From e38ff4c7b05e05145f699c727a80dd3ca4ec7412 Mon Sep 17 00:00:00 2001 From: Luka Date: Fri, 21 Aug 2020 15:05:30 +0200 Subject: [PATCH] Added limit to minimum frequency = 10 + Ordered by frequency --- scripts/recalculate_statistics.py | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/scripts/recalculate_statistics.py b/scripts/recalculate_statistics.py index 4239ab7..12c434a 100644 --- a/scripts/recalculate_statistics.py +++ b/scripts/recalculate_statistics.py @@ -174,6 +174,12 @@ def main(args): write_file_path = os.path.join(args.output, file_name) with open(read_file_path, 'r') as rf, open(write_file_path, 'w') as wf: original_text, stats = get_new_stats(rf) + freq_pos = original_text[0].index('Frequency') + original_text = [original_text[0]] + [l for l in original_text[1:] if int(l[freq_pos]) >= 10] + if len(original_text) > 1: + original_text = [original_text[0]] + sorted(original_text[1:], key=lambda x: -1 * int(x[freq_pos])) + else: + original_text = [original_text[0]] write_new_stats(wf, original_text, stats, file_name, word_order) if __name__ == '__main__':