#!/data/data/com.termux/files/usr/bin/python3 -u
"""批量文章摘要 — 最多 2 篇并行，逐批完成"""
import os, requests, time, sys
from datetime import datetime

BASE = "http://127.0.0.1:5001"
BATCH_SIZE = 2
POLL_INTERVAL = 8   # 每 8 秒检查一次
JOB_TIMEOUT = 360    # 单个任务超时 6 分钟
LOG = "/data/data/com.termux/files/home/wechat-viewer/batch_summarize.log"

# Ensure HOME is set so hermes subprocess can find ~/pixelphone
if not os.environ.get("HOME"):
    os.environ["HOME"] = "/data/data/com.termux/files/home"
if not os.environ.get("HERMES_HOME"):
    os.environ["HERMES_HOME"] = os.path.join(os.environ["HOME"], "pixelphone")

def log(msg):
    ts = datetime.now().strftime("%H:%M:%S")
    line = f"[{ts}] {msg}"
    print(line, flush=True)
    with open(LOG, "a") as f:
        f.write(line + "\n")

def get_unsummarized(limit=200):
    r = requests.get(f"{BASE}/api/articles/unsummarized?limit={limit}", timeout=10)
    return r.json().get("data", [])

def submit_summary(article):
    try:
        r = requests.post(f"{BASE}/api/articles/summarize", json={
            "url": article["url"], "title": article["title"]
        }, timeout=10)
        j = r.json()
        if j.get("ok") and j.get("data", {}).get("status") == "started":
            return j["data"]["task_id"]
        if j.get("ok") and j.get("data", {}).get("status") == "cached":
            return "cached"
    except Exception as e:
        log(f"  提交异常: {e}")
    return None

def wait_for_job(task_id, timeout=JOB_TIMEOUT):
    start = time.time()
    while time.time() - start < timeout:
        try:
            r = requests.get(f"{BASE}/api/job/{task_id}", timeout=5)
            j = r.json()
            status = j.get("data", {}).get("status", "")
            if status == "completed":
                return True
            elif status == "failed":
                return False
        except:
            pass
        time.sleep(POLL_INTERVAL)
    return False

def main():
    articles = get_unsummarized()
    total = len(articles)
    if total == 0:
        log("没有待摘要的文章。")
        return
    
    log(f"开始批量摘要: 共 {total} 篇，每批 {BATCH_SIZE} 篇并行")
    done, failed, skipped = 0, 0, 0
    
    for i in range(0, total, BATCH_SIZE):
        batch = articles[i:i + BATCH_SIZE]
        batch_num = i // BATCH_SIZE + 1
        total_batches = (total + BATCH_SIZE - 1) // BATCH_SIZE
        
        # 提交本批次
        tasks = []
        for a in batch:
            tid = submit_summary(a)
            if tid == "cached":
                log(f"[{batch_num}/{total_batches}] 已有摘要跳过: {a['title'][:40]}")
                skipped += 1
            elif tid:
                tasks.append((tid, a["title"][:40]))
                log(f"[{batch_num}/{total_batches}] 已提交: {a['title'][:40]}")
            else:
                log(f"[{batch_num}/{total_batches}] 提交失败: {a['title'][:40]}")
                failed += 1
        
        # 等待本批次完成
        for tid, title in tasks:
            log(f"  等待: {title}...")
            ok = wait_for_job(tid)
            if ok:
                log(f"  ✅ 完成: {title}")
                done += 1
            else:
                log(f"  ❌ 超时/失败: {title}")
                failed += 1
        
        progress = done + failed + skipped
        log(f"  进度: {progress}/{total} (✅{done} ❌{failed} ⏭{skipped})")
    
    log(f"全部完成: ✅{done} ❌{failed} ⏭{skipped} / {total}")

if __name__ == "__main__":
    main()