breoganpardo / ie_pandas

This project contains the group assignment for Advanced Python, on which we have to create our own Python installable package for own created simplified "Pandas library".

IE Pandas

---

This is Team C's final project in Advanced Python.

A simple implementation of dataframe functionality

The library is available in Pypi

Installing

---

The easiest way to install ie_pandas is through pip

pip install ie_pandas

To use it in your project, you must first import the library

from ie_pandas import Dataframe

You can create a frame by the following 4 methods:

• A list of lists
• A numpy array of lists
• A dictionary of lists with keys being column names and values being the values for that column
• A dictionary of numpy arrays (same as with lists)

dictionary = {'c0': [1, 3, 5], 'c1': [7, 6, 2], 'c2': [2, 4, 7], 'c3': [5, 3, 9]}
df = DataFrame(dictionary)

Functionality

• Create dataframes from list of lists, numpy arrays, dictionaries of lists and numpy arrays

dictionary = {'c0': [1, 3, 5], 'c1': [7, 6, 2], 'c2': [2, 4, 7], 'c3': [5, 3, 9]}
df = DataFrame(dictionary)

# You may optionally pass along two parameters, cols and index
# cols determines the column names (if blank they will be numerical strings)
# index determines the row names (if blank they will be numbers)
df = DataFrame(dictionary, cols = ["col0", "col1", "col2", "col3"], index = ["row1", "row2", "row3"])

• Access columns by name

df['column_1']

• Access rows by position or by row name

df.get_index(1)
# or
df.get_index('row_1')

• Access data like a numpy array by name

df[0:2, 1:3]

• Modify dataframe

df[0,0] = 3

• Sum, median, mean, min, max methods (only work for numerical columns)

df.mean()

Since the underlying object of the dataframe is a numpy array, you may perform aditional functionality like

df[:, 1:2].sum()

• Visualize relationships between 2 entirely numerical columns (only for numerical columns)

df.visualize(df[:, 2], df[:, 3])
# or
df.visualize(df["c1"], df["c2"])

Dependencies

IE_Pandas only requires the following packages:

• Numpy (>=1.16)
• Matplotlib (>=3.0.2)

However, for development purposes, the following packages are needed:

• Pytest (>= 4.2)
• Pytest-cov (>= 2.6)
• Black (for PEP8 compliance)

Development

---

For development purposes, you may download the files directly and install the library locally by placing your terminal in the downloaded folder and doing

pip install --editable .[dev]

Then, to execute the tests you just need to run

pytest --cov

IE_Pandas Coding Style

---

IE_Pandas complies to PEP8 and uses black for coding standards

Versioning

---

SemVer is used for versioning.

License

---

This project is licensed under the MIT License - see the License file for details

ie_pandas's Issues

Add some tests

You are doing a great job with the pull requests and the reviews. However, without looking closely at the code it's clear that there's a lot of functionality, but it's not tested. I recommend you to

• Test what has already been implemented
• Do not include further functionality if it's not tested

Otherwise, it's very common to leave the tests for the end, and believe me - you will find unexpected problems!

Avoid using __getitem__ directly

You implemented the __getitem__ method for the DataFrame, however you don't need to use it: you can use the square brackets instead. For example:

>>> class Foo:
...   def __getitem__(self, item):
...     return f"This {item}"
... 
>>> f = Foo()
>>> f.__getitem__("bar")
'This bar'
>>> f["bar"]
'This bar'