Table Data for comparison:

Name	Caption
Text_1	Text 1
Text_2	Text 2

The Python library we employ offers various fuzzy search methods. The choice of a particular method is determined by the specific task or data type.

Let's look at each method in more detail:

1. fuzz.ratio compares strings character by character, assigning a score of 100 only for exact matches. Any difference, including punctuation, capitalization, or word order, reduces the score.

2. fuzz.token_sort_ratio breaks strings into words, ignores case, word order, and punctuation at the edges, and compares the sorted words. A higher score indicates greater similarity in word composition.

3. fuzz.token_set_ratio similar to token_sort_ratio, but removes duplicate words before comparison, focusing on the unique word set. This can be useful when comparing strings with repeated words.

4. fuzz.partial_ratio checks if one string is a substring of another, ignoring extra words. It evaluates the similarity of the core message, focusing on the essential content rather than exact word order or punctuation.

5. fuzz.WRatio combines multiple fuzziness methods (including ratio, token_sort_ratio, and partial_ratio) with weighted scores to provide a more comprehensive and human-like assessment of string similarity. It considers various factors to achieve a more accurate comparison.

To select a method, we must set the number of the necessary algorithm as the value for the fuzchanger variable. This can be done in the variable port settings of the Fuzzy Search supernode.

In the Python node configuration wizard, we manually set the following columns:

Name	Caption
Text_1	Text 1
Text_2	Text 2
fuzzy_search	Fuzzy Search Score

The checkbox Allow creating output columns in script is checked.

To access port data and other built-in objects in the context of code execution, the following are provided:

• Input datasets (InputTable)
• Input variables (InputVariables)
• Output dataset (OutputTable)
• Required enumerations (DataType)

The above objects are imported from the built-in module builtin_data. By default, an import line is added to the text of the code executed by the node.

from builtin_data import InputTable, InputVariables, OutputTable, DataType
from fuzzywuzzy import fuzz

We create a dictionary equipped with fuzzy search functionality:

fuzz_function_map = {     
    1: fuzz.ratio,
    2: fuzz.token_set_ratio,
    3: fuzz.token_sort_ratio,
    4: fuzz.partial_ratio,
    5: fuzz.WRatio
}

We assign a function, obtained from a dictionary key, to a variable:

fuz_value = fuzz_function_map.get(InputVariables.Items["fuzchanger"].Value, fuzz.ratio) 
     for i in range(InputTable.RowCount):
        text1 = InputTable.Get(i, 0)
        text2 = InputTable.Get(i, 1)
        OutputTable.Append()
        OutputTable.Set(0, text1)
        OutputTable.Set(1, text2)

We apply the fuzzy search algorithm to compare the columns:

OutputTable.Set(2, fuz_value(text1, text2))

Here is the result from the fuzz.ratio method: