"""cutpreds.py : apply loopcut.cut_loops to recognizer outputs, keeping the raw output in pred_raw (the same transformation that produced preds/ocr_*_ocr4b_mixcut.jsonl).""" import json import sys from loopcut import cut_loops with open(sys.argv[2], "w") as f: for l in open(sys.argv[1]): r = json.loads(l) raw = r.get("pred_raw", r["pred"]) f.write(json.dumps(dict(r, pred=cut_loops(raw or ""), pred_raw=raw), ensure_ascii=False) + "\n")