deploy standalone ASR pipeline on tools-pve

2026-09-18 02:32:14 -07:00
parent f30f72b295
commit b23f00f41a
38 changed files with 182773 additions and 0 deletions
@@ -0,0 +1,157 @@
+#!/usr/bin/env python3
+"""Vault side effects: draft file, pending index, git sync, notification.
+
+Ported from the legacy shell implementation with its hard-won details preserved:
+  * a same date+title never overwrites an existing draft (suffix -2, -3, …)
+  * the pending index row is created BEFORE staging, so one commit carries both
+  * the 5-minute `obsidian-sync.sh` cron may have already staged the files; that is
+    tolerated rather than treated as an error
+  * a failed push is reported as "local only", never as success
+"""
+from __future__ import annotations
+
+import re
+import subprocess
+import tempfile
+from pathlib import Path
+
+from .config import CFG
+from .log import log
+
+PENDING_HEADER = """---
+type: index
+title: 待确认草稿
+note: 自动维护，确认定稿后请把对应行删除
+---
+
+# 待确认草稿 (ASR 自动精修)
+
+| 日期 | 标题 | 草稿 |
+|---|---|---|
+"""
+
+
+def slug_parts(slug: str) -> tuple[str, str]:
+    """`2026-07-21-125915录音` -> ('2026-07-21', '125915录音')."""
+    m = re.match(r"^(\d{4}-\d{2}-\d{2})[-_](.+)$", slug)
+    if m:
+        return m.group(1), m.group(2)
+    import time
+    return time.strftime("%Y-%m-%d"), slug
+
+
+def draft_path(date: str, title: str) -> Path:
+    CFG.drafts.mkdir(parents=True, exist_ok=True)
+    p = CFG.drafts / f"{date}-{title}.md"
+    if not p.exists():
+        return p
+    n = 2
+    while (CFG.drafts / f"{date}-{title}-{n}.md").exists():
+        n += 1
+    return CFG.drafts / f"{date}-{title}-{n}.md"
+
+
+def write_draft_file(path: Path, body: str) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    if not body.endswith("\n"):
+        body += "\n"
+    with tempfile.NamedTemporaryFile("w", encoding="utf-8", dir=path.parent,
+                                     prefix=f".{path.name}.", delete=False) as f:
+        f.write(body)
+        tmp = Path(f.name)
+    tmp.replace(path)
+
+
+def find_draft_for_transcript(transcript: str) -> Path | None:
+    """Recover an interrupted run without paying for or creating a second draft."""
+    if not CFG.drafts.exists():
+        return None
+    needle = f"transcript: {transcript}"
+    for p in CFG.drafts.glob("*.md"):
+        try:
+            if needle in p.read_text(errors="ignore")[:2000]:
+                return p
+        except OSError:
+            continue
+    return None
+
+
+def update_pending(date: str, title: str, path: Path) -> None:
+    p = CFG.pending_index
+    if not p.exists():
+        p.parent.mkdir(parents=True, exist_ok=True)
+        p.write_text(PENDING_HEADER, encoding="utf-8")
+    base = path.stem
+    text = p.read_text(errors="ignore")
+    if f"[[{base}]]" in text:
+        return
+    text += f"| {date} | {title} | [[{base}]] |\n"
+    write_draft_file(p, text)
+
+
+def _git(*args, check=False) -> subprocess.CompletedProcess:
+    return subprocess.run(["git", *args], cwd=CFG.vault, capture_output=True,
+                          text=True, timeout=180, check=check)
+
+
+def sync(paths: list[Path], message: str) -> str:
+    """Commit + push. Returns 'pushed' | 'local' | 'noop' | 'disabled' | 'error'."""
+    if not CFG.git_sync:
+        return "disabled"
+    try:
+        rel = [str(p.relative_to(CFG.vault)) for p in paths if p.exists()]
+        if not rel:
+            return "noop"
+        _git("add", "--", *rel)
+        if _git("diff", "--cached", "--quiet").returncode == 0:
+            log("nothing staged (likely already committed by the auto-sync cron)")
+            return "noop"
+        # Commit only our paths. The 5-minute vault sync may already have unrelated
+        # files staged; absorbing those into this transaction makes rollback and
+        # diagnosis needlessly difficult.
+        c = _git("commit", "-q", "--only", "-m", message, "--", *rel)
+        if c.returncode != 0:
+            log(f"git commit failed: {c.stderr.strip()[:200]}", level="ERROR")
+            return "error"
+        p = _git("push", "-q")
+        if p.returncode != 0:
+            log(f"committed but push failed (local only): {p.stderr.strip()[:200]}",
+                level="WARN")
+            return "local"
+        return "pushed"
+    except Exception as e:
+        log(f"git sync error: {e}", level="ERROR")
+        return "error"
+
+
+def notify(title: str, body: str, *, redact_log: bool = False) -> None:
+    if not CFG.notify:
+        return
+    n = CFG.notifier
+    if not n.exists():
+        log(f"NOTIFY (no notifier): {title} — {body}")
+        return
+    try:
+        r = subprocess.run(["python3", str(n), "草稿已生成，待确认", title, body],
+                           capture_output=True, text=True, timeout=30)
+        detail = "[sensitive body redacted]" if redact_log else body
+        log(f"NOTIFY: {title} — {detail} ({r.stdout.strip()[:80]})")
+    except Exception as e:
+        log(f"notify failed: {e}", level="WARN")
+
+
+def mark_done(slug: str) -> Path:
+    CFG.state.mkdir(parents=True, exist_ok=True)
+    m = CFG.state / f"{slug}.done"
+    m.touch()
+    return m
+
+
+def is_done(slug: str) -> bool:
+    return (CFG.state / f"{slug}.done").exists()
+
+
+def pending_transcripts() -> list[Path]:
+    if not CFG.inbox.exists():
+        return []
+    return sorted(p for p in CFG.inbox.glob("*.txt") if not is_done(p.stem))