File size: 35,748 Bytes
2a8faae
 
 
 
 
 
 
 
2587e2c
2a8faae
 
 
 
ddc9c77
 
2587e2c
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0a5dcf9
 
 
2a8faae
 
 
 
 
 
 
ddc9c77
 
2a8faae
ddc9c77
 
 
 
 
 
 
 
 
 
 
 
0a5dcf9
2a8faae
 
 
 
 
ddc9c77
 
2a8faae
ddc9c77
2a8faae
 
 
 
 
 
 
 
 
ddc9c77
0a5dcf9
ddc9c77
 
 
 
0a5dcf9
ddc9c77
 
0a5dcf9
 
 
ddc9c77
 
 
 
 
 
 
2a8faae
ddc9c77
2a8faae
ddc9c77
2a8faae
 
 
 
ddc9c77
 
 
 
 
 
2a8faae
 
 
 
ddc9c77
2a8faae
ddc9c77
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
 
 
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2587e2c
2a8faae
2587e2c
 
2a8faae
2587e2c
 
 
 
 
 
 
 
 
 
 
 
 
2a8faae
 
2587e2c
2a8faae
2587e2c
 
 
 
2a8faae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
import re
import csv
from concurrent.futures import ThreadPoolExecutor
from datetime import datetime
from typing import Optional, List

import pytz
from langchain.schema import Document, HumanMessage, SystemMessage
from langchain.tools import tool
from .retrievers import hybrid_search, vector_search, bm25_search
from .validation import validate_medical_answer
from .github_storage import get_github_storage
from .context_enrichment import enrich_retrieved_documents
from .config import logger
from langchain_openai import ChatOpenAI

CANONICAL_PROVIDERS = {"Manus", "ASCO", "NCCN", "ESMO", "NICE"}

# Global variables to store context for validation
_last_question = None  # Stores the tool query
_last_user_question = None  # Stores the original user question
_last_documents = None
_last_answer = None

TOOL_MAX_WORKERS = max(2, min(8, (os.cpu_count() or 4)))
_tool_executor = ThreadPoolExecutor(max_workers=TOOL_MAX_WORKERS)

def store_user_question(user_question: str):
    """Store the original user question for validation purposes."""
    global _last_user_question
    _last_user_question = user_question

def _get_llm_safe(temperature: float = 0.0, model: str = "gpt-4o"):
    """Create a ChatOpenAI client if API key/config is available, else return None."""
    try:
        # ChatOpenAI will read OPENAI_API_KEY from env as in validation.py
        return ChatOpenAI(model=model, temperature=temperature, max_tokens=512, request_timeout=30)
    except Exception:
        return None

def _is_side_effect_report_llm(user_input: str) -> Optional[bool]:
    """Use LLM to classify if input is an adverse drug reaction/side-effect report.
    Returns True/False if confident, or None if unavailable/uncertain.
    """
    llm = _get_llm_safe()
    if not llm:
        return None
    try:
        system = SystemMessage(content=(
            "You are a medical triage classifier. Decide if the user's text is a report of an adverse drug reaction (side effect) about a medication.\n"
            "Criteria: mentions a medication/drug and symptoms or adverse effects experienced by a patient.\n"
            "Respond with exactly one token: yes or no."
        ))
        human = HumanMessage(content=user_input[:1500])
        resp = llm.invoke([system, human])
        ans = (resp.content or "").strip().lower()
        if ans.startswith("yes"):
            return True
        if ans.startswith("no"):
            return False
        return None
    except Exception:
        return None

# Map lowercase variants and full names to canonical provider codes
_PROVIDER_ALIASES = {
    # NCCN
    "nccn": "NCCN",
    "national comprehensive cancer network": "NCCN",
    "nccn guidelines": "NCCN",
    # ESMO
    "esmo": "ESMO",
    "european society for medical oncology": "ESMO",
    "esmo guidelines": "ESMO",
    # ASCO
    "asco": "ASCO",
    "american society of clinical oncology": "ASCO",
    "asco guidelines": "ASCO",
    # NICE
    "nice": "NICE",
    "national institute for health and care excellence": "NICE",
    "nice guidelines": "NICE",
    # Manus (custom provider)
    "manus": "Manus",
    "by manus": "Manus",
}


def _normalize_provider_from_text(text: str) -> Optional[str]:
    if not text:
        return None
    t = text.lower()
    # Quick direct hits for canonical providers
    for canon in CANONICAL_PROVIDERS:
        if re.search(rf"\b{re.escape(canon.lower())}\b", t):
            return canon
    # Alias-based detection
    for alias, canon in _PROVIDER_ALIASES.items():
        if alias in t:
            return canon
    return None


def _normalize_provider(provider: Optional[str], query: str) -> Optional[str]:
    # If explicit provider given, normalize it first
    if provider:
        p = provider.strip().lower()
        # Exact canonical match
        for canon in CANONICAL_PROVIDERS:
            if p == canon.lower():
                return canon
        # Alias match
        if p in _PROVIDER_ALIASES:
            return _PROVIDER_ALIASES[p]
        # Try to find within text like "according to NCCN guidelines"
        norm = _normalize_provider_from_text(provider)
        if norm:
            return norm
    # Fall back to inferring from query text
    return _normalize_provider_from_text(query)


def clear_text(text: str, max_chars: int = 1200) -> str:
    """Reduce token bloat by removing heavy markdown and collapsing whitespace.
    - Convert [title](url) -> title (url)
    - Remove images ![alt](url)
    - Strip code fences/backticks and most markdown emphasis
    - Collapse multiple newlines/spaces
    - Trim to max_chars
    """
    if not text:
        return ""
    t = text
    # Normalize newlines
    t = t.replace("\r\n", "\n").replace("\r", "\n")
    # Links: keep title and URL
    t = re.sub(r"\[([^\]]+)\]\(([^)]+)\)", r"\1 (\2)", t)
    # Images: drop entirely
    t = re.sub(r"!\[[^\]]*\]\([^)]*\)", "", t)
    # Remove headers/quotes markers at line starts
    t = re.sub(r"(?m)^[>\s]*#{1,6}\s*", "", t)
    # Remove backticks/code fences and emphasis
    t = t.replace("```", "").replace("`", "")
    t = t.replace("**", "").replace("*", "").replace("_", "")
    # Collapse spaces before newlines
    t = re.sub(r"[ \t]+\n", "\n", t)
    # Collapse multiple newlines and spaces
    t = re.sub(r"\n{3,}", "\n\n", t)
    t = re.sub(r"[ \t]{2,}", " ", t)
    # Trim and truncate
    t = t.strip()
    if max_chars and len(t) > max_chars:
        t = t[:max_chars].rstrip() + " ..."
    return t

def _format_docs_with_citations(docs: List[Document]) -> str:
    if not docs:
        return "No results."
    
    parts = []
    for i, d in enumerate(docs, start=1):
        meta = d.metadata or {}
        source = meta.get("source", "unknown")
        page = meta.get("page_number", "?")
        provider = meta.get("provider", "unknown")
        disease = meta.get("disease", "unknown")
        is_context = meta.get("context_enrichment", False)
        
        snippet = clear_text(d.page_content)
        
        # Build citation header
        citation = f"Result {i}:\n"
        citation += f"Provider: {provider} | Disease: {disease} | Source: {source} | Page: {page}"
        
        # Add context enrichment marker if this is a context page
        if is_context:
            citation += " [CONTEXT PAGE]"
        
        citation += f"\nText:\n{snippet}\n"
        parts.append(citation)
    
    return "\n\n".join(parts)


@tool
def medical_guidelines_knowledge_tool(query: str, provider: Optional[str] = None) -> str:
    """
    Retrieve comprehensive medical guideline knowledge with enriched context.
    Includes surrounding pages (before/after) for top results to provide complete clinical context.
    If provider is provided (e.g., "NCCN", "ASCO", "ESMO", "NICE"), results will be filtered by metadata provider.
    Returns detailed text with full metadata and contextual information for expert analysis.
    """
    global _last_question, _last_documents
    try:
        # Store question for validation context
        _last_question = query
        
        # Normalize provider name from either explicit arg or query text
        normalized_provider = _normalize_provider(provider, query)
        
        # Use hybrid search with query expansion for comprehensive retrieval
        # Uses global defaults: DEFAULT_K_VECTOR=10, DEFAULT_K_BM25=5 (configurable in core/retrievers.py)
        docs = hybrid_search(query=query, provider=normalized_provider)
        
        # Enrich top documents with surrounding pages for richer context
        # This provides complete clinical context including adjacent information
        # Increased pages_before/after and max_enriched for more comprehensive answers
        enriched_docs = enrich_retrieved_documents(
            documents=docs,
            pages_before=2,  # Include 2 pages before for fuller context
            pages_after=2,   # Include 2 pages after for fuller context
            max_enriched=8   # Enrich top 8 most relevant documents
        )
        
        # Count context pages added
        context_pages_count = sum(1 for doc in enriched_docs if doc.metadata.get("context_enrichment", False))
        logger.info(f"Retrieved {len(docs)} documents, added {context_pages_count} context pages")
        
        # Store documents for validation context with enrichment metadata
        _last_documents = []
        for doc in enriched_docs:
            doc_dict = {
                "doc_id": getattr(doc, 'id', None),
                "source": doc.metadata.get("source", "unknown"),
                "provider": doc.metadata.get("provider", "unknown"),
                "page_number": doc.metadata.get("page_number", "unknown"),
                "disease": doc.metadata.get("disease", "unknown"),
                "context_enrichment": doc.metadata.get("context_enrichment", False),
                "enriched": doc.metadata.get("enriched", False),
                "pages_included": doc.metadata.get("pages_included", []),
                "primary_page": doc.metadata.get("primary_page"),
                "context_pages_before": doc.metadata.get("context_pages_before"),
                "context_pages_after": doc.metadata.get("context_pages_after"),
                "content": doc.page_content
            }
            _last_documents.append(doc_dict)
        
        return _format_docs_with_citations(enriched_docs)
    except Exception as e:
        logger.error(f"Retrieval error: {str(e)}")
        return f"Retrieval error: {str(e)}"


@tool
def compare_providers_tool(query: str, provider_a: str, provider_b: str) -> str:
    """
    Compare guideline answers between two providers (e.g., provider_a="NCCN", provider_b="ESMO").
    Retrieves provider-filtered results independently, then returns a structured text block suited for comparison.
    Output includes citations (source file, page number, provider, disease) for each side.
    """
    try:
        canon_a = _normalize_provider(provider_a, query) or provider_a
        canon_b = _normalize_provider(provider_b, query) or provider_b

        a_future = _tool_executor.submit(hybrid_search, query, canon_a, 5, 5)
        b_future = _tool_executor.submit(hybrid_search, query, canon_b, 5, 5)

        a_docs = a_future.result()
        b_docs = b_future.result()

        format_a_future = _tool_executor.submit(_format_docs_with_citations, a_docs)
        format_b_future = _tool_executor.submit(_format_docs_with_citations, b_docs)

        a_text = format_a_future.result()
        b_text = format_b_future.result()

        return (
            f"Comparison for query: {query}\n\n"
            f"Provider A: {canon_a}\n"
            f"{'-'*40}\n"
            f"{a_text}\n\n"
            f"Provider B: {canon_b}\n"
            f"{'-'*40}\n"
            f"{b_text}\n"
        )
    except Exception as e:
        return f"Comparison retrieval error: {str(e)}"



@tool
def get_current_datetime_tool() -> str:
    """
    Returns the current date, time, and day of the week for Egypt (Africa/Cairo).
    This is the only reliable source for date and time information. Use this tool
    whenever a user asks about 'today', 'now', or any other time-sensitive query.
    The output is always in English and in standard 12-hour format.
    """
    try:
        # Define the timezone for Egypt
        egypt_tz = pytz.timezone('Africa/Cairo')
        
        # Get the current time in that timezone
        now_egypt = datetime.now(egypt_tz)
        
        # Manual mapping to ensure English output regardless of system locale
        days_en = {
            0: "Monday", 1: "Tuesday", 2: "Wednesday", 3: "Thursday",
            4: "Friday", 5: "Saturday", 6: "Sunday"
        }
        months_en = {
            1: "January", 2: "February", 3: "March", 4: "April",
            5: "May", 6: "June", 7: "July", 8: "August",
            9: "September", 10: "October", 11: "November", 12: "December"
        }
        
        # Get English names using manual mapping
        day_name = days_en[now_egypt.weekday()]
        month_name = months_en[now_egypt.month]
        day = now_egypt.day
        year = now_egypt.year
        
        # Format time manually to avoid locale issues
        hour = now_egypt.hour
        minute = now_egypt.minute
        
        # Convert to 12-hour format
        if hour == 0:
            hour_12 = 12
            period = "AM"
        elif hour < 12:
            hour_12 = hour
            period = "AM"
        elif hour == 12:
            hour_12 = 12
            period = "PM"
        else:
            hour_12 = hour - 12
            period = "PM"
        
        time_str = f"{hour_12:02d}:{minute:02d} {period}"
        
        # Create the final string
        return f"Current date and time in Egypt: {day_name}, {month_name} {day}, {year} at {time_str}"
    
    except Exception as e:
        return f"Error getting current datetime: {str(e)}"


@tool
def side_effect_recording_tool(user_input: str) -> str:
    """
    Detects when a doctor reports or mentions discovering a side effect related to a drug.
    First asks for missing critical information (drug name, side effects) and optional details
    (patient_age, patient_gender, dosage, duration, severity). If user cannot provide optional
    information, saves the report with NaN values for unknown data.
    
    This tool should be used when the input contains:
    - Reports of adverse drug reactions or side effects
    - Patient experiencing unexpected symptoms after medication
    - Drug-related complications or adverse events
    - Medical professionals reporting medication issues
    
    Args:
        user_input (str): The doctor's input describing the side effect or adverse reaction
        
    Returns:
        str: Interactive form for collecting missing information or confirmation of data recording
    """
    try:
        # LLM classification (preferred), with keyword fallback to preserve behavior
        side_effect_keywords = [
            'side effect', 'adverse reaction', 'adverse event', 'drug reaction',
            'medication reaction', 'allergic reaction', 'complication', 'toxicity',
            'intolerance', 'hypersensitivity', 'contraindication', 'withdrawal',
            'overdose', 'poisoning', 'drug-induced', 'medication-induced',
            'experienced after taking', 'developed after', 'caused by medication',
            'drug-related', 'medication-related', 'pharmaceutical reaction',
            'kidney problems', 'liver problems', 'heart problems', 'breathing problems',
            'skin problems', 'stomach problems', 'nausea', 'vomiting', 'diarrhea',
            'headache', 'dizziness', 'fatigue', 'weakness', 'rash', 'swelling',
            'pain', 'fever', 'cough', 'infection', 'bleeding', 'bruising',
            'has these', 'has serious', 'causes', 'resulted in', 'led to',
            'problems with', 'issues with', 'complications from'
        ]
        input_lower = user_input.lower().strip()
        llm_decision = _is_side_effect_report_llm(user_input)
        
        # Check for special commands first
        if input_lower in ['save report', 'save', 'submit report', 'submit']:
            # Create minimal data for saving
            extracted_data = _extract_side_effect_data(user_input)
            return _save_side_effect_report(extracted_data)
        
        if input_lower in ['cancel', 'cancel report', 'abort']:
            return "**Side Effect Report Cancelled**\n\nThe adverse drug reaction report has been cancelled and no data was saved."
        
        # Check if this is a follow-up with additional information or user saying they can't provide info
        if _is_followup_response(user_input):
            # For follow-up responses, we need to get the base data from somewhere
            # Since we don't have session state, treat this as a new report
            extracted_data = _extract_side_effect_data(user_input)
            return _process_followup_response(user_input, extracted_data)
        
        # Combine LLM decision with keyword fallback to avoid behavior regression
        keyword_detected = any(keyword in input_lower for keyword in side_effect_keywords)
        contains_side_effect = (llm_decision is True) or (llm_decision is not False and keyword_detected)
        
        if not contains_side_effect:
            return "This input does not appear to contain a side effect report. If you are reporting an adverse drug reaction, please include specific details about the medication and symptoms."
        
        # Extract information using pattern matching and keyword analysis
        extracted_data = _extract_side_effect_data(user_input)
        
        # Check if we have the critical information (drug name and side effects)
        missing_critical = _identify_missing_information(extracted_data)
        
        if missing_critical:
            # Missing critical info, ask for it first
            return _generate_information_request(extracted_data, missing_critical)
        else:
            # Have critical info, now ask for optional information
            missing_optional = _identify_missing_optional_information(extracted_data)
            if missing_optional:
                return _generate_optional_information_request(extracted_data, missing_optional)
            else:
                # Have all available info, save the report
                return _save_side_effect_report(extracted_data)
        
    except Exception as e:
        return f"Error processing side effect report: {str(e)}. Please ensure your report includes drug name and symptoms."


def _is_followup_response(user_input: str) -> bool:
    """Check if the input appears to be a follow-up response with additional information."""
    followup_indicators = [
        'patient age:', 'age:', 'gender:', 'dosage:', 'dose:', 'duration:', 
        'severity:', 'outcome:', 'additional:', 'reporter:', 'notes:',
        'male', 'female', 'years old', 'mg', 'ml', 'tablets', 'capsules',
        'mild', 'moderate', 'severe', 'recovered', 'ongoing', 'hospitalized',
        # Add indicators for when user can't provide info
        "can't provide", "cannot provide", "don't have", "do not have", 
        "not available", "unavailable", "missing", "no information",
        "just save", "save them", "save it", "save anyway", "will not provide",
        "won't provide", "don't know", "unknown", "not sure"
    ]
    input_lower = user_input.lower()
    return any(indicator in input_lower for indicator in followup_indicators)


def _process_followup_response(user_input: str, base_data: dict) -> str:
    """Process follow-up response and update the extracted data."""
    
    # Check if user is indicating they can't provide information
    cant_provide_indicators = [
        "can't provide", "cannot provide", "don't have", "do not have", 
        "not available", "unknown", "unavailable", "missing", "no information",
        "will not provide", "won't provide", "save anyway", "just save"
    ]
    
    input_lower = user_input.lower()
    if any(indicator in input_lower for indicator in cant_provide_indicators):
        # User can't provide additional info, save with what we have
        return _save_side_effect_report(base_data)
    
    # Extract additional information from the follow-up
    additional_data = _extract_side_effect_data(user_input)
    
    # Merge with base data, prioritizing new information
    merged_data = base_data.copy()
    for key, value in additional_data.items():
        if value and value != 'NaN' and str(value).strip():
            merged_data[key] = value
    
    # Check if there are still critical missing fields (only drug_name and side_effects are truly critical)
    critical_missing = []
    truly_critical_fields = ['drug_name', 'side_effects']
    
    for field in truly_critical_fields:
        value = merged_data.get(field, '')
        if not value or value == 'NaN' or not str(value).strip():
            critical_missing.append(field)
    
    # If critical information is missing, ask for it
    if critical_missing:
        return _generate_information_request(merged_data, [(field, field.replace('_', ' ').title()) for field in critical_missing])
    
    # Always save automatically after processing follow-up information
    # This ensures we save after any follow-up response, whether complete or partial
    return _save_side_effect_report(merged_data)


def _identify_missing_information(extracted_data: dict) -> list:
    """Identify which critical information is missing from the extracted data."""
    missing = []
    
    # Only truly critical fields - drug name and side effects
    critical_fields = {
        'drug_name': 'Drug/Medication Name',
        'side_effects': 'Side Effects/Symptoms'
    }
    
    for field, display_name in critical_fields.items():
        value = extracted_data.get(field, '')
        if not value or value == 'NaN' or not value.strip():
            missing.append((field, display_name))
    
    return missing


def _identify_missing_optional_information(extracted_data: dict) -> list:
    """Identify which optional information is missing from the extracted data."""
    missing = []
    
    # Optional fields that we should ask for
    optional_fields = {
        'patient_age': 'Patient Age',
        'patient_gender': 'Patient Gender', 
        'dosage': 'Medication Dosage',
        'duration': 'Treatment Duration',
        'severity': 'Severity Level'
    }
    
    for field, display_name in optional_fields.items():
        value = extracted_data.get(field, '')
        if not value or value == 'NaN' or not value.strip():
            missing.append((field, display_name))
    
    return missing


# Remove this function as we no longer ask for optional information


def _generate_information_request(extracted_data: dict, missing_info: list) -> str:
    """Generate a medical-professional request for missing critical information."""
    
    # Only ask for truly critical missing information
    critical_missing = []
    for field, display_name in missing_info:
        if field in ['drug_name', 'side_effects']:
            critical_missing.append((field, display_name))
    
    if not critical_missing:
        # No critical info missing, save the report
        return _save_side_effect_report(extracted_data)
    
    # Create a concise request for only critical missing information
    response = "**Adverse Drug Reaction Report**\n\n"
    
    if any(field == 'drug_name' for field, _ in critical_missing):
        response += "Please specify the **medication/drug name** involved in this adverse reaction.\n\n"
    
    if any(field == 'side_effects' for field, _ in critical_missing):
        response += "Please describe the **side effects or symptoms** experienced.\n\n"
    
    response += "**Note**: All other details (age, gender, dosage, etc.) are optional. If you cannot provide them, I'll save the report with the available information."
    
    return response.strip()


def _generate_optional_information_request(extracted_data: dict, missing_optional: list) -> str:
    """Generate a request for optional information that would enhance the side effect report."""
    
    # Show what we already have
    response = "**Adverse Drug Reaction Report**\n\n"
    response += "**Recorded Information:**\n"
    
    if extracted_data.get('drug_name') and extracted_data['drug_name'] != 'NaN':
        response += f"- **Drug:** {extracted_data['drug_name']}\n"
    
    if extracted_data.get('side_effects') and extracted_data['side_effects'] != 'NaN':
        response += f"- **Side Effects:** {extracted_data['side_effects']}\n"
    
    response += "\n**Additional Information (Optional):**\n"
    response += "To enhance this report, please provide any of the following details if available:\n\n"
    
    for field, display_name in missing_optional:
        if field == 'patient_age':
            response += "- **Patient Age:** (e.g., 45 years old)\n"
        elif field == 'patient_gender':
            response += "- **Patient Gender:** (Male/Female)\n"
        elif field == 'dosage':
            response += "- **Dosage:** (e.g., 10mg daily, 2 tablets)\n"
        elif field == 'duration':
            response += "- **Duration:** (e.g., 3 months, 2 weeks)\n"
        elif field == 'severity':
            response += "- **Severity:** (Mild/Moderate/Severe)\n"
    
    response += "\n**Note:** If you don't have this information or cannot provide it, just reply with \"I don't have that information\" or \"save anyway\" and I'll save the report with the available data."
    
    return response.strip()


def _save_side_effect_report(extracted_data: dict) -> str:
    """Save the side effect report to CSV file."""
    try:
        # Ensure all fields have values (use 'NaN' for empty fields)
        fieldnames = [
            'timestamp', 'drug_name', 'side_effects', 'patient_age', 
            'patient_gender', 'dosage', 'duration', 'severity', 
            'outcome', 'additional_details', 'reporter_info', 'raw_input'
        ]
        
        # Fill missing fields with 'NaN' and ensure proper data types
        for field in fieldnames:
            value = extracted_data.get(field, '')
            if not value or value == '' or not str(value).strip():
                extracted_data[field] = 'NaN'
            else:
                # Ensure the value is properly formatted
                extracted_data[field] = str(value).strip()
        
        # Save to GitHub repository (fallback to local if needed)
        github_storage = get_github_storage()
        success = github_storage.save_side_effects_report(extracted_data)
        if not success:
            csv_filename = "side_effects_reports.csv"
            csv_path = os.path.join(os.getcwd(), csv_filename)
            file_exists = os.path.exists(csv_path)
            with open(csv_path, 'a', newline='', encoding='utf-8') as csvfile:
                writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
                if not file_exists:
                    writer.writeheader()
                writer.writerow(extracted_data)
            storage_location = "locally to side_effects_reports.csv (GitHub upload failed)"
        else:
            storage_location = "to GitHub cloud repository"

        # Generate confirmation message
        drug_name = extracted_data.get('drug_name', 'NaN')
        side_effects = extracted_data.get('side_effects', 'NaN')
        report_id = extracted_data['timestamp'].replace(':', '').replace('-', '').replace(' ', '_')
        
        # Create a summary of provided vs missing information
        provided_info = []
        missing_info = []
        
        info_fields = {
            'drug_name': 'Drug/Medication',
            'side_effects': 'Side Effects', 
            'patient_age': 'Patient Age',
            'patient_gender': 'Patient Gender',
            'dosage': 'Dosage',
            'duration': 'Duration',
            'severity': 'Severity',
            'outcome': 'Outcome'
        }
        
        for field, display_name in info_fields.items():
            value = extracted_data.get(field, 'NaN')
            if value and value != 'NaN':
                provided_info.append(f"- **{display_name}:** {value}")
            else:
                missing_info.append(display_name)
        
        confirmation = f"""
**✅ Adverse Drug Reaction Report Saved**

**Report ID:** {report_id}

**Documented Information:**
{chr(10).join(provided_info) if provided_info else '- Basic side effect report recorded'}

**Pharmacovigilance Status:** Report successfully saved {storage_location} for regulatory review.

**Clinical Recommendations:**
- Monitor patient for symptom progression
- Consider dose adjustment or alternative therapy if appropriate
- Document in patient medical record
- Report serious reactions to pharmacovigilance authorities

How can I assist you further with clinical guidance for this case?
"""
        
        return confirmation.strip()
        
    except Exception as e:
        return f"Error saving side effect report: {str(e)}"


def _extract_side_effect_data_with_llm(user_input: str) -> dict:
    """
    Extract structured data from side effect report text using LLM-based extraction.
    
    Args:
        user_input (str): Raw input text containing side effect report
        
    Returns:
        dict: Structured data extracted from the input
    """
    import json
    # Get current timestamp
    egypt_tz = pytz.timezone('Africa/Cairo')
    current_time = datetime.now(egypt_tz).strftime('%Y-%m-%d %H:%M:%S')
    # Initialize extracted data with defaults
    extracted_data = {
        'timestamp': current_time,
        'drug_name': 'NaN',
        'side_effects': 'NaN',
        'patient_age': 'NaN',
        'patient_gender': 'NaN',
        'dosage': 'NaN',
        'duration': 'NaN',
        'severity': 'NaN',
        'outcome': 'NaN',
        'additional_details': 'NaN',
        'reporter_info': 'NaN',
        'raw_input': user_input[:500]
    }
    llm = _get_llm_safe()
    if llm:
        try:
            system = SystemMessage(content=(
                "Extract medical side effect information. Return ONLY a JSON object with these exact fields: "
                "drug_name, side_effects, patient_age, patient_gender, dosage, duration, severity, outcome. "
                "If missing/unclear, use 'NaN'."
            ))
            human = HumanMessage(content=user_input[:2000])
            response = llm.invoke([system, human])
            text = (response.content or "").strip()
            # Try parse; if fails, fallback regex
            try:
                extracted_json = json.loads(text)
            except json.JSONDecodeError:
                extracted_json = _extract_with_improved_regex(user_input)
        except Exception:
            extracted_json = _extract_with_improved_regex(user_input)
    else:
        extracted_json = _extract_with_improved_regex(user_input)
    # Update extracted_data
    for key, value in extracted_json.items():
        if key in extracted_data and value and str(value).strip() and str(value).strip().lower() != 'nan':
            extracted_data[key] = str(value).strip()
    return extracted_data


def _extract_with_improved_regex(user_input: str) -> dict:
    """
    Improved regex-based extraction with better duration handling.
    """
    extracted = {
        'drug_name': 'NaN',
        'side_effects': 'NaN',
        'patient_age': 'NaN',
        'patient_gender': 'NaN',
        'dosage': 'NaN',
        'duration': 'NaN',
        'severity': 'NaN',
        'outcome': 'NaN'
    }
    
    input_lower = user_input.lower()
    
    # Extract drug names with improved patterns
    drug_patterns = [
        r'\b(afinitor|cisplatin|afatinib|imatinib|dasatinib|nilotinib|bosutinib|ponatinib|bevacizumab|cetuximab|trastuzumab)\b',
        r'(?:found that|that)\s+([A-Za-z]{4,20})\s+(?:has|have)',
        r'([A-Za-z]{4,20})\s+(?:has|have)\s+(?:these\s+)?(?:side\s+effects?|adverse\s+effects?)',
        r'(?:taking|prescribed|given|on)\s+([A-Za-z][A-Za-z0-9\s\-]{2,20}?)(?:\s+(?:mg|mcg|g|ml)|\s+for|\.|,)',
        r'(?:side effects?|adverse effects?)\s+(?:of|from)\s+([A-Za-z][A-Za-z0-9\s\-]{2,20}?)(?:\s|,|\.|;)'
    ]
    
    for pattern in drug_patterns:
        matches = re.findall(pattern, user_input, re.IGNORECASE)
        if matches:
            drug_name = matches[0].strip()
            if len(drug_name) > 2 and not drug_name.lower() in ['that', 'these', 'those', 'found', 'have']:
                extracted['drug_name'] = drug_name
                break
    
    # Extract side effects
    symptom_patterns = [
        r'(?:side effects?|symptoms?|adverse effects?)\s*[:\-]?\s*([^.!?]+?)(?:\.|!|\?|patient|$)',
        r'(?:has|have)\s+(?:these\s+)?(?:side effects?[:\s]+)?([A-Za-z][^.!?]*?)(?:\.|!|\?|patient|$)',
        r'(?:experienced|developed|suffered|had)\s+([^.!?]+?)(?:\.|!|\?|after|following|$)'
    ]
    
    for pattern in symptom_patterns:
        matches = re.findall(pattern, user_input, re.IGNORECASE)
        if matches:
            symptoms = matches[0].strip()
            if len(symptoms) > 3:
                extracted['side_effects'] = symptoms
                break
    
    # Extract patient age with better patterns
    age_patterns = [
        r'patient\'?s?\s+age\s*[:\-]?\s*(\d{1,3})',
        r'age\s*[:\-]?\s*(\d{1,3})',
        r'(\d{1,3})\s*(?:years?\s+old|y/?o)',
        r'aged\s+(\d{1,3})'
    ]
    
    for pattern in age_patterns:
        matches = re.findall(pattern, user_input, re.IGNORECASE)
        if matches:
            age = int(matches[0])
            if 0 <= age <= 120:
                extracted['patient_age'] = str(age)
                break
    
    # Extract patient gender
    if re.search(r'\b(?:male|man|boy|gentleman|he|his|him)\b', input_lower):
        extracted['patient_gender'] = 'Male'
    elif re.search(r'\b(?:female|woman|girl|lady|she|her)\b', input_lower):
        extracted['patient_gender'] = 'Female'
    
    # Extract dosage
    dosage_patterns = [
        r'(?:medication\s+)?dosage\s*[:\-]?\s*([\d\.]+\s*(?:mg|mcg|g|ml|units?|tablets?|capsules?))',
        r'dosage\s*[:\-]?\s*([\d\.]+\s*ml)',
        r'(\d+(?:\.\d+)?\s*(?:mg|mcg|g|ml|units?|tablets?|capsules?))'
    ]
    
    for pattern in dosage_patterns:
        matches = re.findall(pattern, user_input, re.IGNORECASE)
        if matches:
            extracted['dosage'] = matches[0].strip()
            break
    
    # Extract duration with improved patterns
    duration_patterns = [
        r'treatment\s+duration\s*[:\-]?\s*(\d+\s*(?:days?|weeks?|months?|years?))',
        r'duration\s*[:\-]?\s*(\d+\s*(?:days?|weeks?|months?|years?))',
        r'(?:for|over|during)\s+(\d+\s*(?:days?|weeks?|months?|years?))',
        r'(\d+\s*(?:days?|weeks?|months?|years?))\s+(?:of\s+)?(?:treatment|therapy)',
        r'(?:lasted|continuing for|ongoing for)\s+(\d+\s*(?:days?|weeks?|months?|years?))'
    ]
    
    for pattern in duration_patterns:
        matches = re.findall(pattern, user_input, re.IGNORECASE)
        if matches:
            extracted['duration'] = matches[0].strip()
            break
    
    # Extract severity
    severity_keywords = {
        'mild': ['mild', 'slight', 'minor', 'light'],
        'moderate': ['moderate', 'medium', 'noticeable'],
        'severe': ['severe', 'serious', 'major', 'significant', 'intense', 'extreme']
    }
    
    for severity, keywords in severity_keywords.items():
        if any(keyword in input_lower for keyword in keywords):
            extracted['severity'] = severity.capitalize()
            break
    
    # Extract outcome
    outcome_keywords = {
        'recovered': ['recovered', 'resolved', 'better', 'improved'],
        'ongoing': ['ongoing', 'continuing', 'persistent', 'current status: ongoing'],
        'worsened': ['worsened', 'deteriorated', 'worse'],
        'hospitalized': ['hospitalized', 'admitted', 'emergency']
    }
    
    for outcome, keywords in outcome_keywords.items():
        if any(keyword in input_lower for keyword in keywords):
            extracted['outcome'] = outcome.capitalize()
            break
    
    return extracted


def _extract_side_effect_data(user_input: str) -> dict:
    """
    Extract structured data from side effect report text.
    
    Args:
        user_input (str): Raw input text containing side effect report
        
    Returns:
        dict: Structured data extracted from the input
    """
    # Use the new LLM-based extraction
    return _extract_side_effect_data_with_llm(user_input)