+
+past_fights["Fighter 2 Probability"] = (
+    past_fights["Fighter 2 Odds"]
+    .apply(ufc.convert_odds, probability=True)
+)
+
+## Create the data for cross validation and final testing
+folds = 15
+fold_names = [f"Fold {i}" for i in range(1, folds + 1)]
+cross_validation, final_test = ufc.train_test_split(past_fights, 0.2)
+fifteen_fold_cv = ufc.get_expanding_window(cross_validation, folds, 500)
+
+fold_accuracy = []
+## Loop through each fold to get the accuracy
+for fold in fold_names:
+  test = fifteen_fold_cv[fold]['test']
+  test["Predicted Winner"] = np.where(
+    test["Fighter 1 Probability"] > test["Fighter 2 Probability"],
+    test["Fighter 1 Link"],
+    test["Fighter 2 Link"])
+  accuracy = float((test["Predicted Winner"] == test["Winner"]).mean())
+  fold_accuracy.append(accuracy)
+
+## Plot cross validation accuracy
+ufc.plot_cv(fold_names, fold_accuracy)
+```
+<br>
+
+![Cross Validation Plot](https://i.imgur.com/sjCpnRE.png)
+
+<br>
+
+```python
+## Evaluate on final test set
+final_test["Predicted Winner"] = np.where(
+    final_test["Fighter 1 Probability"] > final_test["Fighter 2 Probability"],
+    final_test["Fighter 1 Link"],
+    final_test["Fighter 2 Link"])
+test_accuracy = float((final_test["Predicted Winner"] == final_test["Winner"]).mean())
+ufc.plot_test(test_accuracy)
+```
+
+<br>
+
+![Test Plot](https://i.imgur.com/5dHR59X.png)
+
+<br>
+
+```python
+## Deploy the model
+future_fights = data["future_fights"].copy()
+
+future_fights["Fighter 1 Probability"] = (
+    future_fights["Fighter 1 Odds"]
+    .apply(ufc.convert_odds, probability=True)
+)
+
+future_fights["Fighter 2 Probability"] = (
+    future_fights["Fighter 2 Odds"]
+    .apply(ufc.convert_odds, probability=True)
+)
+
+future_fights["Predicted Winner"] = np.where(
+    future_fights["Fighter 1 Probability"] > future_fights["Fighter 2 Probability"],
+    future_fights["Fighter 1 Link"],
+    future_fights["Fighter 2 Link"])
+
+future_fights["Predicted Winner"] = np.where(
+    future_fights["Predicted Winner"] == future_fights["Fighter 1 Link"],
+    future_fights["Fighter 1"],
+    future_fights["Fighter 2"]
+)
+
+future_fights = future_fights[
+    [
+        "Date",
+        "Event Link",
+        "Fight Link",
+        "Fighter 1",
+        "Fighter 2",
+        "Predicted Winner"
+    ]
+]
+
+future_fights
+
+```
+<br>
+
+![Deployed Predictions](https://i.imgur.com/jgZzNOV.png)
+
+<br>
+
+---
+
+<br>
+
 ## Online Sources
 
 Odds and birthplace data was obtained from https://www.tapology.com
@@ -603,12 +757,16 @@ Venue and attendance data was obtained from https://en.wikipedia.org/wiki/List_o
 
 All other data was obtained from http://ufcstats.com
 
+<br>
+
 ## Discrepancies in Data
 
 UFCStats is treated as the authoritative source for UFCData. When discrepancies exist between UFCStats and other sources, such as Wikipedia or Tapology, the UFCStats data is used.
 
 The UFCStats completed events page serves as the authoritative source for event, fight, and round data. Individual fighter profiles may contain fights from organizations or events that are not included in the completed events database, including WEC, Strikeforce, and PRIDE. These events are therefore excluded from UFCData.
 
+<br>
+
 ## Update Frequency
 
 The dataset is updated at the start of the scheduled broadcast time for each event. This update captures changes to betting odds, as well as any cancelled, postponed, or otherwise modified fights.

{ufcdata-0.7.0 → ufcdata-0.7.1}/UFCData.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: UFCData
-Version: 0.7.0
+Version: 0.7.1
 Summary: Obtain UFC data and functions to manipulate it
 Requires-Python: >=3.10
 License-File: LICENSE.md

{ufcdata-0.7.0 → ufcdata-0.7.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "UFCData"
-version = "0.7.0"
+version = "0.7.1"
 description = "Obtain UFC data and functions to manipulate it"
 requires-python = ">=3.10"
 

{ufcdata-0.7.0 → ufcdata-0.7.1}/ufcdata/fighter.py

@@ -181,7 +181,7 @@ def get_fighter_statistic(fighter_link, fighter_bio, fights_df, rounds_df,
 
     record = get_fighter_history(fighter_link, fighter_bio, fights_df)
 
-    date = ("2001-01-01")
+    date = ("2201-01-01")
 
     if len(record) == 0:
         columns = [