Refactor Google Meet Transcripts Extension for Local Use
- Removed all cloud-related functionalities, including login prompts and token handling. - Disabled Laxis cloud features, ensuring no data is sent to external servers. - Updated manifest to reflect the new local-only functionality. - Added a new Python server to handle transcripts locally, including WebSocket support. - Implemented storage management for transcripts, including deduplication and file writing. - Created a smoke test for the WebSocket server to simulate transcript updates. - Updated README with setup instructions and usage details for the new local server.
This commit is contained in:
vahidaskari
@@ -0,0 +1,234 @@
|
||||
"""
|
||||
storage.py — مدل داده، dedup، و خواندن/نوشتن transcript روی دیسک.
|
||||
|
||||
این ماژول هیچ I/O شبکهای ندارد؛ فقط منطقِ segment و فایلها:
|
||||
- segmentهای هر جلسه را نگه میدارد (با dedup بر اساس startedAt)
|
||||
- برای هر جلسه سه فایل مینویسد: <sid>.srt / <sid>.txt / <sid>.json
|
||||
- توابع خواندن برای لایهی MCP فراهم میکند
|
||||
|
||||
چرا segmentمحور با dedup؟ افزونه برای هر «حرف» چند بار پیام میفرستد و هر بار متن
|
||||
کمی بلندتر میشود (snapshotِ روبهرشد) ولی startedAt ثابت میماند. پس startedAt کلیدِ
|
||||
هویتِ segment است: تکرار شد → همان segment بهروز میشود (نه append). این هم تکرار را
|
||||
حذف میکند و هم چون startedAt/endedAt داریم، زمانبندیِ SRT دقیق میشود.
|
||||
"""
|
||||
|
||||
import json
|
||||
import re
|
||||
import sys
|
||||
import time
|
||||
from pathlib import Path
|
||||
|
||||
TRANSCRIPTS_DIR = Path(__file__).parent / "transcripts"
|
||||
TRANSCRIPTS_DIR.mkdir(exist_ok=True)
|
||||
|
||||
# segmentهای هر جلسه در حافظه: { sid: [ {seq, speaker, text, start, end, start_key}, ... ] }
|
||||
sessions_segments: "dict[str, list[dict]]" = {}
|
||||
|
||||
|
||||
def log(*args) -> None:
|
||||
"""همهی لاگها به stderr میروند؛ stdout برای پروتکل MCP رزرو است."""
|
||||
print(*args, file=sys.stderr, flush=True)
|
||||
|
||||
|
||||
def safe_session_id(session_id: str) -> str:
|
||||
cleaned = re.sub(r"[^a-zA-Z0-9_-]", "", session_id or "")
|
||||
return cleaned or "default"
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# زمان و رندر
|
||||
# ---------------------------------------------------------------------------
|
||||
def to_epoch_s(v) -> "float | None":
|
||||
"""startedAt/endedAt را به ثانیهی epoch تبدیل میکند (epoch ms، رشتهی عددی یا ISO)."""
|
||||
if v is None or isinstance(v, bool):
|
||||
return None
|
||||
if isinstance(v, (int, float)):
|
||||
return v / 1000.0 if v > 1e11 else float(v)
|
||||
if isinstance(v, str):
|
||||
s = v.strip()
|
||||
if not s:
|
||||
return None
|
||||
if s.isdigit():
|
||||
n = float(s)
|
||||
return n / 1000.0 if n > 1e11 else n
|
||||
try:
|
||||
from datetime import datetime
|
||||
return datetime.fromisoformat(s.replace("Z", "+00:00")).timestamp()
|
||||
except ValueError:
|
||||
return None
|
||||
return None
|
||||
|
||||
|
||||
def is_continuation(prev: str, cur: str) -> bool:
|
||||
"""fallback وقتی timestamp نداریم: آیا cur ادامه/سوپرستِ همان حرفِ prev است؟"""
|
||||
if not prev or not cur:
|
||||
return False
|
||||
short, long = (prev, cur) if len(prev) <= len(cur) else (cur, prev)
|
||||
common = 0
|
||||
for a, b in zip(short, long):
|
||||
if a == b:
|
||||
common += 1
|
||||
else:
|
||||
break
|
||||
return common >= max(8, int(len(short) * 0.7))
|
||||
|
||||
|
||||
def _fmt_srt(seconds: float) -> str:
|
||||
if seconds < 0:
|
||||
seconds = 0
|
||||
ms = int(round(seconds * 1000))
|
||||
h, ms = divmod(ms, 3600000)
|
||||
m, ms = divmod(ms, 60000)
|
||||
s, ms = divmod(ms, 1000)
|
||||
return f"{h:02d}:{m:02d}:{s:02d},{ms:03d}"
|
||||
|
||||
|
||||
def fmt_clock(seconds: float) -> str:
|
||||
if seconds < 0:
|
||||
seconds = 0
|
||||
total = int(seconds)
|
||||
h, rem = divmod(total, 3600)
|
||||
m, s = divmod(rem, 60)
|
||||
return f"{h:02d}:{m:02d}:{s:02d}" if h else f"{m:02d}:{s:02d}"
|
||||
|
||||
|
||||
def render_srt(segments: "list[dict]") -> str:
|
||||
if not segments:
|
||||
return ""
|
||||
base = min(seg["start"] for seg in segments)
|
||||
out = []
|
||||
for i, seg in enumerate(segments, 1):
|
||||
start = seg["start"] - base
|
||||
end = max(seg["end"], seg["start"]) - base
|
||||
if end <= start:
|
||||
end = start + 1.0 # حداقل ۱ ثانیه نمایش
|
||||
speaker = (seg.get("speaker") or "").strip()
|
||||
text = (seg.get("text") or "").strip()
|
||||
line = f"{speaker}: {text}" if speaker else text
|
||||
out.append(f"{i}\n{_fmt_srt(start)} --> {_fmt_srt(end)}\n{line}\n")
|
||||
return "\n".join(out)
|
||||
|
||||
|
||||
def render_txt(segments: "list[dict]") -> str:
|
||||
if not segments:
|
||||
return ""
|
||||
base = min(seg["start"] for seg in segments)
|
||||
lines = []
|
||||
for seg in segments:
|
||||
ts = fmt_clock(seg["start"] - base)
|
||||
speaker = (seg.get("speaker") or "").strip()
|
||||
text = (seg.get("text") or "").strip()
|
||||
lines.append(f"[{ts}] {speaker}: {text}" if speaker else f"[{ts}] {text}")
|
||||
return "\n".join(lines)
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# نوشتن / بهروزرسانی
|
||||
# ---------------------------------------------------------------------------
|
||||
def ensure_loaded(sid: str) -> "list[dict]":
|
||||
"""segmentهای جلسه را میدهد؛ اگر در حافظه نبود از <sid>.json میخواند (resume)."""
|
||||
if sid not in sessions_segments:
|
||||
p = TRANSCRIPTS_DIR / f"{sid}.json"
|
||||
if p.exists():
|
||||
try:
|
||||
sessions_segments[sid] = json.loads(p.read_text(encoding="utf-8"))
|
||||
except Exception:
|
||||
sessions_segments[sid] = []
|
||||
else:
|
||||
sessions_segments[sid] = []
|
||||
return sessions_segments[sid]
|
||||
|
||||
|
||||
def persist(sid: str) -> None:
|
||||
segs = sessions_segments.get(sid, [])
|
||||
(TRANSCRIPTS_DIR / f"{sid}.json").write_text(
|
||||
json.dumps(segs, ensure_ascii=False), encoding="utf-8")
|
||||
(TRANSCRIPTS_DIR / f"{sid}.srt").write_text(render_srt(segs), encoding="utf-8")
|
||||
(TRANSCRIPTS_DIR / f"{sid}.txt").write_text(render_txt(segs), encoding="utf-8")
|
||||
|
||||
|
||||
def upsert_segment(session_id, speaker, text, started_at, ended_at) -> None:
|
||||
"""یک caption را اضافه یا (اگر همان حرف باشد) بهروز میکند، سپس روی دیسک مینویسد."""
|
||||
sid = safe_session_id(session_id)
|
||||
text = (text or "").strip()
|
||||
if not text:
|
||||
return
|
||||
segs = ensure_loaded(sid)
|
||||
|
||||
start_s = to_epoch_s(started_at)
|
||||
end_s = to_epoch_s(ended_at)
|
||||
recv = time.time()
|
||||
|
||||
matched = False
|
||||
if segs:
|
||||
last = segs[-1]
|
||||
if last.get("speaker") == speaker:
|
||||
lk = last.get("start_key")
|
||||
if start_s is not None and lk is not None:
|
||||
matched = abs(lk - start_s) < 0.001
|
||||
elif start_s is None and lk is None:
|
||||
matched = is_continuation(last.get("text", ""), text)
|
||||
if matched:
|
||||
last = segs[-1]
|
||||
last["text"] = text # جدیدترین snapshot کاملترین است
|
||||
last["end"] = end_s if end_s is not None else recv
|
||||
else:
|
||||
seq = (segs[-1]["seq"] + 1) if segs else 1
|
||||
segs.append({
|
||||
"seq": seq,
|
||||
"speaker": speaker,
|
||||
"text": text,
|
||||
"start": start_s if start_s is not None else recv,
|
||||
"end": end_s if end_s is not None else (start_s if start_s is not None else recv),
|
||||
"start_key": start_s,
|
||||
})
|
||||
persist(sid)
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# خواندن (برای لایهی MCP)
|
||||
# ---------------------------------------------------------------------------
|
||||
def _read(sid: str, ext: str) -> "str | None":
|
||||
p = TRANSCRIPTS_DIR / f"{safe_session_id(sid)}.{ext}"
|
||||
if p.exists():
|
||||
try:
|
||||
return p.read_text(encoding="utf-8")
|
||||
except Exception:
|
||||
return None
|
||||
return None
|
||||
|
||||
|
||||
def read_txt(sid: str) -> "str | None":
|
||||
return _read(sid, "txt")
|
||||
|
||||
|
||||
def read_srt(sid: str) -> "str | None":
|
||||
return _read(sid, "srt")
|
||||
|
||||
|
||||
def read_segments(sid: str) -> "list[dict]":
|
||||
p = TRANSCRIPTS_DIR / f"{safe_session_id(sid)}.json"
|
||||
if p.exists():
|
||||
try:
|
||||
return json.loads(p.read_text(encoding="utf-8"))
|
||||
except Exception:
|
||||
return []
|
||||
return []
|
||||
|
||||
|
||||
def json_mtime(sid: str) -> float:
|
||||
p = TRANSCRIPTS_DIR / f"{safe_session_id(sid)}.json"
|
||||
return p.stat().st_mtime if p.exists() else 0.0
|
||||
|
||||
|
||||
def latest_session() -> "str | None":
|
||||
files = list(TRANSCRIPTS_DIR.glob("*.txt"))
|
||||
if not files:
|
||||
return None
|
||||
return max(files, key=lambda p: p.stat().st_mtime).stem
|
||||
|
||||
|
||||
def list_session_files():
|
||||
"""فایلهای .txt جلسهها را بهترتیبِ جدیدترین برمیگرداند."""
|
||||
return sorted(TRANSCRIPTS_DIR.glob("*.txt"),
|
||||
key=lambda p: p.stat().st_mtime, reverse=True)
|
||||
Reference in New Issue
Block a user