File size: 40,321 Bytes
720dfc2
188f6b3
f8e1a1c
 
 
b3f045c
f8e1a1c
219c443
8587347
219c443
f8e1a1c
e0f232f
 
 
 
 
 
 
4e60f97
 
e0f232f
720dfc2
 
 
 
dbfeb5e
 
 
 
 
05ed9d3
 
 
 
 
f8e1a1c
 
 
e0f232f
f8e1a1c
 
 
 
 
 
 
 
720dfc2
 
f8e1a1c
720dfc2
dbfeb5e
 
720dfc2
f8e1a1c
 
 
 
 
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
05ed9d3
 
 
f8e1a1c
05197b2
 
 
 
 
 
 
 
 
 
 
 
 
 
f8e1a1c
 
 
 
 
 
b3f045c
f8e1a1c
b3f045c
 
f8e1a1c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
720dfc2
219c443
f8e1a1c
 
 
 
720dfc2
f8e1a1c
 
720dfc2
f8e1a1c
e0f232f
 
 
 
 
 
f8e1a1c
 
720dfc2
f8e1a1c
 
e0f232f
 
 
f8e1a1c
c167dff
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
b3f045c
f8e1a1c
 
e0f232f
 
 
f8e1a1c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e0f232f
 
 
 
 
 
f8e1a1c
 
 
 
 
 
4e60f97
 
 
 
 
 
 
 
 
 
720dfc2
 
b1d3be1
 
 
 
 
720dfc2
 
 
899bfab
 
 
 
 
 
 
 
 
 
 
 
 
 
 
720dfc2
 
f8e1a1c
a7f0469
 
 
b3f045c
e0f232f
 
 
 
 
b3f045c
e0f232f
 
b3f045c
e0f232f
 
 
b3f045c
 
 
e0f232f
 
 
 
f8e1a1c
 
 
 
e0f232f
 
 
 
 
 
 
 
 
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
 
b3f045c
 
 
 
 
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e0f232f
 
 
 
 
 
 
f8e1a1c
 
 
e0f232f
 
 
219c443
e0f232f
f8e1a1c
720dfc2
 
f8e1a1c
720dfc2
 
f8e1a1c
720dfc2
 
 
b3f045c
 
 
 
 
 
 
f8e1a1c
720dfc2
 
b3f045c
 
 
f8e1a1c
b3f045c
 
 
 
 
 
 
 
 
 
720dfc2
 
e0f232f
b3f045c
 
 
 
 
 
 
 
 
 
e0f232f
 
720dfc2
e0f232f
 
 
f8e1a1c
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
219c443
 
 
 
720dfc2
 
 
f8e1a1c
720dfc2
f8e1a1c
720dfc2
f8e1a1c
 
720dfc2
b3f045c
 
 
 
 
 
f8e1a1c
 
e0f232f
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
b3f045c
 
 
 
 
f8e1a1c
 
 
 
 
 
 
b3f045c
f8e1a1c
 
b3f045c
 
 
 
 
 
e0f232f
f8e1a1c
 
 
 
e0f232f
4e60f97
 
 
 
 
 
 
e0f232f
 
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4e60f97
 
 
 
 
 
 
 
 
f8e1a1c
 
 
 
 
 
 
 
 
720dfc2
f8e1a1c
 
 
 
 
e0f232f
 
 
f8e1a1c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
219c443
f8e1a1c
 
 
 
e0f232f
 
 
 
 
 
f8e1a1c
 
 
 
 
 
e0f232f
f8e1a1c
 
 
 
 
 
 
 
 
219c443
f8e1a1c
 
 
 
 
 
 
720dfc2
f8e1a1c
 
 
 
05ed9d3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9b7783e
 
b3f045c
 
 
 
9b7783e
 
 
 
b3f045c
 
 
 
 
 
 
9b7783e
8587347
 
 
 
 
 
 
 
 
9b7783e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b3f045c
9b7783e
 
 
 
 
219c443
9b7783e
 
 
 
 
 
b3f045c
9b7783e
b3f045c
 
 
9b7783e
 
 
 
 
8587347
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b3f045c
 
 
8587347
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b3f045c
 
 
8587347
 
 
 
 
 
 
b3f045c
f8e1a1c
b3f045c
 
f8e1a1c
 
e0f232f
 
 
f8e1a1c
b3f045c
 
 
 
f8e1a1c
 
b3f045c
 
 
 
f8e1a1c
e0f232f
 
 
b3f045c
 
 
 
 
f8e1a1c
720dfc2
 
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
219c443
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
219c443
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
219c443
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
219c443
b3f045c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
720dfc2
188f6b3
e0f232f
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
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
import logging
import os
import time
import json
import random
import hashlib
import tempfile
from typing import cast
import requests as _http
from openai.types.chat import ChatCompletionMessageParam
from collections import defaultdict
from flask import (
    Flask,
    render_template,
    request,
    jsonify,
    Response,
    stream_with_context,
    send_from_directory,
    make_response,
)
from openai import OpenAI
from config import Config
from flask_cors import CORS

try:
    from auth_module.flask_auth_routes import auth_blueprint
except Exception:
    auth_blueprint = None

try:
    from auth_module.security import decode_access_token as _decode_token
except Exception:
    _decode_token = None

# Groq SDK for Whisper
try:
    from groq import Groq

    GROQ_AVAILABLE = True
except ImportError:
    GROQ_AVAILABLE = False

# --- Logging ---
logging.basicConfig(
    level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s"
)

app = Flask(__name__)
CORS(app)
app.config.from_object(Config)
if auth_blueprint is not None:
    app.register_blueprint(auth_blueprint, url_prefix="/auth")

START_TIME = time.time()

# --- Rate Limiter (in-memory) ---
rate_limits = defaultdict(list)

# --- In-memory response cache (TTL-based) ---
_response_cache: dict = {}  # key -> {"response": ..., "expires_at": float}
CACHE_TTL_SECONDS = 300  # 5 minutes


def _cache_key(*parts: str) -> str:
    """Create a deterministic cache key from multiple parts."""
    raw = "|".join(str(p) for p in parts)
    return hashlib.md5(raw.encode()).hexdigest()


def _cache_get(key: str):
    """Return cached value or None if missing/expired."""
    entry = _response_cache.get(key)
    if entry and entry["expires_at"] > time.time():
        return entry["value"]
    if key in _response_cache:
        del _response_cache[key]
    return None


def _cache_set(key: str, value, ttl: int = CACHE_TTL_SECONDS):
    """Store value in cache with TTL."""
    # Prune old entries if cache is getting large
    if len(_response_cache) > 500:
        now = time.time()
        expired = [k for k, v in _response_cache.items() if v["expires_at"] <= now]
        for k in expired:
            del _response_cache[k]
    _response_cache[key] = {"value": value, "expires_at": time.time() + ttl}


# --- In-memory leaderboard ---
_leaderboard: dict = {}  # keyed by email (from JWT) or IP


def get_client_ip():
    forwarded = (request.headers.get("X-Forwarded-For", "") or "").strip()
    if forwarded:
        first_ip = forwarded.split(",", 1)[0].strip()
        if first_ip:
            return first_ip

    real_ip = (request.headers.get("X-Real-IP", "") or "").strip()
    if real_ip:
        return real_ip

    return request.remote_addr or "unknown"


def is_rate_limited(ip):
    now = time.time()
    window = 60
    limit = app.config.get("RATE_LIMIT", 30)
    rate_limits[ip] = [t for t in rate_limits[ip] if now - t < window]
    if len(rate_limits[ip]) >= limit:
        return True, limit, 0
    rate_limits[ip].append(now)
    remaining = limit - len(rate_limits[ip])
    return False, limit, remaining


# --- OpenAI Client Factory (supports Groq key rotation) ---
def get_client():
    provider = Config.AI_PROVIDER
    if provider == "groq" and Config.GROQ_API_KEYS:
        api_key = random.choice(Config.GROQ_API_KEYS)
        base_url = Config.PROVIDERS["groq"]["base_url"]
    elif Config.NVIDIA_API_KEY:
        api_key = Config.NVIDIA_API_KEY
        base_url = Config.PROVIDERS["nvidia"]["base_url"]
    else:
        return None
    try:
        return OpenAI(base_url=base_url, api_key=api_key)
    except Exception as e:
        logging.error(f"Failed to create client: {e}")
        return None


# --- Groq Client for Whisper ---
def get_groq_client():
    if not GROQ_AVAILABLE or not Config.GROQ_API_KEYS:
        return None
    try:
        api_key = random.choice(Config.GROQ_API_KEYS)
        return Groq(api_key=api_key)
    except Exception as e:
        logging.error(f"Failed to create Groq client: {e}")
        return None


# Test initial connection
_test_client = get_client()
if _test_client:
    logging.info(f"AI Provider: {Config.AI_PROVIDER} | Model: {Config.MODEL_NAME}")
else:
    logging.warning("No AI provider configured. AI features unavailable.")


# --- Load Exercises ---
EXERCISES = {}
try:
    exercises_path = os.path.join(app.static_folder or "", "data", "exercises.json")
    with open(exercises_path, "r", encoding="utf-8") as f:
        EXERCISES = json.load(f)
    total = sum(len(v) for v in EXERCISES.values())
    logging.info(f"Loaded {total} exercises across {len(EXERCISES)} types")
except Exception as e:
    logging.warning(f"Could not load exercises: {e}")


# --- System Prompt Builder (with language + scenario) ---
def build_system_prompt(
    level="intermediate", topic="free", language="en", scenario=None
):
    level_config = Config.DIFFICULTY_LEVELS.get(
        level, Config.DIFFICULTY_LEVELS["intermediate"]
    )
    lang_config = Config.LANGUAGES.get(language, Config.LANGUAGES["en"])
    lang_name = lang_config["label"]

    # Scenario mode overrides regular topic
    if scenario:
        scenario_config = next(
            (s for s in Config.SCENARIOS if s["id"] == scenario), None
        )
        if scenario_config:
            extra = ""
            if scenario == "ielts_speaking":
                extra = (
                    "\n\nIELTS SPEAKING FORMAT:\n"
                    "- Part 1 (4-5 min): Ask familiar topic questions (hometown, work, hobbies)\n"
                    "- Part 2 (3-4 min): Give a cue card topic, allow 1 min prep, then user speaks 2 min\n"
                    "- Part 3 (4-5 min): Abstract discussion related to Part 2 topic\n"
                    "After each response, give a brief band-score tip (e.g. 'To improve your score, try using more linking words like however, although...')\n"
                    "Give a final band score estimate at the end.\n"
                )
            elif scenario == "toefl_speaking":
                extra = (
                    "\n\nTOEFL SPEAKING FORMAT:\n"
                    "- Independent Task: Opinion question (15s prep, 45s response)\n"
                    "- After each response, give feedback on: Task completion, Delivery, Language use, Topic development\n"
                    "- Estimate a TOEFL speaking score (0-30) after the practice task.\n"
                )
            return (
                f"You are playing the role of {scenario_config['ai_role']}. "
                f"The user is {scenario_config['user_role']}. "
                f"Stay in character throughout the conversation. "
                f"If the user makes language mistakes, gently correct them while staying in character. "
                f"Respond naturally as your character would.\n\n"
                f"LANGUAGE: Speak in {lang_name}.\n"
                f"DIFFICULTY: {level_config['label']} β€” {level_config['description']}.\n"
                f"DO NOT use emoji symbols. Speak naturally. Keep responses concise (2-4 sentences).\n"
                f"You MAY use **bold** for emphasis.\n" + extra
            )

    topic_config = next(
        (t for t in Config.TOPICS if t["id"] == topic), Config.TOPICS[0]
    )

    base = (
        f"You are {lang_config['tutor_name']}, a friendly and expert {lang_name} language tutor. "
        f"Your goal is to help the user practice and improve their {lang_name} through natural conversation.\n\n"
        "RULES:\n"
        "1. **Engage**: Keep the conversation going by asking follow-up questions.\n"
        "2. **Correct Gently**: If the user makes a grammatical mistake, provide a correction naturally "
        "within the conversation. Don't correct every single minor error, only ones that affect meaning. "
        "Example: 'By the way, instead of X, you could say Y β€” it sounds more natural.'\n"
        "3. **Suggest Improvements**: Offer better vocabulary or phrasing as part of your natural response.\n"
        "4. **Formatting**: Keep your responses clean and readable. "
        "DO NOT use emoji symbols because the response is read aloud by text-to-speech. "
        "DO NOT use labels like 'Correction:' or 'Suggestion:' β€” instead weave corrections into your natural reply. "
        "You MAY use **bold** for emphasis on key words.\n"
        "5. **Be Concise**: Keep responses conversational β€” 2-4 sentences max for simple exchanges.\n"
        "6. **Encourage**: Use positive reinforcement.\n"
        "7. **Sound Human**: Write the way a real friendly tutor would speak.\n"
    )

    level_instruction = (
        f"\n**DIFFICULTY**: {level_config['label']} β€” {level_config['description']}.\n"
    )
    topic_instruction = (
        f"**TOPIC**: {topic_config['label']} β€” {topic_config['prompt']}\n"
    )
    lang_instruction = f"**LANGUAGE**: Speak and teach in {lang_name}.\n"

    return base + level_instruction + topic_instruction + lang_instruction


# --- Routes ---
@app.route("/service-worker.js")
def service_worker():
    response = make_response(send_from_directory("static", "service-worker.js"))
    response.headers["Content-Type"] = "application/javascript"
    response.headers["Cache-Control"] = "no-cache, no-store, must-revalidate"
    response.headers["Pragma"] = "no-cache"
    response.headers["Expires"] = "0"
    return response


@app.route("/")
def index():
    return render_template("landing.html")


@app.route("/app")
def app_page():
    return render_template("index.html")


@app.route("/login")
def login_page():
    return render_template("login.html")


@app.route("/signup")
def signup_page():
    return render_template("signup.html")


@app.route("/verify")
def verify_page():
    return render_template("verify.html")


@app.route("/health")
def health():
    uptime = int(time.time() - START_TIME)
    whisper_sdk_available = GROQ_AVAILABLE
    whisper_keys_configured = bool(Config.GROQ_API_KEYS)
    whisper_available = whisper_sdk_available and whisper_keys_configured
    ai_client = get_client()
    return (
        jsonify(
            {
                "status": "ok",
                "uptime_seconds": uptime,
                "uptime_human": f"{uptime // 3600}h {(uptime % 3600) // 60}m",
                "provider": Config.AI_PROVIDER,
                "model": Config.MODEL_NAME,
                "ai_available": ai_client is not None,
                "whisper_available": whisper_available,
                "whisper_sdk_available": whisper_sdk_available,
                "whisper_keys_configured": whisper_keys_configured,
                "cache_entries": len(_response_cache),
                "leaderboard_entries": len(_leaderboard),
                "version": "2.2.0",
            }
        ),
        200,
    )


@app.route("/topics")
def topics():
    return jsonify(
        {
            "topics": Config.TOPICS,
            "levels": {
                k: {"label": v["label"], "description": v["description"]}
                for k, v in Config.DIFFICULTY_LEVELS.items()
            },
            "languages": Config.LANGUAGES,
            "scenarios": Config.SCENARIOS,
            "badges": Config.BADGES,
        }
    )


# ─── Whisper STT Endpoint ───
@app.route("/transcribe", methods=["POST"])
def transcribe():
    """Transcribe audio using Groq Whisper API."""
    groq_client = get_groq_client()
    if not groq_client:
        return jsonify({"error": "Whisper service unavailable"}), 503

    limited, limit, remaining = is_rate_limited(get_client_ip())
    if limited:
        resp = jsonify({"error": "Too many requests. Please slow down."})
        resp.headers["X-RateLimit-Limit"] = str(limit)
        resp.headers["X-RateLimit-Remaining"] = "0"
        resp.headers["Retry-After"] = "60"
        return resp, 429

    if "audio" not in request.files:
        return jsonify({"error": "No audio file provided"}), 400

    audio_file = request.files["audio"]
    language = request.form.get("language", "en")

    try:
        # Save to temp file (Groq needs a file path)
        with tempfile.NamedTemporaryFile(suffix=".webm", delete=False) as tmp:
            audio_file.save(tmp)
            tmp_path = tmp.name

        with open(tmp_path, "rb") as f:
            transcription = groq_client.audio.transcriptions.create(
                file=(os.path.basename(tmp_path), f.read()),
                model="whisper-large-v3-turbo",
                language=language,
                temperature=0,
                response_format="verbose_json",
            )

        os.unlink(tmp_path)  # Clean up

        return jsonify(
            {
                "text": transcription.text,
                "language": getattr(transcription, "language", language),
                "duration": getattr(transcription, "duration", 0),
            }
        )

    except Exception as e:
        logging.error(f"Transcription error: {e}", exc_info=True)
        if "tmp_path" in locals():
            try:
                os.unlink(tmp_path)
            except Exception:
                pass
        return jsonify({"error": str(e)}), 500


# ─── Chat (non-streaming) ───
@app.route("/chat", methods=["POST"])
def chat():
    client = get_client()
    if not client:
        return jsonify({"error": "AI service is currently unavailable."}), 503

    limited, limit, remaining = is_rate_limited(get_client_ip())
    if limited:
        resp = jsonify({"error": "Too many requests. Please slow down."})
        resp.headers["X-RateLimit-Limit"] = str(limit)
        resp.headers["X-RateLimit-Remaining"] = "0"
        resp.headers["Retry-After"] = "60"
        return resp, 429

    data = request.json or {}
    history = data.get("history", [])
    level = str(data.get("level", "intermediate"))[:20]
    topic = str(data.get("topic", "free"))[:50]
    language = str(data.get("language", "en"))[:5]
    scenario = data.get("scenario", None)
    user_name = str(data.get("user_name", "") or "").strip()[:60]

    # Input validation & sanitization
    if not isinstance(history, list):
        return jsonify({"error": "Invalid history format"}), 400
    if len(history) > 100:
        history = history[-100:]  # Prevent context flooding
    for msg in history:
        if isinstance(msg, dict) and len(str(msg.get("content", ""))) > 4000:
            msg["content"] = str(msg["content"])[:4000]

    if not history:
        return (
            jsonify(
                {
                    "response": get_welcome_message(
                        topic,
                        language,
                        scenario,
                        user_name=user_name,
                    )
                }
            ),
            200,
        )

    level_config = Config.DIFFICULTY_LEVELS.get(
        level, Config.DIFFICULTY_LEVELS["intermediate"]
    )
    system_prompt = build_system_prompt(level, topic, language, scenario)

    # Check cache for identical conversation state
    last_user_msg = next(
        (m["content"] for m in reversed(history) if m.get("role") == "user"), ""
    )
    cache_key = _cache_key(last_user_msg, level, topic, language, scenario or "")
    cached = _cache_get(cache_key)
    if cached:
        resp = jsonify({"response": cached})
        resp.headers["X-Cache"] = "HIT"
        resp.headers["X-RateLimit-Remaining"] = str(remaining)
        return resp

    messages_payload: list[ChatCompletionMessageParam] = cast(
        list[ChatCompletionMessageParam],
        [{"role": "system", "content": system_prompt}] + history,
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=messages_payload,
            temperature=level_config["temperature"],
            top_p=0.95,
            max_tokens=level_config["max_tokens"],
            stream=False,
        )
        reply = completion.choices[0].message.content
        _cache_set(cache_key, reply)
        resp = jsonify({"response": reply})
        resp.headers["X-Cache"] = "MISS"
        resp.headers["X-RateLimit-Remaining"] = str(remaining)
        return resp
    except Exception as e:
        logging.error(f"Error calling API: {e}", exc_info=True)
        return (
            jsonify({"error": "I'm having trouble right now. Please try again."}),
            500,
        )


# ─── Chat (streaming) ───
@app.route("/chat/stream", methods=["POST"])
def chat_stream():
    client = get_client()
    if not client:
        return jsonify({"error": "AI service is currently unavailable."}), 503

    limited, limit, remaining = is_rate_limited(get_client_ip())
    if limited:
        resp = jsonify({"error": "Too many requests. Please slow down."})
        resp.headers["Retry-After"] = "60"
        return resp, 429

    data = request.json or {}
    history = data.get("history", [])
    level = data.get("level", "intermediate")
    topic = data.get("topic", "free")
    language = data.get("language", "en")
    scenario = data.get("scenario", None)
    user_name = str(data.get("user_name", "") or "").strip()[:60]

    if not history:
        welcome = get_welcome_message(
            topic,
            language,
            scenario,
            user_name=user_name,
        )

        def welcome_gen():
            yield f"data: {json.dumps({'token': welcome})}\n\n"
            yield "data: [DONE]\n\n"

        return Response(
            stream_with_context(welcome_gen()),
            content_type="text/event-stream",
            headers={
                "Cache-Control": "no-cache",
                "Connection": "keep-alive",
                "X-Accel-Buffering": "no",
            },
        )

    level_config = Config.DIFFICULTY_LEVELS.get(
        level, Config.DIFFICULTY_LEVELS["intermediate"]
    )
    system_prompt = build_system_prompt(level, topic, language, scenario)
    messages_payload = [{"role": "system", "content": system_prompt}] + history

    def generate():
        try:
            stream_client = get_client()
            if not stream_client:
                yield f"data: {json.dumps({'error': 'AI service unavailable'})}\n\n"
                return

            stream = stream_client.chat.completions.create(
                model=Config.MODEL_NAME,
                messages=messages_payload,
                temperature=level_config["temperature"],
                top_p=0.95,
                max_tokens=level_config["max_tokens"],
                stream=True,
            )
            for chunk in stream:
                if chunk.choices and chunk.choices[0].delta.content:
                    token = chunk.choices[0].delta.content
                    yield f"data: {json.dumps({'token': token})}\n\n"
            yield "data: [DONE]\n\n"
        except Exception as e:
            logging.error(f"Streaming error: {e}", exc_info=True)
            yield f"data: {json.dumps({'error': 'Connection interrupted. Please try again.'})}\n\n"

    return Response(
        stream_with_context(generate()),
        content_type="text/event-stream",
        headers={
            "Cache-Control": "no-cache",
            "Connection": "keep-alive",
            "X-Accel-Buffering": "no",
        },
    )


# ─── Exercises Endpoints ───
@app.route("/exercises")
def get_exercises():
    """Return random exercises filtered by level and type."""
    level = request.args.get("level", "intermediate")
    ex_type = request.args.get("type", "all")
    count = min(int(request.args.get("count", 10)), 20)

    result = []
    types_to_fetch = [ex_type] if ex_type != "all" else list(EXERCISES.keys())

    for t in types_to_fetch:
        if t in EXERCISES:
            filtered = [
                e for e in EXERCISES[t] if e.get("level", "intermediate") == level
            ]
            if not filtered:
                filtered = EXERCISES[t]  # fall back to all levels
            result.extend(filtered)

    random.shuffle(result)
    return jsonify({"exercises": result[:count], "total": len(result)})


@app.route("/exercises/generate", methods=["POST"])
def generate_exercises():
    """AI-generate exercises based on user's past errors."""
    client = get_client()
    if not client:
        return jsonify({"error": "AI unavailable"}), 503

    data = request.json or {}
    errors = data.get("errors", [])
    level = data.get("level", "intermediate")
    _language = data.get("language", "en")  # reserved for future i18n use

    if not errors:
        return jsonify({"error": "No errors provided"}), 400

    errors_text = "\n".join(
        [
            f"- Wrong: '{e.get('wrong','')}' β†’ Right: '{e.get('right','')}'"
            for e in errors[:10]
        ]
    )

    prompt = (
        f"Based on these language mistakes a student made:\n{errors_text}\n\n"
        f"Generate 5 fill-in-the-blank exercises at the {level} level that target these specific error patterns. "
        f"Return ONLY valid JSON array, no other text. Each item should have: "
        f'"q" (question with ___ for blank), "options" (4 choices), "answer" (correct), "explain" (brief explanation). '
        f'Example: {{"q":"She ___ to school.","options":["go","goes","going","gone"],"answer":"goes","explain":"Third person singular."}}'
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.3,
            max_tokens=1024,
        )
        raw = (completion.choices[0].message.content or "").strip()
        # Try to parse JSON from response
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
        exercises = json.loads(raw)
        return jsonify({"exercises": exercises, "generated": True})
    except Exception as e:
        logging.error(f"Exercise generation error: {e}")
        return jsonify({"error": "Could not generate exercises"}), 500


@app.route("/leaderboard", methods=["GET"])
def get_leaderboard():
    top = sorted(_leaderboard.values(), key=lambda x: x["xp"], reverse=True)[:10]
    return jsonify(top)


@app.route("/leaderboard", methods=["POST"])
def submit_leaderboard():
    data = request.json or {}
    key = get_client_ip()
    auth_header = request.headers.get("Authorization", "")
    if auth_header.lower().startswith("bearer ") and _decode_token:
        token = auth_header.split(" ", 1)[1]
        payload = _decode_token(token)
        if payload:
            key = payload.get("sub", key)
    _leaderboard[key] = {
        "name": str(data.get("name", "Anonymous"))[:40],
        "xp": max(0, int(data.get("xp", 0))),
        "level": max(1, int(data.get("level", 1))),
        "flag": str(data.get("flag", "\U0001f30d"))[:8],
    }
    return jsonify({"ok": True})


@app.route("/story/generate", methods=["POST"])
def generate_story():
    story_client = get_client()
    if not story_client:
        return jsonify({"error": "AI service unavailable"}), 503

    data = request.json or {}
    language = data.get("language", "en")
    level = data.get("level", "intermediate")
    topic = data.get("topic", "daily life")

    # Check cache first
    story_key = _cache_key("story", language, level, topic)
    cached_story = _cache_get(story_key)
    if cached_story:
        return jsonify(cached_story)

    lang_names = {
        "en": "English",
        "fr": "French",
        "es": "Spanish",
        "de": "German",
        "ar": "Arabic",
        "it": "Italian",
        "pt": "Portuguese",
        "ja": "Japanese",
        "zh": "Chinese",
    }
    lang_name = lang_names.get(language, "English")
    word_counts = {"beginner": 80, "intermediate": 130, "advanced": 180}
    wc = word_counts.get(level, 130)
    prompt = (
        f"Write a short {lang_name} story for a {level} language learner about '{topic}'. "
        f"The story MUST be written entirely in {lang_name}. "
        f"Target length: {wc} words. Make it engaging and use common vocabulary. "
        f"Return ONLY a valid JSON object with these exact fields:\n"
        f'  "title": story title in {lang_name}\n'
        f'  "story": the full story text in {lang_name}\n'
        f'  "vocabulary": array of 6 key words, each with "word" (in {lang_name}), '
        f'"translation" (English), "example" (short sentence in {lang_name})\n'
        f'  "image_prompt": vivid 12-word English description of the main scene for AI image generation\n'
        f"No markdown, no extra text β€” only the JSON object."
    )
    try:
        completion = story_client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.75,
            max_tokens=1400,
        )
        raw = (completion.choices[0].message.content or "").strip()
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
            raw = raw.strip("`").strip()
        story_data = json.loads(raw)
        _cache_set(story_key, story_data, ttl=600)  # cache stories for 10 min
        return jsonify(story_data)
    except json.JSONDecodeError as e:
        logging.error(f"Story JSON parse error: {e} | raw: {raw[:200]}")
        return jsonify({"error": "Story format error. Please try again."}), 500
    except Exception as e:
        logging.error(f"Story generation error: {e}")
        return jsonify({"error": "Could not generate story"}), 500


@app.route("/story/image", methods=["POST"])
def get_story_image():
    """Fetch a relevant image from Unsplash β†’ Pexels β†’ Pollinations fallback chain."""
    data = request.json or {}
    topic = str(data.get("topic", ""))[:60]
    image_prompt = str(data.get("image_prompt", topic))[:80]
    query = topic or image_prompt

    # 1. Unsplash (highest quality, free API)
    if Config.UNSPLASH_ACCESS_KEY:
        try:
            r = _http.get(
                "https://api.unsplash.com/photos/random",
                params={
                    "query": query,
                    "orientation": "landscape",
                    "count": 1,
                    "content_filter": "high",
                },
                headers={"Authorization": f"Client-ID {Config.UNSPLASH_ACCESS_KEY}"},
                timeout=5,
            )
            if r.ok:
                photos = r.json()
                if isinstance(photos, list) and photos:
                    return jsonify(
                        {"url": photos[0]["urls"]["regular"], "source": "unsplash"}
                    )
        except Exception as e:
            logging.warning(f"Unsplash image fetch failed: {e}")

    # 2. Pexels fallback
    if Config.PEXELS_API_KEY:
        try:
            r = _http.get(
                "https://api.pexels.com/v1/search",
                params={"query": query, "per_page": 1, "orientation": "landscape"},
                headers={"Authorization": Config.PEXELS_API_KEY},
                timeout=5,
            )
            if r.ok:
                d = r.json()
                if d.get("photos"):
                    return jsonify(
                        {"url": d["photos"][0]["src"]["large"], "source": "pexels"}
                    )
        except Exception as e:
            logging.warning(f"Pexels image fetch failed: {e}")

    # 3. Signal client to use Pollinations (no server-side key needed)
    return jsonify({"url": None, "source": "pollinations"})


def get_welcome_message(topic="free", language="en", scenario=None, user_name=""):
    lang_config = Config.LANGUAGES.get(language, Config.LANGUAGES["en"])
    safe_name = str(user_name or "").strip()[:40]
    first_name = safe_name.split()[0] if safe_name else ""

    if scenario:
        scenario_config = next(
            (s for s in Config.SCENARIOS if s["id"] == scenario), None
        )
        if scenario_config:
            opening = scenario_config["opening"]
            if first_name:
                return f"{opening} {first_name}, let's begin."
            return opening

    if topic == "free":
        welcome = lang_config["welcome"]
        if first_name:
            return f"{welcome} Nice to meet you, {first_name}."
        return welcome

    topic_config = next(
        (t for t in Config.TOPICS if t["id"] == topic), Config.TOPICS[0]
    )
    if first_name:
        return (
            f"{lang_config['welcome']} {topic_config['prompt']} "
            f"Ready to start, {first_name}?"
        )
    return f"{lang_config['welcome']} {topic_config['prompt']} Ready to start?"


@app.route("/api/status")
def api_status():
    """Detailed status endpoint for monitoring and debugging."""
    now = time.time()
    uptime = int(now - START_TIME)
    active_ips = len(rate_limits)
    cache_live = sum(1 for v in _response_cache.values() if v["expires_at"] > now)
    return jsonify(
        {
            "status": "operational",
            "version": "2.2.0",
            "uptime_seconds": uptime,
            "uptime_human": f"{uptime // 3600}h {(uptime % 3600) // 60}m {uptime % 60}s",
            "ai_provider": Config.AI_PROVIDER,
            "model": Config.MODEL_NAME,
            "ai_available": get_client() is not None,
            "whisper_available": GROQ_AVAILABLE and bool(Config.GROQ_API_KEYS),
            "cache": {
                "total_entries": len(_response_cache),
                "live_entries": cache_live,
                "ttl_seconds": CACHE_TTL_SECONDS,
            },
            "rate_limiter": {
                "tracked_ips": active_ips,
                "limit_per_minute": app.config.get("RATE_LIMIT", 30),
            },
            "leaderboard_entries": len(_leaderboard),
        }
    )


# ─── Vocabulary Suggestion Endpoint ───
@app.route("/vocab/suggest", methods=["POST"])
def vocab_suggest():
    """AI-powered vocabulary suggestions based on topic, level, and language."""
    client = get_client()
    if not client:
        return jsonify({"error": "AI service unavailable"}), 503

    limited, _, _ = is_rate_limited(get_client_ip())
    if limited:
        return jsonify({"error": "Too many requests. Please slow down."}), 429

    data = request.json or {}
    topic = str(data.get("topic", "daily life"))[:100]
    level = str(data.get("level", "intermediate"))
    language = str(data.get("language", "en"))[:5]
    count = min(int(data.get("count", 8)), 15)

    lang_names = {
        "en": "English",
        "fr": "French",
        "es": "Spanish",
        "de": "German",
        "ar": "Arabic",
        "it": "Italian",
    }
    lang_name = lang_names.get(language, "English")

    cache_key = _cache_key("vocab_suggest", topic, level, language, str(count))
    cached = _cache_get(cache_key)
    if cached:
        resp = jsonify(cached)
        resp.headers["X-Cache"] = "HIT"
        return resp

    prompt = (
        f"Generate {count} useful {lang_name} vocabulary words for a {level} learner studying '{topic}'.\n"
        f"Return ONLY a valid JSON array. Each item must have:\n"
        f'  "word": the word in {lang_name}\n'
        f'  "translation": English translation\n'
        f'  "example": short natural sentence using the word in {lang_name}\n'
        f'  "tip": one memory tip or usage note (in English)\n'
        f"No markdown, no extra text β€” only the JSON array."
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.5,
            max_tokens=900,
        )
        raw = (completion.choices[0].message.content or "").strip()
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
            raw = raw.strip("`").strip()
        words = json.loads(raw)
        result = {"words": words, "topic": topic, "level": level, "language": language}
        _cache_set(cache_key, result, ttl=600)
        return jsonify(result)
    except json.JSONDecodeError as e:
        logging.error(f"Vocab suggest JSON parse error: {e}")
        return jsonify({"error": "Could not parse vocabulary. Try again."}), 500
    except Exception as e:
        logging.error(f"Vocab suggest error: {e}")
        return jsonify({"error": "Could not generate vocabulary."}), 500


# ─── Grammar Check Endpoint ───
@app.route("/grammar/check", methods=["POST"])
def grammar_check():
    """Check a sentence or paragraph for grammar errors and return corrections."""
    client = get_client()
    if not client:
        return jsonify({"error": "AI service unavailable"}), 503

    limited, _, _ = is_rate_limited(get_client_ip())
    if limited:
        return jsonify({"error": "Too many requests. Please slow down."}), 429

    data = request.json or {}
    text = str(data.get("text", "")).strip()
    language = str(data.get("language", "en"))[:5]

    if not text:
        return jsonify({"error": "No text provided"}), 400
    if len(text) > 1000:
        return jsonify({"error": "Text too long (max 1000 characters)"}), 400

    lang_names = {
        "en": "English",
        "fr": "French",
        "es": "Spanish",
        "de": "German",
        "ar": "Arabic",
        "it": "Italian",
    }
    lang_name = lang_names.get(language, "English")

    cache_key = _cache_key("grammar", text[:200], language)
    cached = _cache_get(cache_key)
    if cached:
        resp = jsonify(cached)
        resp.headers["X-Cache"] = "HIT"
        return resp

    prompt = (
        f"You are a {lang_name} grammar expert. Analyze this text and return a grammar check report.\n\n"
        f'Text: "{text}"\n\n'
        f"Return ONLY valid JSON with these fields:\n"
        f'  "is_correct": boolean\n'
        f'  "corrected": the fully corrected version (same as input if no errors)\n'
        f'  "errors": array of {{"original": str, "corrected": str, "explanation": str}}\n'
        f'  "score": integer 1-10 (grammar quality)\n'
        f'  "tip": one concise improvement tip\n'
        f"No markdown, no extra text."
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.1,
            max_tokens=600,
        )
        raw = (completion.choices[0].message.content or "").strip()
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
            raw = raw.strip("`").strip()
        result = json.loads(raw)
        _cache_set(cache_key, result, ttl=300)
        return jsonify(result)
    except json.JSONDecodeError as e:
        logging.error(f"Grammar check JSON parse error: {e}")
        return jsonify({"error": "Could not parse grammar check response."}), 500
    except Exception as e:
        logging.error(f"Grammar check error: {e}")
        return jsonify({"error": "Grammar check failed. Try again."}), 500


# ─── Pronunciation Tip Endpoint ───
@app.route("/pronunciation/tip", methods=["POST"])
def pronunciation_tip():
    """Get pronunciation tips for a word or phrase."""
    client = get_client()
    if not client:
        return jsonify({"error": "AI service unavailable"}), 503

    limited, _, _ = is_rate_limited(get_client_ip())
    if limited:
        return jsonify({"error": "Too many requests."}), 429

    data = request.json or {}
    word = str(data.get("word", "")).strip()[:100]
    language = str(data.get("language", "en"))[:5]

    if not word:
        return jsonify({"error": "No word provided"}), 400

    cache_key = _cache_key("pron_tip", word.lower(), language)
    cached = _cache_get(cache_key)
    if cached:
        resp = jsonify(cached)
        resp.headers["X-Cache"] = "HIT"
        return resp

    lang_names = {
        "en": "English",
        "fr": "French",
        "es": "Spanish",
        "de": "German",
        "ar": "Arabic",
        "it": "Italian",
    }
    lang_name = lang_names.get(language, "English")

    prompt = (
        f'Give a pronunciation guide for the {lang_name} word or phrase: "{word}".\n'
        f"Return ONLY valid JSON:\n"
        f'  "word": the word\n'
        f'  "ipa": IPA phonetic transcription\n'
        f'  "phonetic": simple phonetic spelling (e.g. SEE-ren-DIP-ih-tee)\n'
        f'  "tips": array of 2-3 short tips for native English speakers to pronounce it correctly\n'
        f'  "similar_sound": an English word that has a similar sound (if applicable)\n'
        f"No markdown, no extra text."
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.2,
            max_tokens=400,
        )
        raw = (completion.choices[0].message.content or "").strip()
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
            raw = raw.strip("`").strip()
        result = json.loads(raw)
        _cache_set(cache_key, result, ttl=3600)  # cache 1 hour
        return jsonify(result)
    except Exception as e:
        logging.error(f"Pronunciation tip error: {e}")
        return jsonify({"error": "Could not generate pronunciation tip."}), 500


# ─── Daily Challenge Endpoint ───
@app.route("/challenge/daily", methods=["GET"])
def daily_challenge():
    """Return a daily language challenge based on level and language."""
    level = request.args.get("level", "intermediate")
    language = request.args.get("language", "en")

    client = get_client()
    if not client:
        return jsonify({"error": "AI service unavailable"}), 503

    # Use date as seed for consistent daily challenge
    today = time.strftime("%Y-%m-%d")
    cache_key = _cache_key("daily_challenge", today, level, language)
    cached = _cache_get(cache_key)
    if cached:
        resp = jsonify(cached)
        resp.headers["X-Cache"] = "HIT"
        return resp

    lang_names = {
        "en": "English",
        "fr": "French",
        "es": "Spanish",
        "de": "German",
        "ar": "Arabic",
        "it": "Italian",
    }
    lang_name = lang_names.get(language, "English")

    prompt = (
        f"Create a short daily language challenge for a {level} {lang_name} learner.\n"
        f"Return ONLY valid JSON:\n"
        f'  "title": short catchy title (max 6 words)\n'
        f'  "description": what the learner should do (1-2 sentences)\n'
        f'  "prompt": the exact conversation starter or task prompt to send to the AI tutor\n'
        f'  "xp_reward": integer 15-50 based on difficulty\n'
        f'  "category": one of ["speaking", "vocabulary", "grammar", "roleplay", "pronunciation"]\n'
        f"No markdown, no extra text."
    )

    try:
        completion = client.chat.completions.create(
            model=Config.MODEL_NAME,
            messages=[{"role": "user", "content": prompt}],
            temperature=0.8,
            max_tokens=300,
        )
        raw = (completion.choices[0].message.content or "").strip()
        if raw.startswith("```"):
            raw = raw.split("```")[1]
            if raw.startswith("json"):
                raw = raw[4:]
            raw = raw.strip("`").strip()
        result = json.loads(raw)
        result["date"] = today
        _cache_set(cache_key, result, ttl=86400)  # cache 24 hours
        return jsonify(result)
    except Exception as e:
        logging.error(f"Daily challenge error: {e}")
        return jsonify(
            {
                "title": "Free Talk Practice",
                "description": "Have a natural conversation with your AI tutor.",
                "prompt": "Let's have a natural conversation. Tell me about your day.",
                "xp_reward": 20,
                "category": "speaking",
                "date": today,
            }
        )


if __name__ == "__main__":
    port = int(os.environ.get("PORT", 7860))
    app.run(host="0.0.0.0", port=port, debug=Config.FLASK_DEBUG)