| """Turn click cases into realistic DONE cases by actually executing the click in the browser. |
| |
| python finetune/make_done_cases.py out/pages.jsonl out/cases.jsonl out/done_cases.jsonl [max=250] |
| |
| For a click case (page A, goal, gold element): open A, click the gold element, observe the landing page B. |
| If the page changed, emit {goal, gold_op: DONE, page_obj: B, history: [that click]} with the same goal phrasing. |
| """ |
| import json, os, random, sys, time |
| sys.path.insert(0, "/home/ckl/projects/S/jev-ultrafast") |
| os.environ.setdefault("BU_CDP_URL", "http://127.0.0.1:9222") |
| from jev_ultrafast.browser import Browser |
|
|
| def main(): |
| pages_f, cases_f, out = sys.argv[1:4]; mx = int(sys.argv[4]) if len(sys.argv) > 4 else 250 |
| pages = [json.loads(l) for l in open(pages_f)] |
| cases = [c for c in (json.loads(l) for l in open(cases_f)) if c["kind"] == "click"] |
| rng = random.Random(3); rng.shuffle(cases) |
| n_ok = n_try = 0 |
| with open(out, "w") as f: |
| for c in cases: |
| if n_ok >= mx: break |
| src = pages[c["page"]]; n_try += 1 |
| try: |
| b = Browser(src["url"]) |
| try: |
| page = b.observe(screenshot=False) |
| act = next((a for a in page["actions"] if a["kind"] == "click" and a["label"] == c["label"]), None) |
| if act is None: |
| continue |
| b.act(act, page); time.sleep(0.3) |
| dest = b.observe(screenshot=False) |
| finally: |
| b.close() |
| except Exception as e: |
| print("fail", type(e).__name__, str(e)[:60], flush=True); continue |
| if dest["url"] == page["url"] and dest["title"] == page["title"]: |
| continue |
| hist = [{"action": act["label"], "kind": "click", "text": None, "page_changed": True}] |
| f.write(json.dumps({"page": c["page"], "url": dest["url"], "title": dest["title"], "goal": c["goal"], "gold_op": "DONE", |
| "gold_id": "DONE", "kind": "done", "label": act["label"], "history": hist, |
| "page_obj": {k: dest[k] for k in ("url", "title", "text", "actions")}}, ensure_ascii=False) + "\n"); f.flush() |
| n_ok += 1 |
| if n_ok % 25 == 0: print(f" {n_ok} done cases from {n_try} tries", flush=True) |
| print("wrote", n_ok, "DONE cases") |
|
|
| if __name__ == "__main__": |
| main() |
|
|
