# /// script
# requires-python = ">=3.12"
# dependencies = ["websockets>=13"]
# ///
"""Sample post texts from a jetstream-compatible feed into a local JSONL corpus.

usage: uv run scripts/sample_posts.py OUT.jsonl [COUNT]
"""

import asyncio
import json
import sys

import websockets

URL = "wss://stream.waow.tech/subscribe?wantedCollections=app.bsky.feed.post"


async def main(out_path: str, count: int) -> None:
    seen = 0
    with open(out_path, "w") as out:
        async with websockets.connect(URL, max_size=2**22) as ws:
            async for raw in ws:
                event = json.loads(raw)
                commit = event.get("commit") or {}
                if commit.get("operation") != "create":
                    continue
                text = (commit.get("record") or {}).get("text")
                if not text:
                    continue
                out.write(json.dumps({"text": text}, ensure_ascii=False) + "\n")
                seen += 1
                if seen >= count:
                    return


if __name__ == "__main__":
    asyncio.run(main(sys.argv[1], int(sys.argv[2]) if len(sys.argv) > 2 else 5000))