File size: 62,352 Bytes
5c64bb0
e9e9e3a
 
 
 
 
cd72c3f
 
 
ecd9eb2
e9e9e3a
6a18fe1
cd72c3f
f55e243
 
 
6a18fe1
 
cd72c3f
e9e9e3a
 
 
 
 
 
4629ea6
e9e9e3a
6a18fe1
b3e1b64
3ac7b5f
e9e9e3a
6a18fe1
7c987d0
b3e1b64
e9e9e3a
b3e1b64
f55e243
 
 
 
 
 
eeeef0c
 
 
5c64bb0
eeeef0c
 
 
 
 
 
5c64bb0
 
eeeef0c
e9e9e3a
9412ba0
7e4a1c6
6a18fe1
 
073dd22
6a18fe1
7e4a1c6
 
eeeef0c
5c64bb0
f55e243
b5cdfb8
b51c818
6a18fe1
cd72c3f
 
e9e9e3a
3ac7b5f
 
e9e9e3a
 
5c64bb0
 
e9e9e3a
 
 
5c64bb0
 
 
 
9412ba0
5c64bb0
3ac7b5f
c83a4a7
 
eeeef0c
c83a4a7
7c987d0
 
c83a4a7
 
b5cdfb8
 
c83a4a7
 
 
b5cdfb8
c83a4a7
 
68d4bf1
5c64bb0
 
 
 
3ac7b5f
 
5c64bb0
 
 
 
b51c818
eeeef0c
5c64bb0
 
 
d6e3c43
7c987d0
5c64bb0
b3e1b64
5c64bb0
b3e1b64
5c64bb0
 
 
 
 
f55e243
7c987d0
e9e9e3a
5c64bb0
 
e9e9e3a
5c64bb0
b3e1b64
5c64bb0
 
 
3ac7b5f
7c987d0
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b3e1b64
c83a4a7
 
 
5c64bb0
 
b3e1b64
 
c83a4a7
 
3ac7b5f
 
c83a4a7
 
 
3ac7b5f
f996499
5c64bb0
 
 
 
 
 
 
c83a4a7
 
 
 
5c64bb0
3ac7b5f
 
5c64bb0
2f82894
5c64bb0
 
 
 
c83a4a7
 
5c64bb0
c83a4a7
5c64bb0
b3e1b64
332d5fc
5c64bb0
 
c83a4a7
5c64bb0
 
332d5fc
3ac7b5f
c83a4a7
332d5fc
c83a4a7
 
5c64bb0
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
203a74a
5c64bb0
 
 
 
 
 
 
 
 
 
 
3ac7b5f
 
7c987d0
5c64bb0
 
c83a4a7
 
b3e1b64
5c64bb0
 
 
 
c83a4a7
 
5c64bb0
 
c83a4a7
 
 
 
 
5c64bb0
 
 
c83a4a7
5c64bb0
c83a4a7
 
5c64bb0
 
 
 
 
 
 
 
 
0f34792
5c64bb0
b3e1b64
 
c83a4a7
5c64bb0
 
 
 
c83a4a7
5c64bb0
 
b3e1b64
c83a4a7
5c64bb0
 
 
3ac7b5f
 
5c64bb0
 
 
c83a4a7
 
 
b3e1b64
c83a4a7
5c64bb0
c83a4a7
5c64bb0
c83a4a7
 
 
 
5c64bb0
 
 
 
 
 
 
 
 
 
0f34792
 
5c64bb0
b5cdfb8
5c64bb0
b5cdfb8
5c64bb0
 
 
 
c83a4a7
5c64bb0
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
d6e3c43
5c64bb0
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
5c64bb0
 
b5cdfb8
 
5c64bb0
b5cdfb8
5c64bb0
 
 
 
c83a4a7
5c64bb0
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
d6e3c43
5c64bb0
 
 
 
 
c83a4a7
 
 
 
5c64bb0
 
 
 
 
 
 
 
 
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
 
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
 
b5cdfb8
 
5c64bb0
 
 
 
b5cdfb8
5c64bb0
 
 
c83a4a7
 
5c64bb0
c83a4a7
5c64bb0
 
 
 
 
 
c83a4a7
5c64bb0
c83a4a7
5c64bb0
 
c83a4a7
5c64bb0
c83a4a7
5c64bb0
c83a4a7
 
5c64bb0
 
c83a4a7
 
5c64bb0
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
c83a4a7
 
5c64bb0
 
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
045ba8b
 
5c64bb0
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
d19297a
3ac7b5f
b3e1b64
5c64bb0
 
d19297a
 
 
b3e1b64
5c64bb0
 
 
d19297a
 
 
 
5c64bb0
d19297a
5c64bb0
 
 
 
 
 
d19297a
 
 
5c64bb0
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
c83a4a7
d19297a
5c64bb0
 
 
72be728
 
5c64bb0
 
203a74a
72be728
5c64bb0
 
 
203a74a
5c64bb0
 
 
 
 
 
6e780a2
5c64bb0
 
d6e3c43
0f34792
5c64bb0
 
 
 
6a18fe1
5c64bb0
 
 
 
 
 
 
 
6a18fe1
768889e
5c64bb0
 
 
6a18fe1
5c64bb0
 
 
 
a1a607a
5c64bb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
073dd22
b3e1b64
5c64bb0
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
68d4bf1
b3e1b64
5c64bb0
 
 
c83a4a7
5c64bb0
 
 
 
 
 
 
 
 
 
a4284b2
e9e9e3a
c83a4a7
5c64bb0
c83a4a7
5c64bb0
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
# main.py (Corrected SyntaxError at line 424 - Now with Gemini 2.0 as primary)
import os
import re
import logging
import asyncio
import json
import html
import contextlib
import traceback
from typing import Optional, Dict, Any, Tuple

# --- Frameworks ---
from starlette.applications import Starlette
from starlette.routing import Route
from starlette.responses import PlainTextResponse, JSONResponse, Response
from starlette.requests import Request

# --- Telegram Bot ---
from telegram import Update, InlineKeyboardButton, InlineKeyboardMarkup, Bot
from telegram.ext import (
    Application,
    CommandHandler,
    MessageHandler,
    filters,
    ContextTypes,
    CallbackQueryHandler,
)
from telegram.constants import ParseMode
from telegram.error import NetworkError, RetryAfter, TimedOut, BadRequest, TelegramError
from telegram.request import HTTPXRequest, BaseRequest

# --- Other Libraries ---
import httpx
from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, NoTranscriptFound
from bs4 import BeautifulSoup
from tenacity import retry, stop_after_attempt, wait_exponential, retry_if_exception_type, before_sleep_log
try:
    import lxml
    DEFAULT_PARSER = 'lxml'
except ImportError:
    DEFAULT_PARSER = 'html.parser'

# --- Google Gemini ---
try:
    import google.generativeai as genai
    from google.generativeai.types import HarmCategory, HarmBlockThreshold
    _gemini_available = True
except ImportError:
    genai = None
    HarmCategory = None
    HarmBlockThreshold = None
    _gemini_available = False
    # logger will be defined later, log warning after logger setup


# --- Logging Setup ---
logging.basicConfig( format='%(asctime)s - %(name)s - %(levelname)s - %(message)s', level=logging.INFO )
logging.getLogger("httpx").setLevel(logging.WARNING)
logging.getLogger("telegram.ext").setLevel(logging.INFO)
logging.getLogger('telegram.bot').setLevel(logging.INFO)
logging.getLogger("urllib3").setLevel(logging.INFO)
logging.getLogger('gunicorn.error').setLevel(logging.INFO)
logging.getLogger('uvicorn').setLevel(logging.INFO)
logging.getLogger('starlette').setLevel(logging.INFO)
if _gemini_available: logging.getLogger("google.ai.generativelanguage").setLevel(logging.WARNING)
logger = logging.getLogger(__name__)
logger.info(f"Logging configured. Using BS4 parser: {DEFAULT_PARSER}")
if not _gemini_available: logger.warning("google-generativeai library not found. Gemini functionality disabled.")


# --- Global variable for PTB app ---
ptb_app: Optional[Application] = None

# --- Environment Variable Loading & Configuration ---
logger.info("Attempting to load secrets and configuration...")
def get_secret(secret_name):
    value = os.environ.get(secret_name)
    if value: status = "Found"; log_length = min(len(value), 8); value_start = value[:log_length]; logger.info(f"Secret '{secret_name}': {status} (Value starts with: {value_start}...)")
    else: status = "Not Found"; logger.warning(f"Secret '{secret_name}': {status}")
    return value

TELEGRAM_TOKEN = get_secret('TELEGRAM_TOKEN')
OPENROUTER_API_KEY = get_secret('OPENROUTER_API_KEY') # Now Fallback
URLTOTEXT_API_KEY = get_secret('URLTOTEXT_API_KEY')
SUPADATA_API_KEY = get_secret('SUPADATA_API_KEY')
APIFY_API_TOKEN = get_secret('APIFY_API_TOKEN')
WEBHOOK_SECRET = get_secret('WEBHOOK_SECRET')
GEMINI_API_KEY = get_secret('GEMINI_API_KEY') # Now Primary

# Models (User can still configure via env vars)
OPENROUTER_MODEL = os.environ.get("OPENROUTER_MODEL", "deepseek/deepseek-chat-v3-0324:free") # Fallback Model
APIFY_ACTOR_ID = os.environ.get("APIFY_ACTOR_ID", "karamelo~youtube-transcripts")
GEMINI_MODEL = os.environ.get("GEMINI_MODEL", "gemini-2.0-flash-001") # Primary Model

if not TELEGRAM_TOKEN: logger.critical("❌ FATAL: TELEGRAM_TOKEN not found."); raise RuntimeError("Exiting: Telegram token missing.")
if not GEMINI_API_KEY: logger.error("❌ ERROR: GEMINI_API_KEY not found. Primary summarization (Gemini) will fail.")
if not OPENROUTER_API_KEY: logger.warning("⚠️ WARNING: OPENROUTER_API_KEY not found. Fallback summarization will fail.")

_gemini_primary_enabled = _gemini_available and bool(GEMINI_API_KEY)
if not _gemini_available: logger.warning("⚠️ WARNING: google-generativeai library missing. Gemini disabled.")
elif not GEMINI_API_KEY: logger.warning("⚠️ WARNING: GEMINI_API_KEY not found or empty. Gemini disabled.")

_openrouter_fallback_enabled = bool(OPENROUTER_API_KEY)
if not _openrouter_fallback_enabled: logger.warning("⚠️ WARNING: OPENROUTER_API_KEY not found. Fallback disabled.")


if not URLTOTEXT_API_KEY: pass
if not SUPADATA_API_KEY: pass
if not APIFY_API_TOKEN: pass
if not WEBHOOK_SECRET: logger.info("Optional secret 'WEBHOOK_SECRET' not found. Webhook security disabled.")

logger.info("Secret loading and configuration check finished.")
logger.info(f"Using Gemini Model (Primary): {GEMINI_MODEL if _gemini_primary_enabled else 'DISABLED'}")
logger.info(f"Using OpenRouter Model (Fallback): {OPENROUTER_MODEL if _openrouter_fallback_enabled else 'DISABLED'}")
logger.info(f"Using Apify Actor (via REST): {APIFY_ACTOR_ID}")
_apify_token_exists = bool(APIFY_API_TOKEN)


if _gemini_primary_enabled:
    try: genai.configure(api_key=GEMINI_API_KEY); logger.info("Google GenAI client configured successfully.")
    except Exception as e: logger.error(f"Failed to configure Google GenAI client: {e}"); _gemini_primary_enabled = False

# --- Retry Decorator ---
@retry( stop=stop_after_attempt(4), wait=wait_exponential(multiplier=1, min=2, max=15), retry=retry_if_exception_type((NetworkError, RetryAfter, TimedOut, BadRequest)), before_sleep=before_sleep_log(logger, logging.WARNING), reraise=True )
async def retry_bot_operation(func, *args, **kwargs):
    try: return await func(*args, **kwargs)
    except BadRequest as e:
        ignore_errors = [ "message is not modified", "query is too old", "message to edit not found", "chat not found", "bot was blocked by the user", ]
        if any(err in str(e).lower() for err in ignore_errors): logger.warning(f"Ignoring non-critical BadRequest: {e}"); return None
        logger.error(f"Potentially critical BadRequest: {e}"); raise
    except TelegramError as e: logger.warning(f"TelegramError (will retry if applicable): {e}"); raise
    except Exception as e: logger.error(f"Unexpected error during bot operation: {e}", exc_info=True); raise

# --- Helper Functions ---
def is_youtube_url(url):
    youtube_regex = re.compile( r'(?:https?://)?(?:www\.)?(?:m\.)?(?:youtube(?:-nocookie)?\.com|youtu\.be)/' r'(?:watch\?v=|embed/|v/|shorts/|live/|attribution_link\?a=.*&u=/watch\?v=)?' r'([\w-]{11})' r'(?:\S+)?', re.IGNORECASE)
    match = youtube_regex.search(url); logger.debug(f"is_youtube_url '{url}': {bool(match)}"); return bool(match)
def extract_youtube_id(url):
    youtube_regex = re.compile( r'(?:https?://)?(?:www\.)?(?:m\.)?(?:youtube(?:-nocookie)?\.com|youtu\.be)/' r'(?:watch\?v=|embed/|v/|shorts/|live/|attribution_link\?a=.*&u=/watch\?v=)?' r'([\w-]{11})' r'(?:\S+)?', re.IGNORECASE)
    match = youtube_regex.search(url)
    if match: video_id = match.group(1); logger.debug(f"Extracted YT ID '{video_id}' from {url}"); return video_id
    else: logger.warning(f"Could not extract YT ID from {url}"); return None


# --- Content Fetching Functions ---
# (fetch_url_content_for_scrape, get_transcript_via_supadata, get_transcript_via_apify, get_youtube_transcript, get_website_content, get_website_content_via_api remain the same as previous version)
async def fetch_url_content_for_scrape(url: str, timeout: int = 25) -> Optional[str]:
    headers = { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36', 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8', 'Accept-Language': 'en-US,en;q=0.5', 'Connection': 'keep-alive', 'DNT': '1', 'Upgrade-Insecure-Requests': '1' }
    try:
        async with httpx.AsyncClient(follow_redirects=True, timeout=timeout, headers=headers) as client:
            logger.debug(f"[Web Scrape] Sending request to {url}")
            response = await client.get(url)
            logger.debug(f"[Web Scrape] Received response {response.status_code} from {url}")
            response.raise_for_status()
            content_type = response.headers.get('content-type', '').lower()
            if 'html' not in content_type: logger.warning(f"[Web Scrape] Non-HTML content type from {url}: {content_type}"); return None
            try: return response.text
            except Exception as e: logger.error(f"[Web Scrape] Error decoding response for {url}: {e}"); return None
    except httpx.HTTPStatusError as e: logger.error(f"[Web Scrape] HTTP error {e.response.status_code} fetching {url}: {e}")
    except httpx.TimeoutException: logger.error(f"[Web Scrape] Timeout error fetching {url}")
    except httpx.TooManyRedirects: logger.error(f"[Web Scrape] Too many redirects fetching {url}")
    except httpx.RequestError as e: logger.error(f"[Web Scrape] Request error fetching {url}: {e}")
    except Exception as e: logger.error(f"[Web Scrape] Unexpected error fetching {url}: {e}", exc_info=True)
    return None

async def get_transcript_via_supadata(video_id: str, api_key: str) -> Optional[str]:
    if not video_id: logger.error("[Supadata] No video_id provided"); return None
    if not api_key: logger.error("[Supadata] API key missing."); return None
    logger.info(f"[Supadata] Attempting fetch for video ID: {video_id}")
    api_endpoint = "https://api.supadata.ai/v1/youtube/transcript" # Corrected URL
    params = {"videoId": video_id, "format": "text"}; headers = {"X-API-Key": api_key}
    try:
        async with httpx.AsyncClient(timeout=30.0) as client:
            response = await client.get(api_endpoint, headers=headers, params=params)
            logger.debug(f"[Supadata] Status code {response.status_code} for {video_id}")
            if response.status_code == 200:
                try:
                    try: data = response.json()
                    except json.JSONDecodeError: data = None
                    content = None
                    if data: content = data if isinstance(data, str) else data.get("transcript") or data.get("text") or data.get("data")
                    if not content and response.text: content = response.text
                    if content and isinstance(content, str): logger.info(f"[Supadata] Success for {video_id}. Length: {len(content)}"); return content.strip()
                    else: logger.warning(f"[Supadata] Success but content empty/invalid for {video_id}. Response: {response.text[:200]}"); return None
                except Exception as e: logger.error(f"[Supadata] Error processing success response for {video_id}: {e}", exc_info=True); return None
            elif response.status_code in [401, 403]: logger.error(f"[Supadata] Auth error ({response.status_code}). Check API key."); return None
            elif response.status_code == 404: logger.warning(f"[Supadata] Not found (404) for {video_id}."); return None
            else: logger.error(f"[Supadata] Unexpected status {response.status_code} for {video_id}. Resp: {response.text[:200]}"); return None
    except httpx.TimeoutException: logger.error(f"[Supadata] Timeout connecting for {video_id}"); return None
    except httpx.RequestError as e:
        if "CERTIFICATE_VERIFY_FAILED" in str(e): logger.error(f"[Supadata] SSL Cert Verify Failed for {video_id}: {e}")
        else: logger.error(f"[Supadata] Request error for {video_id}: {e}")
        return None
    except Exception as e: logger.error(f"[Supadata] Unexpected error for {video_id}: {e}", exc_info=True); return None

async def get_transcript_via_apify(video_url: str, api_token: str) -> Optional[str]:
    """Fetches YouTube transcript using Apify REST API (run-sync-get-dataset-items endpoint)."""
    global APIFY_ACTOR_ID
    if not video_url: logger.error("[Apify SyncItems] No video_url provided"); return None
    if not api_token: logger.error("[Apify SyncItems] API token missing."); return None
    logger.info(f"[Apify SyncItems] Attempting fetch for URL: {video_url} (Actor: {APIFY_ACTOR_ID})")

    sync_items_endpoint = f"https://api.apify.com/v2/acts/{APIFY_ACTOR_ID}/run-sync-get-dataset-items"
    params = {"token": api_token}
    payload = { "urls": [video_url], "outputFormat": "singleStringText", "maxRetries": 5, "channelHandleBoolean": False, "channelNameBoolean": False, "datePublishedBoolean": False, "relativeDateTextBoolean": False, }
    headers = {"Content-Type": "application/json"}

    try:
        async with httpx.AsyncClient(timeout=120.0) as client:
            log_headers = {k: v for k, v in headers.items()}
            logger.debug(f"[Apify SyncItems] POST Request Details:\nURL: {sync_items_endpoint}\nParams: {params}\nHeaders: {log_headers}\nPayload: {json.dumps(payload)}")
            response = await client.post(sync_items_endpoint, headers=headers, params=params, json=payload)
            logger.debug(f"[Apify SyncItems] Received status code {response.status_code} for {video_url}")

            if response.status_code == 200:
                try:
                    results = response.json()
                    if isinstance(results, list) and len(results) > 0:
                        item = results[0]
                        content = None
                        if "captions" in item and isinstance(item["captions"], str): logger.info("[Apify SyncItems] Found 'captions' key with string content."); content = item["captions"]
                        elif "text" in item and isinstance(item["text"], str): logger.info("[Apify SyncItems] Found 'text' key with string content."); content = item["text"]
                        elif "transcript" in item and isinstance(item["transcript"], str): logger.info("[Apify SyncItems] Found 'transcript' key with string content."); content = item["transcript"]
                        elif "captions" in item and isinstance(item["captions"], list):
                             logger.warning("[Apify SyncItems] Received list format for 'captions' unexpectedly. Processing...")
                             if len(item["captions"]) > 0 and isinstance(item["captions"][0], dict) and 'text' in item["captions"][0]: content = " ".join(line.get("text", "") for line in item["captions"] if line.get("text"))
                             elif len(item["captions"]) > 0 and isinstance(item["captions"][0], str): content = " ".join(item["captions"])
                        if content and isinstance(content, str): logger.info(f"[Apify SyncItems] Success via REST for {video_url}. Length: {len(content)}"); return content.strip()
                        else: logger.warning(f"[Apify SyncItems] Dataset item parsed but transcript content empty/invalid format for {video_url}. Item keys: {list(item.keys())}"); return None
                    else: logger.warning(f"[Apify SyncItems] Actor success but dataset was empty for {video_url}. Response: {results}"); return None
                except json.JSONDecodeError: logger.error(f"[Apify SyncItems] Failed JSON decode. Status:{response.status_code}. Resp:{response.text[:200]}"); return None
                except Exception as e: logger.error(f"[Apify SyncItems] Error processing success response for {video_url}: {e}", exc_info=True); return None
            elif response.status_code == 400: logger.error(f"[Apify SyncItems] Bad Request (400) for {video_url}. Check payload. Resp:{response.text[:200]}"); return None
            elif response.status_code == 401: logger.error("[Apify SyncItems] Auth error (401). Check token."); return None
            elif response.status_code == 404:
                 error_info = "";
                 try: error_info = response.json().get("error", {}).get("message", "")
                 except Exception: pass
                 logger.error(f"[Apify SyncItems] Endpoint/Actor Not Found (404). Error: '{error_info}' Resp:{response.text[:200]}");
                 return None
            else: logger.error(f"[Apify SyncItems] Unexpected status {response.status_code} for {video_url}. Resp:{response.text[:200]}"); return None

    except httpx.TimeoutException as e: logger.error(f"[Apify SyncItems] Timeout during API interaction for {video_url}: {e}"); return None
    except httpx.HTTPStatusError as e: logger.error(f"[Apify SyncItems] HTTP Status Error during API interaction for {video_url}: {e}"); return None
    except httpx.RequestError as e: logger.error(f"[Apify SyncItems] Request error during API interaction for {video_url}: {e}"); return None
    except Exception as e: logger.error(f"[Apify SyncItems] Unexpected error during Apify SyncItems REST call for {video_url}: {e}", exc_info=True); return None

async def get_youtube_transcript(video_id: str, video_url: str) -> Optional[str]:
    global SUPADATA_API_KEY, APIFY_API_TOKEN
    if not video_id: logger.error("get_youtube_transcript: No video_id"); return None
    logger.info(f"Fetching transcript for video ID: {video_id} (URL: {video_url})")
    transcript_text = None
    logger.info("[Primary YT] Attempting youtube-transcript-api...")
    try:
        transcript_list = await asyncio.to_thread( YouTubeTranscriptApi.get_transcript, video_id, languages=['en', 'en-GB', 'en-US'] )
        if transcript_list: transcript_text = " ".join([item['text'] for item in transcript_list if 'text' in item])
        if transcript_text: logger.info(f"[Primary YT] Success via lib for {video_id} (len: {len(transcript_text)})"); return transcript_text
        else: logger.warning(f"[Primary YT] Transcript list/text empty for {video_id}"); transcript_text = None
    except Exception as e:
        logger.warning(f"[Primary YT] Error via lib for {video_id}: {e}")
        if "No transcript found" in str(e): logger.warning(f"[Primary YT] No transcript found.")
        elif "disabled" in str(e): logger.warning(f"[Primary YT] Transcripts disabled.")
        transcript_text = None
    if transcript_text is None:
        logger.info("[Fallback YT 1] Trying Supadata API...")
        if SUPADATA_API_KEY:
            transcript_text = await get_transcript_via_supadata(video_id, SUPADATA_API_KEY)
            if transcript_text: logger.info(f"[Fallback YT 1] Success via Supadata for {video_id}"); return transcript_text
            else: logger.warning(f"[Fallback YT 1] Supadata failed or no content for {video_id}.")
        else: logger.warning("[Fallback YT 1] Supadata API key unavailable. Skipping.")
    if transcript_text is None:
        logger.info("[Fallback YT 2] Trying Apify REST API (SyncItems)...")
        if APIFY_API_TOKEN:
            transcript_text = await get_transcript_via_apify(video_url, APIFY_API_TOKEN)
            if transcript_text: logger.info(f"[Fallback YT 2] Success via Apify SyncItems REST for {video_url}"); return transcript_text
            else: logger.warning(f"[Fallback YT 2] Apify SyncItems REST failed or no content for {video_url}.")
        else: logger.warning("[Fallback YT 2] Apify API token unavailable. Skipping.")
    if transcript_text is None: logger.error(f"All methods failed for YT transcript: {video_id}"); return None
    return transcript_text

async def get_website_content(url: str) -> Optional[str]:
    if not url: logger.error("get_website_content: No URL"); return None
    logger.info(f"[Primary Web] Fetching website content for: {url}")
    html_content = await fetch_url_content_for_scrape(url)
    if not html_content: return None
    try:
        def parse_html(content):
            soup = BeautifulSoup(content, DEFAULT_PARSER)
            for element in soup(["script", "style", "header", "footer", "nav", "aside", "form", "button", "input", "iframe", "img", "svg", "link", "meta", "noscript", "figure"]): element.extract()
            main_content = soup.find('main') or soup.find('article') or soup.find(id='content') or soup.find(class_='content') or soup.find(id='main-content') or soup.find(class_='main-content') or soup.find(role='main')
            target_element = main_content if main_content else soup.body
            if not target_element: logger.warning(f"[Primary Web] Could not find body/main for parsing {url}"); return None
            lines = [line.strip() for line in target_element.get_text(separator='\n', strip=True).splitlines() if line.strip()]
            text = " ".join(lines)
            if not text: logger.warning(f"[Primary Web] Extracted text empty after clean for {url}"); return None
            return text
        text_content = await asyncio.to_thread(parse_html, html_content)
        if text_content: logger.info(f"[Primary Web] Success scrape for {url} (final len: {len(text_content)})"); return text_content
        else: return None
    except Exception as e: logger.error(f"[Primary Web] Error scraping/parsing {url}: {e}", exc_info=True); return None

async def get_website_content_via_api(url: str, api_key: str) -> Optional[str]:
    if not url: logger.error("[Fallback Web API] No URL"); return None
    if not api_key: logger.error("[Fallback Web API] urltotext.com API key missing."); return None
    logger.info(f"[Fallback Web API] Attempting fetch for: {url} using urltotext.com API")
    api_endpoint = "https://urltotext.com/api/v1/urltotext/"
    payload = { "url": url, "output_format": "text", "extract_main_content": True, "render_javascript": True, "residential_proxy": False }
    headers = { "Authorization": f"Token {api_key}", "Content-Type": "application/json" }
    try:
        async with httpx.AsyncClient(timeout=45.0) as client:
            logger.debug(f"[Fallback Web API] Sending request to urltotext.com API for {url}")
            response = await client.post(api_endpoint, headers=headers, json=payload)
            logger.debug(f"[Fallback Web API] Received status {response.status_code} from urltotext.com API for {url}")
            if response.status_code == 200:
                try:
                    data = response.json()
                    content = data.get("data", {}).get("content"); credits = data.get("credits_used", "N/A"); warning = data.get("data", {}).get("warning")
                    if warning: logger.warning(f"[Fallback Web API] urltotext.com API Warning for {url}: {warning}")
                    if content: logger.info(f"[Fallback Web API] Success via urltotext.com API for {url}. Len: {len(content)}. Credits: {credits}"); return content.strip()
                    else: logger.warning(f"[Fallback Web API] urltotext.com API success but content empty for {url}. Resp: {data}"); return None
                except json.JSONDecodeError: logger.error(f"[Fallback Web API] Failed JSON decode urltotext.com for {url}. Resp:{response.text[:500]}"); return None
                except Exception as e: logger.error(f"[Fallback Web API] Error processing urltotext.com success response for {url}: {e}", exc_info=True); return None
            elif response.status_code in [400, 401, 402, 403, 422, 500]: logger.error(f"[Fallback Web API] Error {response.status_code} from urltotext.com API for {url}. Resp:{response.text[:200]}"); return None
            else: logger.error(f"[Fallback Web API] Unexpected status {response.status_code} from urltotext.com API for {url}. Resp:{response.text[:200]}"); return None
    except httpx.TimeoutException: logger.error(f"[Fallback Web API] Timeout connecting to urltotext.com API for {url}"); return None
    except httpx.RequestError as e: logger.error(f"[Fallback Web API] Request error connecting to urltotext.com API for {url}: {e}"); return None
    except Exception as e: logger.error(f"[Fallback Web API] Unexpected error during urltotext.com API call for {url}: {e}", exc_info=True); return None

# --- Summarization Functions ---

async def _call_gemini(text: str, summary_type: str) -> Tuple[Optional[str], Optional[str]]:
    """Internal function to call Gemini API. Returns (summary, error_message)."""
    global GEMINI_MODEL, _gemini_primary_enabled
    if not _gemini_primary_enabled:
        logger.error("[Gemini Primary] Called but is disabled.");
        return None, "Error: Primary AI service (Gemini) not configured/available."
    logger.info(f"[Gemini Primary] Generating {summary_type} summary using {GEMINI_MODEL}. Input length: {len(text)}")

    # Define prompts
    if summary_type == "paragraph":
        prompt = ("You are an AI model designed to provide concise summaries using British English spellings. Your output MUST be:\n"
                  "β€’ Clear and simple language suitable for someone unfamiliar with the topic.\n"
                  "β€’ Uses British English spellings throughout.\n"
                  "β€’ Straightforward and understandable vocabulary; avoid complex terms.\n"
                  "β€’ Presented as ONE SINGLE PARAGRAPH.\n"
                  "β€’ No more than 85 words maximum; but does not have to be exactly 85.\n"
                  "β€’ Considers the entire text content equally.\n"
                  "β€’ Uses semicolons (;) instead of em dashes (– or β€”).\n"
                  "β€’ **Focus ONLY on the main content; strictly EXCLUDE information about website features, subscriptions, ads, cookie notices, or navigation elements. Do not include things like free/paid tiers; basic/premium memberships. Especially for ACS membership.**\n\n" # Added instruction
                  "Here is the text to summarise:")
    else: # points summary
        prompt = ("You are an AI model designed to provide concise summaries using British English spellings. Your output MUST strictly follow this format:\n\n"
                  "β€’ For each distinct topic or section identified in the text, create a heading.\n"
                  "β€’ Each heading MUST be plain text without any formatting (e.g., Section Title).\n"
                  "β€’ Immediately following each heading, list the key points as a bulleted list.\n"
                  "β€’ Each bullet point MUST start with a hyphen and a space (- ) on a new line.\n"
                  "β€’ The text within each bullet point should NOT contain any bold formatting.\n"
                  "β€’ IMPORTANT: Never use bold formatting (double asterisks) within the text of the bullet points themselves.\n"
                  "β€’ Use clear, simple, and straightforward language suitable for someone unfamiliar with the topic.\n"
                  "β€’ Use British English spellings throughout.\n"
                  "β€’ Avoid overly complex or advanced vocabulary.\n"
                  "β€’ Keep bullet points concise.\n"
                  "β€’ Ensure the entire summary takes no more than two minutes to read.\n"
                  "β€’ Consider the entire text's content, not just the beginning or a few topics.\n"
                  "β€’ Use semicolons (;) instead of em dashes (– or β€”).\n"
                  "β€’ **Focus ONLY on the main content; strictly EXCLUDE information about website features, subscriptions, ads, cookie notices, or navigation elements. Do not include things like free/paid tiers; basic/premium memberships. Especially for ACS membership.**\n\n" # Added instruction
                  "Here is the text to summarise:")

    # Input Length Check (Gemini-specific limits if known, otherwise use a large default)
    MAX_INPUT_LENGTH_GEMINI = 1000000 # Default limit for Gemini
    if len(text) > MAX_INPUT_LENGTH_GEMINI:
        logger.warning(f"[Gemini Primary] Input length ({len(text)}) exceeds limit ({MAX_INPUT_LENGTH_GEMINI}). Truncating.");
        text = text[:MAX_INPUT_LENGTH_GEMINI] + "... (Content truncated)"
    full_prompt = f"{prompt}\n\n{text}"

    # Safety Settings (Block None)
    safety_settings = {
        HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
        getattr(HarmCategory, 'HARM_CATEGORY_CIVIC_INTEGRITY', None): HarmBlockThreshold.BLOCK_NONE
    }
    safety_settings = {k: v for k, v in safety_settings.items() if k is not None}
    logger.debug(f"[Gemini Primary] Using safety settings: {safety_settings}")

    try:
        logger.debug(f"[Gemini Primary] Initializing model {GEMINI_MODEL}")
        model = genai.GenerativeModel(GEMINI_MODEL)
        logger.info(f"[Gemini Primary] Sending request to Gemini ({GEMINI_MODEL})...")
        request_options = {"timeout": 120} # Generous timeout for Gemini
        response = await model.generate_content_async( full_prompt, safety_settings=safety_settings, request_options=request_options )
        logger.info("[Gemini Primary] Received response from Gemini.")

        # Check for immediate blocking reasons
        if response.prompt_feedback.block_reason:
            logger.warning(f"[Gemini Primary] Request blocked by API. Reason: {response.prompt_feedback.block_reason}");
            return None, f"Sorry, the primary AI model ({GEMINI_MODEL}) blocked the request (Reason: {response.prompt_feedback.block_reason})."

        # Check candidate-level blocking
        for cand in response.candidates:
            if cand.finish_reason == 'SAFETY':
                logger.warning(f"[Gemini Primary] Candidate blocked due to SAFETY. Ratings: {cand.safety_ratings}")
                return None, f"Sorry, the primary AI model ({GEMINI_MODEL}) blocked the response due to safety filters."

        # Try to get text, handle potential ValueError if blocked
        try:
            summary = response.text
        except ValueError as e:
            logger.warning(f"[Gemini Primary] Error accessing response text (likely blocked content): {e}");
            summary = None

        if summary:
            logger.info(f"[Gemini Primary] Success generating summary. Output len: {len(summary)}");
            # Escape Markdown for Telegram only if necessary (Removed escaping as it might conflict with plain heading)
            # escaped_summary = summary.strip().replace('_', r'\_').replace('*', r'\*').replace('[', r'\[').replace('`', r'\`')
            # Let's see if the AI respects the instruction without further escaping. If not, we might need selective escaping.
            return summary.strip(), None
        else:
            finish_reason = response.candidates[0].finish_reason if response.candidates else 'N/A'
            logger.warning(f"[Gemini Primary] Gemini returned empty summary or content was blocked. Finish reason: {finish_reason}");
            return None, f"Sorry, the primary AI model ({GEMINI_MODEL}) did not provide a summary (Finish Reason: {finish_reason})."

    except Exception as e:
        logger.error(f"[Gemini Primary] Unexpected error during Gemini API call: {e}", exc_info=True);
        return None, f"Sorry, an unexpected error occurred while using the primary AI service ({GEMINI_MODEL})."

async def _call_openrouter(text: str, summary_type: str) -> Tuple[Optional[str], Optional[str]]:
    """Internal function to call OpenRouter API (Fallback). Returns (summary, error_message)."""
    global OPENROUTER_API_KEY, OPENROUTER_MODEL, _openrouter_fallback_enabled
    if not _openrouter_fallback_enabled:
        logger.error("[OpenRouter Fallback] Called but is disabled.");
        return None, "Error: Fallback AI service (OpenRouter) not configured/available."
    logger.info(f"[OpenRouter Fallback] Generating {summary_type} summary using {OPENROUTER_MODEL}. Input length: {len(text)}")

    # Define prompts
    if summary_type == "paragraph":
        prompt = ("You are an AI model designed to provide concise summaries using British English spellings. Your output MUST be:\n"
                  "β€’ Clear and simple language suitable for someone unfamiliar with the topic.\n"
                  "β€’ Uses British English spellings throughout.\n"
                  "β€’ Straightforward and understandable vocabulary; avoid complex terms.\n"
                  "β€’ Presented as ONE SINGLE PARAGRAPH.\n"
                  "β€’ No more than 85 words maximum; but does not have to be exactly 85.\n"
                  "β€’ Considers the entire text content equally.\n"
                  "β€’ Uses semicolons (;) instead of em dashes (– or β€”).\n"
                  "β€’ **Focus ONLY on the main content; strictly EXCLUDE information about website features, subscriptions, ads, cookie notices, or navigation elements. Do not include things like free/paid tiers; basic/premium memberships. Especially for ACS membership.**\n\n" # Added instruction
                  "Here is the text to summarise:")
    else: # points summary
        prompt = ("You are an AI model designed to provide concise summaries using British English spellings. Your output MUST strictly follow this format:\n\n"
                  "β€’ For each distinct topic or section identified in the text, create a heading.\n"
                  "β€’ Each heading MUST be plain text without any formatting (e.g., Section Title).\n"
                  "β€’ Immediately following each heading, list the key points as a bulleted list.\n"
                  "β€’ Each bullet point MUST start with a hyphen and a space (- ) on a new line.\n"
                  "β€’ The text within each bullet point should NOT contain any bold formatting.\n"
                  "β€’ IMPORTANT: Never use bold formatting (double asterisks) within the text of the bullet points themselves.\n"
                  "β€’ Use clear, simple, and straightforward language suitable for someone unfamiliar with the topic.\n"
                  "β€’ Use British English spellings throughout.\n"
                  "β€’ Avoid overly complex or advanced vocabulary.\n"
                  "β€’ Keep bullet points concise.\n"
                  "β€’ Ensure the entire summary takes no more than two minutes to read.\n"
                  "β€’ Consider the entire text's content, not just the beginning or a few topics.\n"
                  "β€’ Use semicolons (;) instead of em dashes (– or β€”).\n"
                  "β€’ **Focus ONLY on the main content; strictly EXCLUDE information about website features, subscriptions, ads, cookie notices, or navigation elements. Do not include things like free/paid tiers; basic/premium memberships. Especially for ACS membership.**\n\n" # Added instruction
                  "Here is the text to summarise:")

    # Input Length Check (OpenRouter-specific limit)
    MAX_INPUT_LENGTH_OR = 500000
    if len(text) > MAX_INPUT_LENGTH_OR:
        logger.warning(f"[OpenRouter Fallback] Input length ({len(text)}) exceeds limit ({MAX_INPUT_LENGTH_OR}). Truncating.");
        text = text[:MAX_INPUT_LENGTH_OR] + "... (Content truncated)"
    full_prompt = f"{prompt}\n\n{text}"

    headers = { "Authorization": f"Bearer {OPENROUTER_API_KEY}", "Content-Type": "application/json" }
    payload = { "model": OPENROUTER_MODEL, "messages": [{"role": "user", "content": full_prompt}] }
    openrouter_api_endpoint = "https://openrouter.ai/api/v1/chat/completions"
    # Use reasonable timeouts for fallback
    api_timeouts = httpx.Timeout(25.0, read=20.0, write=25.0, pool=60.0)
    response = None

    try:
        async with httpx.AsyncClient(timeout=api_timeouts) as client:
            logger.info(f"[OpenRouter Fallback] Sending request to OpenRouter ({OPENROUTER_MODEL}) with read timeout {api_timeouts.read}s...")
            response = await client.post(openrouter_api_endpoint, headers=headers, json=payload)
            if response: logger.info(f"[OpenRouter Fallback] Received response from OpenRouter. Status code: {response.status_code}")
            else: logger.error("[OpenRouter Fallback] No response from OpenRouter (unexpected)."); return None, "Sorry, fallback AI service failed unexpectedly."

            if response.status_code == 200:
                try:
                    data = response.json()
                    if data.get("choices") and isinstance(data["choices"], list) and len(data["choices"]) > 0:
                        message = data["choices"][0].get("message")
                        if message and isinstance(message, dict):
                            summary = message.get("content")
                            if summary:
                                logger.info(f"[OpenRouter Fallback] Success via OpenRouter. Output len: {len(summary)}")
                                # Escape Markdown for Telegram only if necessary (Removed escaping)
                                # escaped_summary = summary.strip().replace('_', r'\_').replace('*', r'\*').replace('[', r'\[').replace('`', r'\`')
                                return summary.strip(), None
                            else:
                                logger.warning(f"[OpenRouter Fallback] OpenRouter success but content empty. Resp: {data}")
                                return None, "Sorry, the fallback AI model returned an empty summary."
                        else:
                            logger.error(f"[OpenRouter Fallback] Unexpected message structure: {message}. Full: {data}")
                            return None, "Sorry, could not parse fallback AI response (format)."
                    else:
                        logger.error(f"[OpenRouter Fallback] Unexpected choices structure: {data.get('choices')}. Full: {data}")
                        return None, "Sorry, could not parse fallback AI response (choices)."
                except json.JSONDecodeError:
                    logger.error(f"[OpenRouter Fallback] Failed JSON decode OpenRouter. Status:{response.status_code}. Resp:{response.text[:500]}")
                    return None, "Sorry, failed to understand fallback AI response."
                except Exception as e:
                    logger.error(f"[OpenRouter Fallback] Error processing OpenRouter success response: {e}", exc_info=True)
                    return None, "Sorry, error processing fallback AI response."
            # Handle specific error codes
            elif response.status_code == 401: logger.error("[OpenRouter Fallback] API key invalid (401)."); return None, "Error: Fallback AI model configuration key is invalid."
            elif response.status_code == 402: logger.error("[OpenRouter Fallback] Payment Required (402)."); return None, "Sorry, fallback AI service limits/payment issue."
            elif response.status_code == 429: logger.warning("[OpenRouter Fallback] Rate Limit Exceeded (429)."); return None, "Sorry, fallback AI model is busy. Try again."
            elif response.status_code == 500: logger.error(f"[OpenRouter Fallback] Internal Server Error (500). Resp:{response.text[:500]}"); return None, "Sorry, fallback AI service internal error."
            else:
                error_info = "";
                try: error_info = response.json().get("error", {}).get("message", "")
                except Exception: pass
                logger.error(f"[OpenRouter Fallback] Unexpected status {response.status_code}. Error: '{error_info}' Resp:{response.text[:500]}");
                return None, f"Sorry, fallback AI service returned unexpected status ({response.status_code})."

    except httpx.TimeoutException as e:
        logger.error(f"[OpenRouter Fallback] Timeout error ({type(e)}) connecting/reading from OpenRouter API: {e}")
        return None, f"Sorry, the fallback AI service ({OPENROUTER_MODEL}) timed out."
    except httpx.RequestError as e:
        logger.error(f"[OpenRouter Fallback] Request error connecting to OpenRouter API: {e}")
        return None, "Sorry, there was an error connecting to the fallback AI model service."
    except Exception as e:
        logger.error(f"[OpenRouter Fallback] Unexpected error during OpenRouter call: {e}", exc_info=True)
        return None, "Sorry, an unexpected error occurred while using the fallback AI service."

async def generate_summary(text: str, summary_type: str) -> str:
    """
    Generates summary using Gemini (Primary) and falls back to OpenRouter if needed.
    Ensures point-based summaries have plain text headings.
    """
    global _gemini_primary_enabled, _openrouter_fallback_enabled, GEMINI_MODEL, OPENROUTER_MODEL
    logger.info(f"[Summary Generation] Starting process. Primary: Gemini ({GEMINI_MODEL}), Fallback: OpenRouter ({OPENROUTER_MODEL})")

    final_summary: Optional[str] = None
    error_message: Optional[str] = None

    # --- Attempt 1: Gemini (Primary) ---
    if _gemini_primary_enabled:
        logger.info(f"[Summary Generation] Attempting primary AI: Gemini ({GEMINI_MODEL})")
        final_summary, error_message = await _call_gemini(text, summary_type)
        if final_summary:
            logger.info(f"[Summary Generation] Success with primary AI (Gemini).")
            # Return directly - prompt handles formatting
            return final_summary
        else:
            logger.warning(f"[Summary Generation] Primary AI (Gemini) failed or returned unusable result. Error: {error_message}. Proceeding to fallback.")
    else:
        logger.warning("[Summary Generation] Primary AI (Gemini) is disabled or unavailable. Proceeding directly to fallback.")
        error_message = "Primary AI (Gemini) unavailable." # Set initial error message

    # --- Attempt 2: OpenRouter (Fallback) ---
    if _openrouter_fallback_enabled:
        logger.info(f"[Summary Generation] Attempting fallback AI: OpenRouter ({OPENROUTER_MODEL})")
        fallback_summary, fallback_error = await _call_openrouter(text, summary_type)
        if fallback_summary:
            logger.info(f"[Summary Generation] Success with fallback AI (OpenRouter).")
            # Return directly - prompt handles formatting
            return fallback_summary
        else:
            logger.error(f"[Summary Generation] Fallback AI (OpenRouter) also failed. Error: {fallback_error}")
            # Combine error messages if possible
            if error_message: # Keep the primary error if it exists
                 return f"{error_message} Fallback AI ({OPENROUTER_MODEL}) also failed: {fallback_error}"
            else: # If primary was skipped
                 return f"Fallback AI ({OPENROUTER_MODEL}) failed: {fallback_error}"
    else:
        logger.error("[Summary Generation] Fallback AI (OpenRouter) is disabled or unavailable. Cannot proceed.")
        if error_message: # Return the primary error
            return f"{error_message} Fallback AI is also unavailable."
        else: # Should not happen if logic is correct, but safeguard
            return "Error: Both primary and fallback AI services are unavailable."

    # Should ideally not be reached if logic above is sound, but as a final catch
    logger.error("[Summary Generation] Reached end of function without returning a summary or specific error.")
    return "Sorry, an unknown error occurred during summary generation after trying all available models."


# (process_summary_task, handlers, setup, lifespan, routes, etc. remain the same)
async def process_summary_task( user_id: int, chat_id: int, message_id_to_edit: Optional[int], url: str, summary_type: str, bot_token: str ) -> None:
    task_id = f"{user_id}-{message_id_to_edit or 'new'}"; logger.info(f"[Task {task_id}] Starting processing for URL: {url}")
    background_request: Optional[BaseRequest] = None; bot: Optional[Bot] = None
    try: background_request = HTTPXRequest( connect_timeout=15.0, read_timeout=60.0, write_timeout=60.0, pool_timeout=60.0 ); bot = Bot(token=bot_token, request=background_request)
    except Exception as e: logger.critical(f"[Task {task_id}] Failed to create background bot: {e}", exc_info=True); return
    content = None; user_feedback_message = None; success = False; status_message_id = message_id_to_edit; message_to_delete_later_id : Optional[int] = None
    try:
        processing_message_text = f"Got it! Generating '{summary_type}' summary for:\n`{url}`\n\nThis might take a moment..."
        if status_message_id:
            try: await retry_bot_operation( bot.edit_message_text, chat_id=chat_id, message_id=status_message_id, text=processing_message_text, parse_mode=ParseMode.MARKDOWN, reply_markup=None ); logger.debug(f"[Task {task_id}] Edited message {status_message_id} to 'Processing'")
            except Exception as e: logger.warning(f"[Task {task_id}] Could not edit original message {status_message_id}: {e}. Sending new."); status_message_id = None
        if not status_message_id:
             try:
                 status_message = await retry_bot_operation( bot.send_message, chat_id=chat_id, text=processing_message_text, parse_mode=ParseMode.MARKDOWN )
                 if status_message: message_to_delete_later_id = status_message.message_id; logger.debug(f"[Task {task_id}] Sent new status message {message_to_delete_later_id}")
                 else: raise RuntimeError("Failed to send status message after retries.")
             except Exception as e: logger.error(f"[Task {task_id}] CRITICAL: Failed to send new status message: {e}"); raise
        try:
            await retry_bot_operation(bot.send_chat_action, chat_id=chat_id, action='typing')
            is_youtube = is_youtube_url(url); logger.debug(f"[Task {task_id}] URL type: {'YouTube' if is_youtube else 'Website'}")
            if is_youtube:
                video_id = extract_youtube_id(url)
                if video_id: content = await get_youtube_transcript(video_id, url)
                else: user_feedback_message = "Sorry, I couldn't understand that YouTube URL format."
                if not content and not user_feedback_message: user_feedback_message = "Sorry, I couldn't get the transcript for that YouTube video using any available method (unavailable/private/no captions?)."
            else:
                content = await get_website_content(url)
                if not content:
                    logger.warning(f"[Task {task_id}] Primary web scrape failed for {url}. Trying fallback API.")
                    global URLTOTEXT_API_KEY
                    if URLTOTEXT_API_KEY:
                         await retry_bot_operation(bot.send_chat_action, chat_id=chat_id, action='typing')
                         content = await get_website_content_via_api(url, URLTOTEXT_API_KEY)
                         if not content: user_feedback_message = "Sorry, I couldn't fetch content from that website using either method (blocked/inaccessible/empty?)."
                    else: user_feedback_message = "Sorry, I couldn't fetch content from that website (blocked/inaccessible/empty?). The fallback method is not configured."
            if content:
                logger.info(f"[Task {task_id}] Content fetched (len:{len(content)}). Generating summary.")
                await retry_bot_operation(bot.send_chat_action, chat_id=chat_id, action='typing')
                final_summary = await generate_summary(content, summary_type) # This now calls Gemini first, then OpenRouter
                if final_summary.startswith("Error:") or final_summary.startswith("Sorry,"): user_feedback_message = final_summary; logger.warning(f"[Task {task_id}] Summary generation failed: {final_summary}")
                else:
                    max_length = 4096; summary_parts = [final_summary[i:i+max_length] for i in range(0, len(final_summary), max_length)]
                    # Sending potentially unescaped summary directly - assuming AI follows instructions
                    # ParseMode.MARKDOWN might interpret unintended things, test needed. Consider ParseMode=None if issues arise.
                    await retry_bot_operation( bot.send_message, chat_id=chat_id, text=summary_parts[0], parse_mode=None, link_preview_options={'is_disabled': True} )
                    for part in summary_parts[1:]: await asyncio.sleep(0.5); await retry_bot_operation( bot.send_message, chat_id=chat_id, text=part, parse_mode=None, link_preview_options={'is_disabled': True} )
                    success = True; logger.info(f"[Task {task_id}] Successfully sent summary ({len(summary_parts)} parts)."); user_feedback_message = None
            elif user_feedback_message: logger.warning(f"[Task {task_id}] Sending failure feedback: {user_feedback_message}"); await retry_bot_operation( bot.send_message, chat_id=chat_id, text=user_feedback_message, link_preview_options={'is_disabled': True} )
        except Exception as e:
            logger.error(f"[Task {task_id}] Unexpected error during processing: {e}", exc_info=True); user_feedback_message = "Oops! Something went really wrong. Please try again later."
            try: await retry_bot_operation( bot.send_message, chat_id=chat_id, text=user_feedback_message )
            except Exception: logger.error(f"[Task {task_id}] Failed to send unexpected error feedback.")
    except Exception as outer_e:
        logger.critical(f"[Task {task_id}] Critical outer error: {outer_e}", exc_info=True)
        try:
             if bot: await retry_bot_operation( bot.send_message, chat_id=chat_id, text="❌ Critical internal error occurred." )
        except Exception: logger.exception(f"[Task {task_id}] Failed even to send critical error message.")
    finally:
        delete_target_id = message_to_delete_later_id if message_to_delete_later_id else status_message_id
        if delete_target_id and bot:
            try: await retry_bot_operation(bot.delete_message, chat_id=chat_id, message_id=delete_target_id); logger.debug(f"[Task {task_id}] Deleted status/button message {delete_target_id}")
            except Exception as del_e: logger.warning(f"[Task {task_id}] Failed to delete status/button message {delete_target_id}: {del_e}")
        if background_request and hasattr(background_request, '_client') and background_request._client:
             try: await background_request._client.aclose(); logger.debug(f"[Task {task_id}] Background bot's HTTPX client closed.")
             except Exception as e: logger.warning(f"[Task {task_id}] Error closing background bot's client: {e}")
        logger.info(f"[Task {task_id}] Task completed. Success: {success}")

async def start(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
    user = update.effective_user; mention = user.mention_html()
    if not user or not update.message: return
    logger.info(f"User {user.id} used /start.")
    await update.message.reply_html( f"πŸ‘‹ Hello {mention}! I can summarise YouTube links or website URLs.\n\nJust send me a link anytime!" ) # Changed 'summarize'

async def help_command(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
    user = update.effective_user
    if not user or not update.message: return
    logger.info(f"User {user.id} used /help.")
    help_text = ( "πŸ” How to use this bot:\n\n" "1. Send me any YouTube video link or website URL.\n" "2. I'll ask you how you want it summarised (paragraph or points).\n" # Changed 'summarized'
                  "3. Click the button for your choice.\n" "4. Wait for the summary!\n\n" "I'll try multiple methods to get content if the first one fails (especially for YouTube transcripts).\n\n" "Commands:\n" "`/start` - Display welcome message\n" "`/help` - Show this help message" )
    await update.message.reply_text(help_text, parse_mode=ParseMode.MARKDOWN)

async def handle_potential_url(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
    if not update.message or not update.message.text: return
    url = update.message.text.strip(); user = update.effective_user
    if not user: return
    if not (url.startswith('http://') or url.startswith('https://')) or '.' not in url[8:]: logger.debug(f"Ignoring non-URL from {user.id}: {url}"); return
    logger.info(f"User {user.id} sent potential URL: {url}")
    context.user_data['url_to_summarize'] = url; context.user_data['original_message_id'] = update.message.message_id
    keyboard = [[ InlineKeyboardButton("Paragraph Summary", callback_data="paragraph"), InlineKeyboardButton("Points Summary", callback_data="points") ]]
    reply_markup = InlineKeyboardMarkup(keyboard)
    await update.message.reply_text( f"Okay, I see this link:\n{url}\n\nHow would you like it summarised?", reply_markup=reply_markup, disable_web_page_preview=True ) # Changed 'summarized'

async def handle_summary_type_callback(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
    query = update.callback_query
    if not query or not query.message or not query.from_user: logger.warning("Callback query missing data."); return
    user = query.from_user; summary_type = query.data; query_id = query.id
    try: await query.answer(); logger.debug(f"Ack callback {query_id} from {user.id}")
    except Exception as e: logger.error(f"Error answering callback {query_id}: {e}", exc_info=True)
    url = context.user_data.get('url_to_summarize'); message_id_to_edit = query.message.message_id
    logger.info(f"User {user.id} chose '{summary_type}' for msg {message_id_to_edit}. URL in context: {'Yes' if url else 'No'}")
    if not url:
        logger.warning(f"No URL in context for user {user.id} (cb {query_id}).")
        try: await query.edit_message_text(text="Sorry, I couldn't find the URL associated with this request. Please send the link again.")
        except Exception as e:
            logger.error(f"Failed edit 'URL not found' msg: {e}")
            try: await context.bot.send_message(chat_id=user.id, text="Sorry, context lost. Send link again.")
            except Exception: pass
        return

    context.user_data.pop('url_to_summarize', None); context.user_data.pop('original_message_id', None); logger.debug(f"Cleared URL context for user {user.id}")

    global TELEGRAM_TOKEN, GEMINI_API_KEY, OPENROUTER_API_KEY, _gemini_primary_enabled, _openrouter_fallback_enabled
    if not TELEGRAM_TOKEN:
        logger.critical("TG TOKEN missing!")
        try: await query.edit_message_text(text="❌ Bot config error.")
        except Exception: pass
        return
    if not _gemini_primary_enabled and not _openrouter_fallback_enabled:
        logger.critical("Neither Gemini nor OpenRouter API keys are configured/valid!")
        try: await query.edit_message_text(text="❌ AI config error: No models available.")
        except Exception: pass
        return
    elif not _gemini_primary_enabled:
         logger.warning("Primary AI (Gemini) is unavailable, will rely on fallback.")
         # No need to inform user unless fallback also fails later
    elif not _openrouter_fallback_enabled:
         logger.warning("Fallback AI (OpenRouter) is unavailable.")
         # No need to inform user unless primary fails later

    logger.info(f"Scheduling task for user {user.id}, chat {query.message.chat_id}, msg {message_id_to_edit}")
    asyncio.create_task( process_summary_task( user_id=user.id, chat_id=query.message.chat_id, message_id_to_edit=message_id_to_edit, url=url, summary_type=summary_type, bot_token=TELEGRAM_TOKEN ), name=f"SummaryTask-{user.id}-{message_id_to_edit}" )

async def error_handler(update: object, context: ContextTypes.DEFAULT_TYPE) -> None:
    ignore_errors = (AttributeError, )
    if isinstance(context.error, ignore_errors) and "object has no attribute" in str(context.error): logger.debug(f"Ignoring known/handled error in error_handler: {context.error}"); return
    logger.error("Exception while handling an update:", exc_info=context.error)

async def setup_bot_config() -> Application:
    logger.info("Configuring Telegram Application..."); global TELEGRAM_TOKEN
    if not TELEGRAM_TOKEN: raise ValueError("TELEGRAM_TOKEN missing.")
    custom_request = HTTPXRequest( connect_timeout=10.0, read_timeout=30.0, write_timeout=30.0, pool_timeout=60.0 )
    application = Application.builder().token(TELEGRAM_TOKEN).request(custom_request).build()
    application.add_handler(CommandHandler("start", start)); application.add_handler(CommandHandler("help", help_command))
    application.add_handler(MessageHandler(filters.TEXT & ~filters.COMMAND, handle_potential_url)); application.add_handler(CallbackQueryHandler(handle_summary_type_callback))
    application.add_error_handler(error_handler); logger.info("Telegram application handlers configured."); return application

@contextlib.asynccontextmanager
async def lifespan(app: Starlette):
    global ptb_app, WEBHOOK_SECRET, TELEGRAM_TOKEN
    logger.info("ASGI Lifespan: Startup initiated...");
    if not TELEGRAM_TOKEN: logger.critical("TG TOKEN missing."); raise RuntimeError("Telegram token missing.")
    try:
        ptb_app = await setup_bot_config(); await ptb_app.initialize(); bot_info = await ptb_app.bot.get_me(); logger.info(f"Bot initialized: @{bot_info.username} (ID: {bot_info.id})")
        current_webhook_info = await ptb_app.bot.get_webhook_info()
        if current_webhook_info and current_webhook_info.url:
            logger.info(f"Found existing webhook: {current_webhook_info.url}. Deleting...")
            try:
                if await ptb_app.bot.delete_webhook(drop_pending_updates=True): logger.info("Webhook deleted.")
                else: logger.warning("Failed delete webhook (API returned False).")
            except Exception as e: logger.warning(f"Could not delete webhook: {e}"); await asyncio.sleep(1)
        space_host = os.environ.get("SPACE_HOST"); webhook_path = "/webhook"; full_webhook_url = None
        if space_host:
            protocol = "https"; host = space_host.split('://')[-1]; full_webhook_url = f"{protocol}://{host.rstrip('/')}{webhook_path}"
            if full_webhook_url:
                logger.info(f"Setting webhook: {full_webhook_url}"); set_webhook_args = { "url": full_webhook_url, "allowed_updates": Update.ALL_TYPES, "drop_pending_updates": True }
                if WEBHOOK_SECRET: set_webhook_args["secret_token"] = WEBHOOK_SECRET; logger.info("Using webhook secret.")
                await asyncio.sleep(1.0)
                try:
                    await ptb_app.bot.set_webhook(**set_webhook_args); webhook_info = await ptb_app.bot.get_webhook_info()
                    if webhook_info.url == full_webhook_url: logger.info(f"Webhook set: URL='{webhook_info.url}', Secret={bool(WEBHOOK_SECRET)}")
                    else: logger.error(f"Webhook URL mismatch! Expected '{full_webhook_url}', Got '{webhook_info.url}'"); raise RuntimeError("Webhook URL mismatch.")
                    await ptb_app.start(); logger.info("PTB Application started (webhook mode).")
                except Exception as e: logger.error(f"FATAL: Failed set webhook: {e}", exc_info=True); raise RuntimeError(f"Failed set webhook: {e}") from e
            else: logger.critical("Could not construct webhook URL."); raise RuntimeError("Webhook URL undetermined.")
        else: logger.critical("SPACE_HOST missing."); raise RuntimeError("SPACE_HOST env var missing.")
        logger.info("ASGI Lifespan: Startup complete."); yield
    except Exception as startup_err:
        logger.critical(f"Application startup failed: {startup_err}", exc_info=True)
        if ptb_app:
             if ptb_app.running: await ptb_app.stop()
             await ptb_app.shutdown()
        raise
    finally:
        logger.info("ASGI Lifespan: Shutdown initiated...")
        if ptb_app:
            if ptb_app.running: logger.info("Stopping PTB..."); await ptb_app.stop()
            logger.info("Shutting down PTB..."); await ptb_app.shutdown(); logger.info("PTB Application shut down.")
        else: logger.info("PTB application not initialized or failed.")
        logger.info("ASGI Lifespan: Shutdown complete.")

async def health_check(request: Request) -> PlainTextResponse:
    global OPENROUTER_MODEL, GEMINI_MODEL, APIFY_ACTOR_ID, _apify_token_exists, _gemini_primary_enabled, _openrouter_fallback_enabled
    bot_status = "Not Initialized"
    if ptb_app and ptb_app.bot:
        try:
             if ptb_app.running: bot_info = await ptb_app.bot.get_me(); bot_status = f"Running (@{bot_info.username})"
             else: bot_status = "Initialized/Not running"
        except Exception as e: bot_status = f"Error checking status: {e}"
    return PlainTextResponse( f"TG Bot Summariser - Status: {bot_status}\n" # Changed 'Summarizer'
                              f"Primary Model: {GEMINI_MODEL if _gemini_primary_enabled else 'N/A (Disabled)'}\n"
                              f"Fallback Model: {OPENROUTER_MODEL if _openrouter_fallback_enabled else 'N/A (Disabled)'}\n"
                              f"Apify Actor: {APIFY_ACTOR_ID if _apify_token_exists else 'N/A (No Token)'}" )

async def telegram_webhook(request: Request) -> Response:
    global WEBHOOK_SECRET
    if not ptb_app: logger.error("Webhook recv but PTB not initialized."); return PlainTextResponse('Bot not initialized', status_code=503)
    if not ptb_app.running: logger.warning("Webhook recv but PTB not running."); return PlainTextResponse('Bot not running', status_code=503)
    try:
        if WEBHOOK_SECRET:
             token_header = request.headers.get("X-Telegram-Bot-Api-Secret-Token")
             if token_header != WEBHOOK_SECRET: logger.warning(f"Webhook invalid secret. Header: '{token_header}'"); return Response(content="Invalid secret token", status_code=403)
        update_data = await request.json(); update = Update.de_json(data=update_data, bot=ptb_app.bot); logger.debug(f"Processing update_id: {update.update_id} via webhook")
        await ptb_app.process_update(update); return Response(status_code=200) # OK
    except json.JSONDecodeError: logger.error("Webhook invalid JSON."); return PlainTextResponse('Bad Request: Invalid JSON', status_code=400)
    except Exception as e: logger.error(f"Error processing webhook update: {e}", exc_info=True); return Response(status_code=200) # OK despite error

app = Starlette( debug=False, lifespan=lifespan, routes=[ Route("/", endpoint=health_check, methods=["GET"]), Route("/webhook", endpoint=telegram_webhook, methods=["POST"]), ] )
logger.info("Starlette ASGI application created with native routes.")

if __name__ == '__main__':
    import uvicorn
    logger.warning("Running in development mode using Uvicorn directly")
    log_level = os.environ.get("LOGGING_LEVEL", "info").lower()
    local_port = int(os.environ.get('PORT', 8080))
    uvicorn.run("__main__:app", host='0.0.0.0', port=local_port, log_level=log_level, reload=True)