Homework on Tic tac toe classification decision tree in python

Homework on Tic tac toe classification decision tree in python.
Note: You have to implement all the related functions by yourself such as the random partition into training and testing data, and the function calculating the information gain. You cannot use functions from other libraries such as sklearn's train test split. 

The homework is about the decision tree algorithm. we will use the Tic Tac Toe dataset tic-tac-toe.data (attached). It contains 958 instances with 2 classes and 9 features. Partition the data randomly into a training set (75% of the file) set and a training set (25% of the file)
Implement a decision tree classification algorithm in python. For attribute test conditions, use multibranch splitting of binary spliting. Use Information gain to measure node impurity. Use the training data to build a fully grown decision tree that is not pruned. Apply then the test data to determine the class labels and the generalization error rate of the test data. 

The output of the program should have:
1- Optimistic error rate of the training data
2- a generalization error rate of the test data. 

Deliverables:
The program source code that is well commented
3- A readme file.
4- a short report with the sample outputs

The program should be well commented on and done on Jupyter.

You have to implement all the related functions by yourself such as the random partition into training and testing data, and the function calculating the information gain. You cannot use functions from other libraries such as sklearn's train test split.