No documentation to run from a pandas dataframe

[rabinadk1]

Pandas Dataframe is the most used tool to load CSVs. Please incorporate the documentation to calculate the reliability matrix from the data frames.

[grigsos]

I have managed to make below Dataframe, same as standardly used reliability_data in this project. (Ignore the decimals, it's necessary for me). Assertion will verify that those are the same if you want to check

I also replaced null values with "N/A", as I'm reading excels

pd.read_excel('file.xlsx', na_values='N/A')

df = pd.DataFrame({'rater1': ["N/A", "N/A", "N/A", "N/A", "N/A",3,4,1,2,1,1,3,3,"N/A",3],
                   'rater2': [1, "N/A", 2, 1, 3,3,4,3,"N/A","N/A","N/A","N/A","N/A","N/A","N/A"],
                   'rater3': ["N/A", "N/A", 2, 1, 3,4,4,"N/A",2,1,1,3,3,"N/A",4]})

data = df.T.values.tolist()

data_tuple = tuple(' '.join(map(str, row)) for row in data)

reliability_data_str = (
        "*    *    *    *    *    3    4    1    2    1    1    3    3    *    3",  # coder A
        "1    *    2    1    3    3    4    3    *    *    *    *    *    *    *",  # coder B
        "*    *    2    1    3    4    4    *    2    1    1    3    3    *    4",  # coder C
    )

print(reliability_data_str)
print(data_tuple)
newlistconvert =[[np.nan if (v == "*" or v=="N/A") else v for v in coder.split()] for coder in data_tuple]
reliability_data = [[np.nan if (v == "*" or v=="N/A") else v for v in coder.split()] for coder in reliability_data_str]

[grigsos]

My full implementation from a while back, with a redundant single rater logic that was fixed after I have wrote the code

def get_krippendorff_DF(inDataFrame):
    data = inDataFrame.T.values.tolist()
    data_tuple = tuple(' '.join(map(str, row)) for row in data)
    if len(data_tuple) == 1:
        return 1 #if only one coder, return 1
    else:
        newlistconvert = [[round(-(1/float(val))+2, 4) if isinstance(val, (int, float)) and float(val) < 1 else round(float(val), 4)
                           if val != "*" and val != "N/A" else np.nan for val in coder.split()] for coder in data_tuple] 
                            #convert into continous scale, instead of 0.25 and so on


        unique_values = list(set([val for sublist in newlistconvert for val in sublist]))
        unique_values.sort()
        
        return (krippendorff.alpha(reliability_data=newlistconvert, level_of_measurement="ordinal",value_domain=unique_values))  #calculate krippendorff alpha

[grigsos]

If you add this the below code into line 310 of krippendorff.krippendorff.py it will process it naturally

if type(reliability_data).__name__ == "DataFrame":
    data = reliability_data.T.values.tolist()
    data_tuple = tuple(' '.join(map(str, row)) for row in data)
    reliability_data = [[round(-(1/float(val))+2, 4) if isinstance(val, (int, float)) and float(val) < 1 else round(float(val), 4)
                    if val != "*" and val != "N/A" else np.nan for val in coder.split()] for coder in data_tuple] 
    value_domain = list(set([val for sublist in reliability_data for val in sublist])).sort()

it should look like this

if (reliability_data is None) == (value_counts is None):
    raise ValueError("Either reliability_data or value_counts must be provided, but not both.")

if type(reliability_data).__name__ == "DataFrame":
    data = reliability_data.T.values.tolist()
    data_tuple = tuple(' '.join(map(str, row)) for row in data)
    reliability_data = [[round(-(1/float(val))+2, 4) if isinstance(val, (int, float)) and float(val) < 1 else round(float(val), 4)
                    if val != "*" and val != "N/A" else np.nan for val in coder.split()] for coder in data_tuple] 
    value_domain = list(set([val for sublist in reliability_data for val in sublist])).sort()

# Don't know if it's a `list` or NumPy array. If it's the latter, the truth value is ambiguous. So, ask for `None`.
if value_counts is None:
    reliability_data = np.asarray(reliability_data)

Expected input, where df is what should be passed as reliability_data

import pandas as pd

data = {
    "coder A": ["*", "*", "*", "*", "*", "3", "4", "1", "2", "1", "1", "3", "3", "*", "3"],
    "coder B": ["1", "*", "2", "1", "3", "3", "4", "3", "*", "*", "*", "*", "*", "*", "*"],
    "coder C": ["*", "*", "2", "1", "3", "4", "4", "*", "2", "1", "1", "3", "3", "*", "4"]
}

df = pd.DataFrame(data)