Keyword to label mapping for list column in pandas

I have a column named surface-wordings in my df named difference, which has values like this:

['paths-modified-/clusters/{cluster_id}/hosts/{host_id}/instructions-operations-modified-POST-parameters-modified-body-reply-schema-properties-modified-step_type-enum-deleted'] 

I have a keyword to label mappings defined, and from this column I want to extract the keywords and assign the label under a new column named labels. The mapping is as follows:

keyword_label_mappings = {
    'POST-parameters-modified': 'POST Parameters Modified',
    'PUT-parameters-modified': 'PUT Parameters Modified',
}

I am not sure how could I achieve this, any suggestions or ideas will be greatly appreciated.

I slightly modified your keyword_label_mappings dict to have an output with your second sample:

keyword_label_mappings = {
    'POST-parameters-modified': 'POST Parameters Modified',
    'PUT-parameters-modified': 'PUT Parameters Modified',
    'POST-responses-modified': 'POST Responses Modified',
    'DELETE-summary-from': 'DELETE Summary Changed',
    'POST-responses-deleted': 'POST Responses Deleted',
    'POST-parameters-added': 'POST Parameters Added',
    'POST-parameters-deleted': 'POST Parameters Deleted',
    'GET-summary-to': 'GET Summary To',  # added for demo
    'GET-summary-from': 'GET Summary From',  # added for demo
}

Use str.extract to extract keys of your dict then map to replace with values:

pattern = fr"({'|'.join(re.escape(k) for k in keyword_label_mappings)})"

difference['labels'] = (
    difference['surface_wordings'].explode().str.extractall(pattern)[0]
                                  .map(keyword_label_mappings).droplevel('match')
                                  .groupby(level=0).agg(list)
)

Output:

>>> difference
                                                                      surface_wordings                              labels
63657  [paths-modified-/pets-operations-modified-GET-summary-from, paths-modified-/...  [GET Summary From, GET Summary To]
63658  [info-version-from, info-version-to, paths-modified-/pets-operations-modifie...  [GET Summary From, GET Summary To]
63659  [paths-modified-/pets-operations-modified-GET-summary-from, paths-modified-/...  [GET Summary From, GET Summary To]
63661  [info-title-from, info-title-to, info-license-deleted, info-version-from, in...  [GET Summary From, GET Summary To]
63662  [openAPI-from, openAPI-to, paths-added, paths-deleted, endpoints-added, endp...                                 NaN

Lets go over your algorithm, for example we have one row and the value of the column surface-wordings is:

[paths-modified-/clusters/{cluster_id}/hosts/{host_id}/instructions-operations-modified-POST-parameters-modified-body-reply-schema-properties-modified-step_type-enum-deleted] 

after this line for keyword_list in difference['surface_wordings']: we will have mentioned value inside keyword_list

After that the line for keyword in keyword_list.split(', '): will return the same string because there is no ", " inside the mentioned row, so keyword is the same as keyword list.

This line keyword = keyword.split('-')[0] will return the word [paths because the entire thing you wrote was string I suppose. So this new word is definitely not matching your keys in the keyword_label_mappings dictionary and it will return [] at the end.

First of all try debugging your code, if you are using debugging tools better and if not use print statements to see the values of splits and etc. Also you have to check all the values and should try to find some pattern between data in order to find better algorithm of splitting and getting the key you want for checking inside the dictionary.

Also it would be better to provide difference.head() to get better idea of your df if this one still does not work.