|
1
|
"""Microsoft Teams meeting recording source using the m365 CLI. |
|
2
|
|
|
3
|
Fetches Teams meeting recordings and transcripts via the Microsoft Graph API |
|
4
|
through the `m365` CLI tool. |
|
5
|
|
|
6
|
Requires: npm install -g @pnp/cli-microsoft365 |
|
7
|
Auth: m365 login (interactive) |
|
8
|
Docs: https://pnp.github.io/cli-microsoft365/ |
|
9
|
""" |
|
10
|
|
|
11
|
import logging |
|
12
|
import re |
|
13
|
import shutil |
|
14
|
import subprocess |
|
15
|
from pathlib import Path |
|
16
|
from typing import List, Optional |
|
17
|
|
|
18
|
from video_processor.sources.base import BaseSource, SourceFile |
|
19
|
from video_processor.sources.m365_source import _run_m365 |
|
20
|
|
|
21
|
logger = logging.getLogger(__name__) |
|
22
|
|
|
23
|
|
|
24
|
def _vtt_to_text(vtt: str) -> str: |
|
25
|
"""Strip VTT timing metadata and return plain text. |
|
26
|
|
|
27
|
Removes WEBVTT headers, timestamps (00:00:00.000 --> 00:00:05.000), |
|
28
|
cue identifiers, and deduplicates consecutive identical lines. |
|
29
|
""" |
|
30
|
lines = vtt.splitlines() |
|
31
|
text_lines: list[str] = [] |
|
32
|
prev_line = "" |
|
33
|
|
|
34
|
for line in lines: |
|
35
|
stripped = line.strip() |
|
36
|
# Skip WEBVTT header and NOTE blocks |
|
37
|
if stripped.startswith("WEBVTT") or stripped.startswith("NOTE"): |
|
38
|
continue |
|
39
|
# Skip timestamp lines (e.g. 00:00:01.000 --> 00:00:05.000) |
|
40
|
if re.match(r"\d{2}:\d{2}[:\.][\d.]+ --> \d{2}:\d{2}[:\.][\d.]+", stripped): |
|
41
|
continue |
|
42
|
# Skip numeric cue identifiers |
|
43
|
if re.match(r"^\d+$", stripped): |
|
44
|
continue |
|
45
|
# Skip blank lines |
|
46
|
if not stripped: |
|
47
|
continue |
|
48
|
# Strip inline VTT tags like <v Speaker> |
|
49
|
cleaned = re.sub(r"<[^>]+>", "", stripped).strip() |
|
50
|
if cleaned and cleaned != prev_line: |
|
51
|
text_lines.append(cleaned) |
|
52
|
prev_line = cleaned |
|
53
|
|
|
54
|
return "\n".join(text_lines) |
|
55
|
|
|
56
|
|
|
57
|
class TeamsRecordingSource(BaseSource): |
|
58
|
""" |
|
59
|
Fetch Teams meeting recordings and transcripts via the m365 CLI / Graph API. |
|
60
|
|
|
61
|
Usage: |
|
62
|
source = TeamsRecordingSource(user_id="me") |
|
63
|
source.authenticate() |
|
64
|
recordings = source.list_videos() |
|
65
|
source.download_all(recordings, Path("./recordings")) |
|
66
|
|
|
67
|
# Fetch transcript for a specific meeting |
|
68
|
transcript = source.fetch_transcript(meeting_id) |
|
69
|
""" |
|
70
|
|
|
71
|
def __init__(self, user_id: str = "me"): |
|
72
|
self.user_id = user_id |
|
73
|
|
|
74
|
def authenticate(self) -> bool: |
|
75
|
"""Check if m365 CLI is installed and logged in.""" |
|
76
|
if not shutil.which("m365"): |
|
77
|
logger.error("m365 CLI not found. Install with: npm install -g @pnp/cli-microsoft365") |
|
78
|
return False |
|
79
|
try: |
|
80
|
result = _run_m365(["status"], timeout=10) |
|
81
|
if isinstance(result, dict) and result.get("connectedAs"): |
|
82
|
return True |
|
83
|
if isinstance(result, str) and "Logged in" in result: |
|
84
|
return True |
|
85
|
logger.error("m365 not logged in. Run: m365 login") |
|
86
|
return False |
|
87
|
except (RuntimeError, subprocess.TimeoutExpired): |
|
88
|
logger.error("m365 not logged in. Run: m365 login") |
|
89
|
return False |
|
90
|
|
|
91
|
def list_videos( |
|
92
|
self, |
|
93
|
folder_id: Optional[str] = None, |
|
94
|
folder_path: Optional[str] = None, |
|
95
|
patterns: Optional[List[str]] = None, |
|
96
|
) -> List[SourceFile]: |
|
97
|
"""List Teams meeting recordings available for the user. |
|
98
|
|
|
99
|
Tries multiple approaches: |
|
100
|
1. Graph API onlineMeetings endpoint |
|
101
|
2. m365 teams meeting list command |
|
102
|
3. Fallback: search chat messages for recording links |
|
103
|
""" |
|
104
|
files: List[SourceFile] = [] |
|
105
|
|
|
106
|
# Approach 1: Graph API — list online meetings |
|
107
|
try: |
|
108
|
result = _run_m365( |
|
109
|
[ |
|
110
|
"request", |
|
111
|
"--url", |
|
112
|
f"https://graph.microsoft.com/v1.0/{self.user_id}/onlineMeetings", |
|
113
|
"--method", |
|
114
|
"get", |
|
115
|
], |
|
116
|
timeout=60, |
|
117
|
) |
|
118
|
meetings = self._extract_meetings_list(result) |
|
119
|
for meeting in meetings: |
|
120
|
recording_files = self._get_meeting_recordings(meeting) |
|
121
|
files.extend(recording_files) |
|
122
|
|
|
123
|
if files: |
|
124
|
logger.info(f"Found {len(files)} recording(s) via Graph API onlineMeetings") |
|
125
|
return files |
|
126
|
except RuntimeError as e: |
|
127
|
logger.debug(f"onlineMeetings endpoint failed: {e}") |
|
128
|
|
|
129
|
# Approach 2: m365 teams meeting list |
|
130
|
try: |
|
131
|
result = _run_m365(["teams", "meeting", "list"], timeout=60) |
|
132
|
meetings = result if isinstance(result, list) else [] |
|
133
|
for meeting in meetings: |
|
134
|
recording_files = self._get_meeting_recordings(meeting) |
|
135
|
files.extend(recording_files) |
|
136
|
|
|
137
|
if files: |
|
138
|
logger.info(f"Found {len(files)} recording(s) via m365 teams meeting list") |
|
139
|
return files |
|
140
|
except RuntimeError as e: |
|
141
|
logger.debug(f"teams meeting list failed: {e}") |
|
142
|
|
|
143
|
# Approach 3: Fallback — search chat messages for recording links |
|
144
|
try: |
|
145
|
result = _run_m365( |
|
146
|
[ |
|
147
|
"request", |
|
148
|
"--url", |
|
149
|
( |
|
150
|
f"https://graph.microsoft.com/v1.0/{self.user_id}/chats" |
|
151
|
"?$expand=messages($top=50)" |
|
152
|
"&$filter=chatType eq 'meeting'" |
|
153
|
"&$top=25" |
|
154
|
), |
|
155
|
"--method", |
|
156
|
"get", |
|
157
|
], |
|
158
|
timeout=60, |
|
159
|
) |
|
160
|
chats = self._extract_value_list(result) |
|
161
|
for chat in chats: |
|
162
|
messages = chat.get("messages", []) |
|
163
|
for msg in messages: |
|
164
|
body = msg.get("body", {}).get("content", "") |
|
165
|
if "recording" in body.lower() or ".mp4" in body.lower(): |
|
166
|
topic = chat.get("topic", "Meeting Recording") |
|
167
|
chat_id = chat.get("id", "") |
|
168
|
msg_id = msg.get("id", "") |
|
169
|
files.append( |
|
170
|
SourceFile( |
|
171
|
name=f"{topic}.mp4", |
|
172
|
id=f"{chat_id}:{msg_id}", |
|
173
|
mime_type="video/mp4", |
|
174
|
modified_at=msg.get("createdDateTime"), |
|
175
|
) |
|
176
|
) |
|
177
|
if files: |
|
178
|
logger.info(f"Found {len(files)} recording link(s) in meeting chats") |
|
179
|
except RuntimeError as e: |
|
180
|
logger.debug(f"Chat message fallback failed: {e}") |
|
181
|
|
|
182
|
if not files: |
|
183
|
logger.warning("No Teams meeting recordings found") |
|
184
|
|
|
185
|
return files |
|
186
|
|
|
187
|
def download(self, file: SourceFile, destination: Path) -> Path: |
|
188
|
"""Download a recording via its Graph API download URL.""" |
|
189
|
destination = Path(destination) |
|
190
|
destination.parent.mkdir(parents=True, exist_ok=True) |
|
191
|
|
|
192
|
download_url = file.path |
|
193
|
if download_url: |
|
194
|
# Use the direct download URL from contentUrl / @microsoft.graph.downloadUrl |
|
195
|
_run_m365( |
|
196
|
[ |
|
197
|
"request", |
|
198
|
"--url", |
|
199
|
download_url, |
|
200
|
"--method", |
|
201
|
"get", |
|
202
|
"--filePath", |
|
203
|
str(destination), |
|
204
|
], |
|
205
|
timeout=300, |
|
206
|
) |
|
207
|
else: |
|
208
|
# Try to get download URL from the recording ID |
|
209
|
meeting_id, _, recording_id = file.id.partition(":") |
|
210
|
if recording_id: |
|
211
|
url = ( |
|
212
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
213
|
f"/onlineMeetings/{meeting_id}" |
|
214
|
f"/recordings/{recording_id}/content" |
|
215
|
) |
|
216
|
else: |
|
217
|
url = ( |
|
218
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
219
|
f"/onlineMeetings/{meeting_id}/recordings" |
|
220
|
) |
|
221
|
# Fetch recording list to find the content URL |
|
222
|
result = _run_m365( |
|
223
|
["request", "--url", url, "--method", "get"], |
|
224
|
timeout=60, |
|
225
|
) |
|
226
|
recordings = self._extract_value_list(result) |
|
227
|
if recordings: |
|
228
|
url = ( |
|
229
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
230
|
f"/onlineMeetings/{meeting_id}" |
|
231
|
f"/recordings/{recordings[0].get('id', '')}/content" |
|
232
|
) |
|
233
|
|
|
234
|
_run_m365( |
|
235
|
[ |
|
236
|
"request", |
|
237
|
"--url", |
|
238
|
url, |
|
239
|
"--method", |
|
240
|
"get", |
|
241
|
"--filePath", |
|
242
|
str(destination), |
|
243
|
], |
|
244
|
timeout=300, |
|
245
|
) |
|
246
|
|
|
247
|
logger.info(f"Downloaded {file.name} to {destination}") |
|
248
|
return destination |
|
249
|
|
|
250
|
def fetch_transcript(self, meeting_id: str) -> Optional[str]: |
|
251
|
"""Fetch the transcript for a Teams meeting. |
|
252
|
|
|
253
|
Queries the Graph API transcripts endpoint, downloads the transcript |
|
254
|
content, and converts VTT format to plain text. |
|
255
|
""" |
|
256
|
try: |
|
257
|
result = _run_m365( |
|
258
|
[ |
|
259
|
"request", |
|
260
|
"--url", |
|
261
|
( |
|
262
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
263
|
f"/onlineMeetings/{meeting_id}/transcripts" |
|
264
|
), |
|
265
|
"--method", |
|
266
|
"get", |
|
267
|
], |
|
268
|
timeout=60, |
|
269
|
) |
|
270
|
except RuntimeError as e: |
|
271
|
logger.warning(f"Failed to list transcripts for meeting {meeting_id}: {e}") |
|
272
|
return None |
|
273
|
|
|
274
|
transcripts = self._extract_value_list(result) |
|
275
|
if not transcripts: |
|
276
|
logger.info(f"No transcripts found for meeting {meeting_id}") |
|
277
|
return None |
|
278
|
|
|
279
|
# Download the first available transcript |
|
280
|
transcript_id = transcripts[0].get("id", "") |
|
281
|
try: |
|
282
|
content_result = _run_m365( |
|
283
|
[ |
|
284
|
"request", |
|
285
|
"--url", |
|
286
|
( |
|
287
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
288
|
f"/onlineMeetings/{meeting_id}" |
|
289
|
f"/transcripts/{transcript_id}/content" |
|
290
|
), |
|
291
|
"--method", |
|
292
|
"get", |
|
293
|
"--accept", |
|
294
|
"text/vtt", |
|
295
|
], |
|
296
|
timeout=60, |
|
297
|
) |
|
298
|
except RuntimeError as e: |
|
299
|
logger.warning(f"Failed to download transcript {transcript_id}: {e}") |
|
300
|
return None |
|
301
|
|
|
302
|
# content_result may be raw VTT text or a dict with raw key |
|
303
|
if isinstance(content_result, dict): |
|
304
|
raw = content_result.get("raw", "") |
|
305
|
else: |
|
306
|
raw = str(content_result) |
|
307
|
|
|
308
|
if not raw: |
|
309
|
logger.warning(f"Empty transcript content for meeting {meeting_id}") |
|
310
|
return None |
|
311
|
|
|
312
|
return _vtt_to_text(raw) |
|
313
|
|
|
314
|
# ------------------------------------------------------------------ |
|
315
|
# Internal helpers |
|
316
|
# ------------------------------------------------------------------ |
|
317
|
|
|
318
|
def _extract_meetings_list(self, result) -> list: |
|
319
|
"""Extract meetings list from Graph API response.""" |
|
320
|
if isinstance(result, list): |
|
321
|
return result |
|
322
|
if isinstance(result, dict): |
|
323
|
return result.get("value", []) |
|
324
|
return [] |
|
325
|
|
|
326
|
def _extract_value_list(self, result) -> list: |
|
327
|
"""Extract value list from a Graph API response.""" |
|
328
|
if isinstance(result, list): |
|
329
|
return result |
|
330
|
if isinstance(result, dict): |
|
331
|
return result.get("value", []) |
|
332
|
return [] |
|
333
|
|
|
334
|
def _get_meeting_recordings(self, meeting: dict) -> List[SourceFile]: |
|
335
|
"""Fetch recordings for a single meeting and return SourceFile entries.""" |
|
336
|
meeting_id = meeting.get("id", "") |
|
337
|
subject = meeting.get("subject", meeting.get("topic", "Teams Meeting")) |
|
338
|
start_time = meeting.get("startDateTime", meeting.get("createdDateTime")) |
|
339
|
|
|
340
|
if not meeting_id: |
|
341
|
return [] |
|
342
|
|
|
343
|
try: |
|
344
|
result = _run_m365( |
|
345
|
[ |
|
346
|
"request", |
|
347
|
"--url", |
|
348
|
( |
|
349
|
f"https://graph.microsoft.com/v1.0/{self.user_id}" |
|
350
|
f"/onlineMeetings/{meeting_id}/recordings" |
|
351
|
), |
|
352
|
"--method", |
|
353
|
"get", |
|
354
|
], |
|
355
|
timeout=60, |
|
356
|
) |
|
357
|
except RuntimeError: |
|
358
|
return [] |
|
359
|
|
|
360
|
recordings = self._extract_value_list(result) |
|
361
|
files: List[SourceFile] = [] |
|
362
|
for rec in recordings: |
|
363
|
rec_id = rec.get("id", "") |
|
364
|
download_url = rec.get("content.downloadUrl", rec.get("contentUrl")) |
|
365
|
files.append( |
|
366
|
SourceFile( |
|
367
|
name=f"{subject}.mp4", |
|
368
|
id=f"{meeting_id}:{rec_id}", |
|
369
|
mime_type="video/mp4", |
|
370
|
modified_at=start_time, |
|
371
|
path=download_url, |
|
372
|
) |
|
373
|
) |
|
374
|
|
|
375
|
return files |
|
376
|
|