Should I use classes for pandas.DataFrame? - pandas

I have more of a general question. I've written a couple of functions that transform data successively:
def func1(df):
pass
...
def main():
df = pd.read_csv()
df1 = func1(df)
df2 = func2(df1)
df3 = func3(df2)
df4 = func4(df3)
df4.to_csv()
if __name__ == "__main__":
main()
Is there a better way of organizing the logic of my script?
Should I use classes for cases like this when everything is tied to one dataset?

It depends of your usecase. For what I understand, I would use dictionary of your functions that process a df.
For instance:
function_returning_a_df = { "f1": func1, "f2": func2, "f3": func3}
df = pd.read_csv(csv)
if this df needs 3 functions to be applied
df_processing = ["f1","f2","f3"] #function will be applied in this order
# If you need to keep df at every step you can make a list
dfs_processed = []
for func in df_processing:
dfs_processed.append(df) # if you want to save all steps
df = function_returning_a_df[func](df)

Related

Refactoring code so I dont have to implement 100+ functions

I'm making a crypto scanner which has to scan 100+ different cryptocoins at the same time. Now I'm having a really hard time simplifying this code because if I don't I'm gonna end up with more than 100 functions for something really easy. I'll post down here what I'm trying to refactor.
def main():
twm = ThreadedWebsocketManager(api_key=api_key,api_secret=api_secret)
twm.start()
dic = {'close': [], 'low': [], 'high': []}
dic2 = {'close': [], 'low': [], 'high': []}
def handle_socket_message(msg):
candle = msg['k']
close_price = candle['c']
highest_price = candle['h']
lowest_price = candle['l']
status = candle['x']
if status:
dic['close'].append(close_price)
dic['low'].append(lowest_price)
dic['high'].append(highest_price)
df = pd.DataFrame(dic)
print(df)
def handle_socket_message2(msg):
candle = msg['k']
close_price = candle['c']
highest_price = candle['h']
lowest_price = candle['l']
status = candle['x']
if status:
dic2['close'].append(close_price)
dic2['low'].append(lowest_price)
dic2['high'].append(highest_price)
df = pd.DataFrame(dic2)
print(df)
twm.start_kline_socket(callback=handle_socket_message, symbol='ETHUSDT')
twm.start_kline_socket(callback=handle_socket_message2, symbol='BTCUSDT')
twm.join()
As you can see I getting live data from BTCUSDT and ETHUSDT. Now I append the close,low and high prices to a dictionary and then I make a DataFrame out of those dictionaries. I tried to do this with 1 dictionary and 1 handle_socket_message function. But then it merges the values of both cryptocoins into 1 dataframe which is not what I want. Does anyone know how I can refactor this piece of code? I was thinking about something with a loop but I can't figure it out myself.
If you have any questions, ask away! Thanks in advance!
I don't know exactly what you are trying to do, but the following code might get you started (basically use a dict of dicts):
twm = ThreadedWebsocketManager(api_key=api_key,api_secret=api_secret)
twm.start()
symbols = ['ETHUSDT', 'BTCUSDT']
symbolToMessageKeys = {
'close': 'c',
'high': 'h',
'low': 'l'
}
dictPerSymbol = dict()
for sym in symbols:
d = dict()
dictPerSymbol[sym] = d
for key in symbolToMessageKeys.keys():
d[key] = list()
print(dictPerSymbol)
def handle_socket_message(msg):
candle = msg['k']
if candle['x']:
d = dictPerSymbol[msg['s']]
for (symbolKey, msgKey) in symbolToMessageKeys.items():
d[symbolKey].append(candle[msgKey])
df = pd.DataFrame(d)
print(df)
for sym in symbols:
twm.start_kline_socket(callback=handle_socket_message, symbol=sym)
twm.join()
Luckily, appending to lists seems thread safe. Warning: if it is not, then we have a major race condition in the code of this answer. I should also note that I haven't used neither ThreadedWebsocketManagers nor DataFrames (so the latter may as well introduce thread safety issues if it is meant to write in the provided dictionary).

Copy/assign a Pandas dataframe based on their name in a for loop

I am relatively new with python - and I am struggling to do the following:
I have a set of different data frames, with sequential naming (df_i), which I want to access in a for loop based on their name (with an string), how can I do that? e.g.
df_1 = pd.read_csv('...')
df_2 = pd.read_csv('...')
df_3 = pd.read_csv('...')
....
n_df = 3
for i in range(len(n_df)):
df_namestr= 'df_' + str(i+1)
# ---------------------
df_temp = df_namestr
# ---------------------
# Operate with df_temp. For i+1= 1, df_temp should be df_1
Kind regards,
DF
You can try something like that:
for n in range(1, n_df+1):
df_namestr = f"df_{n}"
df_tmp = locals().get(df_namestr)
if not isinstance(df_tmp, pd.DataFrame):
continue
print(df_namestr)
print(df_tmp)
Refer to the documentation of locals() to know more.
Would it be better to approach the accessing of multiple dataframes by reading them into a list?
You could put all the csv files required in a subfolder and read them all in. Then they are in a list and you can access each one as an item in that list.
Example:
import pandas as pd
import glob
path = r'/Users/myUsername/Documents/subFolder'
csv_files = glob.glob(path + "/*.csv")
dfs = []
for filename in csv_files:
df = pd.read_csv(filename)
dfs.append(df)
print(len(dfs))
print(dfs[1].head())

What exactly to test (unittest) in a larger function containing several dataframe manipulations

Perhaps this is a constraint of my understanding of unittests, but I get quite confused as to what should be tested, patched, etc in a method that has several pandas dataframe manipulations. Many of the unittest examples out there focus on classes and methods that are typically small. For larger methods, I get a bit lost on the typical unittest paradigm. For example:
myscript.py
class Pivot:
def prepare_dfs(self):
df = pd.read_csv(self.file, sep=self.delimiter)
g = df.groupby("Other_Location")
df1 = g.apply(lambda x: x[x["PRN"] == "Free"].count())
locations = ["O12-03-01", "O12-03-02"]
cp = df1["PRN"]
cp = cp[locations].tolist()
data = [locations, cp]
new_df = pd.DataFrame({"Other_Location": data[0], "Free": data[1]})
return new_df, df
test_myscript.py
class TestPivot(unittest.TestCase):
def setUp(self):
args = parse_args(["-f", "test1", "-d", ","])
self.pivot = Pivot(args)
self.pivot.path = "Pivot/path"
#mock.patch("myscript.cp[locations].tolist()", return_value=None)
#mock.patch("myscript.pd.read_csv", return_value=df)
def test_prepare_dfs_1(self, mock_read_csv, mock_cp):
new_df, df = self.pivot.prepare_dfs()
# Here I get a bit lost
For example here I try to circumvent the following error message:
ModuleNotFoundError: No module named 'myscript.cp[locations]'; 'myscript' is not a package
I managed to mock correctly the pd.read_csv in my method, however further down in the code there are groupy, apply, tolist etc. The error message is thrown at the following line:
cp = cp[locations].tolist()
What is the best way to approach unittesting when your method involves several manipulations on a dataframe? Is refactoring the code always advised (into smaller chunks)? In this case, how can I mock correctly the tolist ?

How to merge 4 dataframes into one

I have created a functions that returns a dataframe.Now, i want merge all dataframe into one. First, i called all the function and used reduce and merge function.It did not work as expected.The error i am getting is "cannot combine function.It should be dataframe or series.I checked the type of my df,it is dataframe not functions. I don't know where the error is coming from.
def func1():
return df1
def func2():
return df2
def func3():
return df3
def func4():
return df4
def alldfs():
df_1 = func1()
df_2 = func2()
df_3 = func3()
df_4 = func4()
result = reduce(lambda df_1,d_2,df_3,df_4: pd.merge(df_1,df_2,df_3,df_4,on ="EMP_ID"),[df1,df2,df3,df4)
print(result)
You could try something like this ( assuming that EMP_ID is common across all dataframes and you want the intersection of all dataframes ) -
result = pd.merge(df1, df2, on='EMP_ID').merge(df3, on='EMP_ID').merge(df4, on='EMP_ID')

use Featureunion in scikit-learn to combine two pandas columns for tfidf

While using this as a model for spam classification, I'd like to add an additional feature of the Subject plus the body.
I have all of my features in a pandas dataframe. For example, the subject is df['Subject'], the body is df['body_text'] and the spam/ham label is df['ham/spam']
I receive the following error:
TypeError: 'FeatureUnion' object is not iterable
How can I use both df['Subject'] and df['body_text'] as features all while running them through the pipeline function?
from sklearn.pipeline import FeatureUnion
features = df[['Subject', 'body_text']].values
combined_2 = FeatureUnion(list(features))
pipeline = Pipeline([
('count_vectorizer', CountVectorizer(ngram_range=(1, 2))),
('tfidf_transformer', TfidfTransformer()),
('classifier', MultinomialNB())])
pipeline.fit(combined_2, df['ham/spam'])
k_fold = KFold(n=len(df), n_folds=6)
scores = []
confusion = numpy.array([[0, 0], [0, 0]])
for train_indices, test_indices in k_fold:
train_text = combined_2.iloc[train_indices]
train_y = df.iloc[test_indices]['ham/spam'].values
test_text = combined_2.iloc[test_indices]
test_y = df.iloc[test_indices]['ham/spam'].values
pipeline.fit(train_text, train_y)
predictions = pipeline.predict(test_text)
prediction_prob = pipeline.predict_proba(test_text)
confusion += confusion_matrix(test_y, predictions)
score = f1_score(test_y, predictions, pos_label='spam')
scores.append(score)
FeatureUnion was not meant to be used that way. It instead takes two feature extractors / vectorizers and applies them to the input. It does not take data in the constructor the way it is shown.
CountVectorizer is expecting a sequence of strings. The easiest way to provide it with that is to concatenate the strings together. That would pass both the text in both columns to the same CountVectorizer.
combined_2 = df['Subject'] + ' ' + df['body_text']
An alternative method would be to run CountVectorizer and optionally TfidfTransformer individually on each column, and then stack the results.
import scipy.sparse as sp
subject_vectorizer = CountVectorizer(...)
subject_vectors = subject_vectorizer.fit_transform(df['Subject'])
body_vectorizer = CountVectorizer(...)
body_vectors = body_vectorizer.fit_transform(df['body_text'])
combined_2 = sp.hstack([subject_vectors, body_vectors], format='csr')
A third option is to implement your own transformer that would extract a dataframe column.
class DataFrameColumnExtracter(TransformerMixin):
def __init__(self, column):
self.column = column
def fit(self, X, y=None):
return self
def transform(self, X, y=None):
return X[self.column]
In that case you could use FeatureUnion on two pipelines, each containing your custom transformer, then CountVectorizer.
subj_pipe = make_pipeline(
DataFrameColumnExtracter('Subject'),
CountVectorizer()
)
body_pipe = make_pipeline(
DataFrameColumnExtracter('body_text'),
CountVectorizer()
)
feature_union = make_union(subj_pipe, body_pipe)
This feature union of pipelines will take the dataframe and each pipeline will process its column. It will produce the concatenation of term count matrices from the two columns given.
sparse_matrix_of_counts = feature_union.fit_transform(df)
This feature union can also be added as the first step in a larger pipeline.