import numpy as np from numba import jit, gdb_init import re

import pandas as pd import numpy as np import numba df=pd.read_csv('/kaggle/input/narrative/your_data.csv') df=df.head(10) #text=np.array(df['Narrative'],dtype=np.object_) text=np.array(df['Narrative'], dtype='S1000')

result_dtype = np.dtype([('keyword', 'S50'), ('matches', 'S50', (10,))])

@jit(nopython=False,debug=True) def find_pattern(text, keywords, pattern): text = text.lower() results = np.empty(len(keywords), dtype=result_dtype) result_idx = 0 for keyword in keywords: start_pos = 0 while True: keyword_pos = text.find(keyword, start_pos) if keyword_pos == -1: break # Start searching after the keyword start_search_pos = keyword_pos + len(keyword) subsequent_text = text[start_search_pos:start_search_pos + 50] matches = re.findall(pattern, subsequent_text) if matches: results[result_idx]['keyword'] = keyword results[result_idx]['matches'][:len(matches[:10])] = matches[:10] result_idx += 1 start_pos = keyword_pos + 1 # Move start_pos ahead to continue searching the text return results[:result_idx]

@jit(nopython=False,debug=True) def find_patterns_after_keywords(data, keywords, pattern): gdb_init() results = np.empty(len(data), dtype=result_dtype)

for idx in range(len(data)):
    text = str(data[idx])
    text = re.sub(r'[^\x00-\x7F]+', '', text)
    result = find_pattern(text, keywords, pattern)
    results[idx] = result
return results

pattern = r'\b\w+\s*/\s*\w+\b'

results = find_patterns_after_keywords(text, keywords, pattern) print(results)