| import os | |
| import json | |
| import spacy | |
| TARGET_POS = [ | |
| "NOUN", | |
| "VERB", | |
| "ADJ", | |
| "ADV" | |
| ] | |
| def do_frameid(): | |
| nlp = spacy.load("it_core_news_md") | |
| with open("data/migration/corpus_titoli_all_raw.txt", encoding="utf-8") as f_in, \ | |
| open("output/migration/pos_based_targetid/corpus_titoli_all_raw.jsonl", "w", encoding="utf-8") as f_out: | |
| for line in f_in: | |
| doc = nlp(line.strip()) | |
| out = { | |
| "tokens": [t.text for t in doc], | |
| "predicates": [i for i, t in enumerate(doc) if t.pos_ in TARGET_POS] | |
| } | |
| f_out.write(json.dumps(out) + os.linesep) | |
| if __name__ == "__main__": | |
| do_frameid() | |