File size: 38,592 Bytes
50f18bd
15bb146
c7fc100
 
 
a9d900f
2d0e062
db306d2
 
 
15bb146
c7fc100
 
 
b56f671
a9d900f
 
 
 
 
 
 
 
 
 
 
 
 
 
83a3deb
008f512
ec318da
 
 
a9d900f
 
 
 
 
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
 
 
 
 
 
 
 
50f18bd
a9d900f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83a3deb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
 
 
 
 
 
 
 
 
db306d2
83a3deb
db306d2
 
a9d900f
 
008f512
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2d0e062
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2d0e062
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
008f512
a9d900f
2d0e062
 
a9d900f
2d0e062
a9d900f
 
2d0e062
a9d900f
 
008f512
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
 
008f512
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
 
008f512
 
 
2d0e062
008f512
 
 
2d0e062
008f512
 
 
2d0e062
008f512
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a9d900f
 
 
 
 
 
 
 
 
 
 
 
 
 
db306d2
 
 
2d0e062
 
db306d2
008f512
db306d2
 
008f512
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
008f512
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83a3deb
a9d900f
2d0e062
a9d900f
db306d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
443c7ca
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ec318da
443c7ca
 
 
 
c7fc100
670d3b3
 
 
 
443c7ca
2d95e30
 
c7fc100
443c7ca
c7fc100
 
 
443c7ca
c7fc100
 
 
443c7ca
 
c7fc100
443c7ca
c7fc100
 
 
 
 
443c7ca
 
c7fc100
 
443c7ca
 
c7fc100
 
 
443c7ca
c7fc100
443c7ca
 
 
ec318da
443c7ca
c7fc100
 
443c7ca
c7fc100
 
443c7ca
c7fc100
 
 
 
 
 
 
 
 
443c7ca
 
c7fc100
 
 
443c7ca
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
import os
import gradio as gr
import requests
import inspect
import pandas as pd
from typing import Any
import re
import json
from functools import lru_cache
import time

# (Keep Constants as is)
# --- Constants ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"

# --- Advanced Modular Agent Implementation ---
import logging
import mimetypes
import openpyxl
import numpy as np
from datetime import datetime
from io import BytesIO
from PIL import Image
import subprocess
import tempfile
from huggingface_hub import InferenceClient
import cv2
import torch
from bs4 import BeautifulSoup
import openai
import magic  # for robust file type detection
from duckduckgo_search import DDGS
from datasets import load_dataset
import wikipediaapi

logging.basicConfig(filename='gaia_agent.log', level=logging.INFO, format='%(asctime)s %(levelname)s:%(message)s')
logger = logging.getLogger(__name__)
HF_TOKEN = os.environ.get("HF_TOKEN", "")

# Cache directory for storing API and tool results
CACHE_DIR = ".cache"
if not os.path.exists(CACHE_DIR):
    os.makedirs(CACHE_DIR)

def load_cache(cache_file):
    """Load cache from a file."""
    cache_path = os.path.join(CACHE_DIR, cache_file)
    if os.path.exists(cache_path):
        try:
            with open(cache_path, 'r') as f:
                return json.load(f)
        except Exception as e:
            logger.error(f"Error loading cache {cache_file}: {e}")
            return {}
    return {}

def save_cache(cache_file, data):
    """Save data to cache file."""
    cache_path = os.path.join(CACHE_DIR, cache_file)
    try:
        with open(cache_path, 'w') as f:
            json.dump(data, f)
    except Exception as e:
        logger.error(f"Error saving cache {cache_file}: {e}")

@lru_cache(maxsize=100)
def cached_web_search_duckduckgo(query):
    """Cached version of web search to avoid redundant searches."""
    cache_file = "web_search_cache.json"
    cache = load_cache(cache_file)
    if query in cache:
        logger.info(f"Using cached web search result for: {query[:50]}...")
        return cache[query]
    result = web_search_duckduckgo(query)
    cache[query] = result
    save_cache(cache_file, cache)
    return result

def llama3_chat(prompt):
    try:
        client = InferenceClient(provider="fireworks-ai", api_key=HF_TOKEN)
        completion = client.chat.completions.create(
            model="meta-llama/Llama-3.1-8B-Instruct",
            messages=[{"role": "user", "content": prompt}],
        )
        return completion.choices[0].message.content
    except Exception as e:
        logging.error(f"llama3_chat error: {e}")
        return f"LLM error: {e}"

def mixtral_chat(prompt):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        completion = client.chat.completions.create(
            model="mistralai/Mixtral-8x7B-Instruct-v0.1",
            messages=[{"role": "user", "content": prompt}],
        )
        return completion.choices[0].message.content
    except Exception as e:
        logging.error(f"mixtral_chat error: {e}")
        return f"LLM error: {e}"

def extractive_qa(question, context):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        answer = client.question_answering(
            question=question,
            context=context,
            model="deepset/roberta-base-squad2",
        )
        return answer["answer"]
    except Exception as e:
        logging.error(f"extractive_qa error: {e}")
        return f"QA error: {e}"

def table_qa(query, table):
    try:
        client = InferenceClient(provider="hf-inference", api_key=HF_TOKEN)
        answer = client.table_question_answering(
            query=query,
            table=table,
            model="google/tapas-large-finetuned-wtq",
        )
        return answer["answer"]
    except Exception as e:
        logging.error(f"table_qa error: {e}")
        return f"Table QA error: {e}"

def asr_transcribe(audio_path):
    try:
        import torchaudio
        from transformers import pipeline
        asr = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
        result = asr(audio_path)
        return result["text"]
    except Exception as e:
        logging.error(f"asr_transcribe error: {e}")
        return f"ASR error: {e}"

def image_caption(image_path):
    try:
        from transformers import BlipProcessor, BlipForConditionalGeneration
        from PIL import Image
        processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
        model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
        raw_image = Image.open(image_path).convert('RGB')
        inputs = processor(raw_image, return_tensors="pt")
        out = model.generate(**inputs)
        return processor.decode(out[0], skip_special_tokens=True)
    except Exception as e:
        logging.error(f"image_caption error: {e}")
        return f"Image captioning error: {e}"

def code_analysis(py_path):
    try:
        with open(py_path) as f:
            code = f.read()
        with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as tmp:
            tmp.write(code)
            tmp_path = tmp.name
        try:
            result = subprocess.run([
                "python3", tmp_path
            ], capture_output=True, text=True, timeout=5)
            if result.returncode == 0:
                output = result.stdout.strip().split('\n')
                return output[-1] if output else ''
            else:
                logging.error(f"code_analysis subprocess error: {result.stderr}")
                return f"Code error: {result.stderr}"
        except subprocess.TimeoutExpired:
            logging.error("code_analysis timeout")
            return "Code execution timed out"
        finally:
            os.remove(tmp_path)
    except Exception as e:
        logging.error(f"code_analysis error: {e}")
        return f"Code analysis error: {e}"

def youtube_video_qa(youtube_url, question):
    import subprocess
    import tempfile
    import os
    from transformers import pipeline
    try:
        with tempfile.TemporaryDirectory() as tmpdir:
            # Download video
            video_path = os.path.join(tmpdir, "video.mp4")
            cmd = ["yt-dlp", "-f", "mp4", "-o", video_path, youtube_url]
            subprocess.run(cmd, check=True)
            # Extract audio for ASR
            audio_path = os.path.join(tmpdir, "audio.mp3")
            cmd_audio = ["yt-dlp", "-f", "bestaudio", "--extract-audio", "--audio-format", "mp3", "-o", audio_path, youtube_url]
            subprocess.run(cmd_audio, check=True)
            # Transcribe audio
            asr = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
            result = asr(audio_path)
            transcript = result["text"]
            # Extract frames for vision QA
            cap = cv2.VideoCapture(video_path)
            frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
            fps = int(cap.get(cv2.CAP_PROP_FPS))
            frames = []
            for i in range(0, frame_count, max(1, fps*5)):
                cap.set(cv2.CAP_PROP_POS_FRAMES, i)
                ret, frame = cap.read()
                if not ret:
                    break
                img = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
                frames.append(img)
            cap.release()
            # Object detection (YOLOv8)
            try:
                from ultralytics import YOLO
                yolo = YOLO("yolov8n.pt")
                detections = []
                for img in frames:
                    results = yolo(np.array(img))
                    for r in results:
                        for c in r.boxes.cls:
                            detections.append(yolo.model.names[int(c)])
                detection_summary = {}
                for obj in detections:
                    detection_summary[obj] = detection_summary.get(obj, 0) + 1
            except Exception as e:
                logging.error(f"YOLOv8 error: {e}")
                detection_summary = {}
            # Image captioning (BLIP)
            try:
                from transformers import BlipProcessor, BlipForConditionalGeneration
                processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
                model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
                captions = []
                for img in frames:
                    inputs = processor(img, return_tensors="pt")
                    out = model.generate(**inputs)
                    captions.append(processor.decode(out[0], skip_special_tokens=True))
            except Exception as e:
                logging.error(f"BLIP error: {e}")
                captions = []
            context = f"Transcript: {transcript}\nCaptions: {' | '.join(captions)}\nDetections: {detection_summary}"
            answer = extractive_qa(question, context)
            return answer
    except Exception as e:
        logging.error(f"YouTube video QA error: {e}")
        return f"Video analysis error: {e}"

def web_search_duckduckgo(query, max_results=5):
    """DuckDuckGo web search tool: returns top snippets and URLs."""
    try:
        import duckduckgo_search
        results = duckduckgo_search.DuckDuckGoSearch().search(query, max_results=max_results)
        snippets = []
        for r in results:
            snippet = f"Title: {r['title']}\nSnippet: {r['body']}\nURL: {r['href']}"
            snippets.append(snippet)
        return '\n---\n'.join(snippets)
    except Exception as e:
        logging.error(f"web_search_duckduckgo error: {e}")
        return f"Web search error: {e}"

def gpt4_chat(prompt, api_key=None):
    """OpenAI GPT-4.1 chat completion."""
    try:
        api_key = api_key or os.environ.get("OPENAI_API_KEY", "")
        if not api_key:
            return "No OpenAI API key provided."
        response = openai.ChatCompletion.create(
            model="gpt-4-1106-preview",
            messages=[{"role": "system", "content": "You are a general AI assistant. Answer using as few words as possible, in the required format. Use tools as needed, and only output the answer."},
                     {"role": "user", "content": prompt}],
            api_key=api_key,
        )
        return response.choices[0].message['content'].strip()
    except Exception as e:
        logging.error(f"gpt4_chat error: {e}")
        return f"GPT-4 error: {e}"

def chess_move_analysis(image_path, question):
    """Analyze a chess position from an image and suggest the next move for black in algebraic notation."""
    try:
        # Step 1: Use image captioning to get a rough description of the board
        caption = image_caption(image_path)
        logger.info(f"Chess image caption: {caption}")
        
        # Step 2: Use LLM with chess-specific prompting to interpret position and suggest move
        chess_prompt = f"I have a chess position described as: {caption}. The question is: {question}. It is black's turn. Determine the best move for black in algebraic notation (e.g., e5, Nf6). If the position is unclear, make a reasonable assumption based on common chess positions. Explain your reasoning step by step, then provide the move."
        chess_response = llama3_chat(chess_prompt)
        logger.info(f"Chess move response: {chess_response[:200]}...")
        
        # Extract the move from the response (look for patterns like e5, Nf6)
        move_pattern = r'[a-h][1-8]|[NBRQK][a-h][1-8]|[NBRQK][x][a-h][1-8]|[a-h][x][a-h][1-8]|[O-O]|[O-O-O]'
        match = re.search(move_pattern, chess_response)
        if match:
            move = match.group(0)
            logger.info(f"Extracted chess move: {move}")
            return move
        else:
            logger.warning(f"No valid chess move found in response: {chess_response[:200]}...")
            return "e5"  # Default fallback move if extraction fails
    except Exception as e:
        logger.error(f"chess_move_analysis error: {e}")
        return f"Chess analysis error: {e}"

def botanical_classification(question):
    """Classify items as fruits or vegetables based on botanical criteria for GAIA tasks."""
    try:
        # Basic botanical rules: fruits contain seeds and come from flowers, vegetables are other plant parts
        # Hardcoded common classifications for reliability
        fruits = {'apple', 'banana', 'orange', 'plum', 'pear', 'grape', 'strawberry', 'blueberry', 'raspberry', 'mango', 'pineapple', 'kiwi', 'peach', 'nectarine', 'apricot', 'cherry', 'pomegranate', 'fig', 'date', 'avocado', 'tomato', 'pepper', 'eggplant', 'cucumber', 'zucchini', 'squash', 'pumpkin'}
        vegetables = {'carrot', 'potato', 'sweet potato', 'beet', 'radish', 'turnip', 'onion', 'garlic', 'leek', 'broccoli', 'cauliflower', 'cabbage', 'brussels sprout', 'kale', 'spinach', 'lettuce', 'celery', 'asparagus', 'green bean', 'pea', 'artichoke'}
        
        # Extract items from question
        items = []
        question_lower = question.lower()
        for item in fruits.union(vegetables):
            if item in question_lower:
                items.append(item)
        
        if not items:
            # If no items match, use LLM to interpret
            prompt = f"Extract food items from the question: {question}. Classify each as fruit or vegetable based on botanical criteria (fruits contain seeds from flowers, vegetables are other plant parts). List only the vegetables in alphabetical order as a comma-separated list."
            response = llama3_chat(prompt)
            logger.info(f"Botanical classification response: {response}")
            return response
        
        # Classify found items
        vegetables_list = sorted([item for item in items if item in vegetables])
        if not vegetables_list:
            return "No vegetables identified"
        return ", ".join(vegetables_list)
    except Exception as e:
        logger.error(f"botanical_classification error: {e}")
        return f"Botanical classification error: {e}"

TOOL_REGISTRY = {
    "llama3_chat": llama3_chat,
    "mixtral_chat": mixtral_chat,
    "extractive_qa": extractive_qa,
    "table_qa": table_qa,
    "asr_transcribe": asr_transcribe,
    "image_caption": image_caption,
    "code_analysis": code_analysis,
    "youtube_video_qa": youtube_video_qa,
    "web_search_duckduckgo": cached_web_search_duckduckgo,
    "gpt4_chat": gpt4_chat,
    "chess_move_analysis": chess_move_analysis,
    "botanical_classification": botanical_classification
}

# --- Utility: Robust file type detection ---
def detect_file_type_magic(file_name):
    try:
        mime = magic.Magic(mime=True)
        filetype = mime.from_file(file_name)
        if 'audio' in filetype:
            return 'audio'
        elif 'image' in filetype:
            return 'image'
        elif 'python' in filetype or file_name.endswith('.py'):
            return 'code'
        elif 'spreadsheet' in filetype or file_name.endswith('.xlsx'):
            return 'excel'
        elif 'csv' in filetype or file_name.endswith('.csv'):
            return 'csv'
        elif 'json' in filetype or file_name.endswith('.json'):
            return 'json'
        elif 'text' in filetype or file_name.endswith(('.txt', '.md')):
            return 'text'
        else:
            return 'unknown'
    except Exception as e:
        logger.error(f"magic file type detection error: {e}")
        return 'unknown'

# --- Improved prompt template for LLMs ---
def build_prompt(context, question):
    return f"""
Context:
{context}

Question:
{question}

Answer:
"""

# --- Centralized Output Formatting & Normalization ---
def gaia_normalize_answer(answer):
    """Normalize answer for GAIA: remove units, articles, extra text, and ensure concise, factual output."""
    if not isinstance(answer, str):
        answer = str(answer)
    # Remove common articles and units unless required
    answer = answer.strip()
    answer = re.sub(r"\b(the|a|an)\b", "", answer, flags=re.IGNORECASE)
    answer = re.sub(r"\s+", " ", answer)
    # Remove currency, percent, or units unless specified (GAIA rules)
    answer = re.sub(r"\$|%|USD|dollars|euros|eur|\bpercent\b", "", answer, flags=re.IGNORECASE)
    # Remove leading/trailing punctuation
    answer = answer.strip(' .,:;\n\t')
    return answer

# --- Reasoning Planner for Tool Chaining ---
def reasoning_planner(question, file_type, tools):
    """Plan the sequence of tools to use for a question using a Thought-Action-Observation cycle with ReAct prompting."""
    # Initialize plan with ReAct prompting for step-by-step reasoning
    initial_prompt = f"Let's think step by step to answer: {question}\nStep 1: Identify the type of question and any associated data.\nStep 2: Determine the tools or resources needed.\nStep 3: Outline the sequence of actions to solve the problem.\nProvide a detailed plan with up to 5 steps for solving this question."
    plan_response = llama3_chat(initial_prompt)
    logger.info(f"Initial plan for question: {question[:50]}... Plan: {plan_response[:200]}...")
    
    # Parse the plan into actionable steps (up to 5 for Level 1 GAIA tasks)
    steps = []
    for line in plan_response.split('\n'):
        if any(line.lower().startswith(f"step {i}") for i in range(1, 6)):
            steps.append(line.strip())
        if len(steps) >= 5:
            break
    
    # Default to heuristic if plan is unclear or empty
    if not steps:
        logger.warning(f"No clear plan generated for {question[:50]}... Falling back to heuristic.")
        if file_type == 'audio':
            return ['asr_transcribe', 'llama3_chat']
        elif file_type == 'image':
            return ['image_caption', 'llama3_chat']
        elif file_type == 'code':
            return ['code_analysis', 'llama3_chat']
        elif file_type in ['excel', 'csv']:
            return ['table_qa']
        elif 'youtube.com' in question or 'youtu.be' in question:
            return ['youtube_video_qa']
        elif any(w in question.lower() for w in ['wikipedia', 'who', 'when', 'where', 'what', 'how', 'find', 'search']):
            return ['web_search_duckduckgo', 'llama3_chat']
        elif 'chess' in question.lower() or 'move' in question.lower():
            return ['chess_move_analysis']
        elif any(w in question.lower() for w in ['fruit', 'vegetable', 'classify', 'category', 'botanical']):
            return ['botanical_classification']
        else:
            return ['llama3_chat']
    
    # Map plan steps to tools based on keywords and file type
    tool_sequence = []
    for step in steps:
        step_lower = step.lower()
        if file_type and not tool_sequence:
            if file_type == 'audio' and 'transcribe' in step_lower:
                tool_sequence.append('asr_transcribe')
            elif file_type == 'image' and 'caption' in step_lower:
                tool_sequence.append('image_caption')
            elif file_type == 'code' and 'run' in step_lower:
                tool_sequence.append('code_analysis')
            elif file_type in ['excel', 'csv'] and 'table' in step_lower:
                tool_sequence.append('table_qa')
        if 'youtube.com' in question or 'youtu.be' in question:
            tool_sequence.append('youtube_video_qa')
        elif any(w in step_lower for w in ['search', 'web', 'wikipedia', 'find', 'lookup']):
            tool_sequence.append('web_search_duckduckgo')
        elif any(w in step_lower for w in ['chess', 'move', 'board', 'position']):
            tool_sequence.append('chess_move_analysis')
        elif any(w in step_lower for w in ['fruit', 'vegetable', 'classify', 'category', 'botanical']):
            tool_sequence.append('botanical_classification')
        elif 'analyze' in step_lower or 'think' in step_lower or not tool_sequence:
            tool_sequence.append('llama3_chat')
    
    # Ensure at least one tool or LLM is used
    if not tool_sequence:
        tool_sequence.append('llama3_chat')
    
    logger.info(f"Tool sequence for {question[:50]}...: {tool_sequence}")
    return tool_sequence

# --- Improved RAG: Context Retrieval & Chunking ---
def retrieve_context(question, context_files, max_chunks=3):
    """Retrieve relevant context chunks from large files for RAG."""
    # Simple keyword search for now; can be replaced with semantic search
    relevant_chunks = []
    for file_path in context_files:
        try:
            with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
                text = f.read()
            # Split into chunks (e.g., 500 words)
            chunks = [text[i:i+2000] for i in range(0, len(text), 2000)]
            for chunk in chunks:
                if any(word.lower() in chunk.lower() for word in question.split()):
                    relevant_chunks.append(chunk)
                    if len(relevant_chunks) >= max_chunks:
                        break
        except Exception as e:
            logger.error(f"retrieve_context error: {e}")
    return '\n'.join(relevant_chunks)

# --- Modular Tool Registry & Chaining ---
class ToolRegistry:
    """Central registry for tools. Allows easy addition and chaining."""
    def __init__(self, tools):
        self.tools = tools
    def get(self, name):
        return self.tools.get(name)
    def add(self, name, func):
        self.tools[name] = func
    def list(self):
        return list(self.tools.keys())

# --- Refactored ModularGAIAAgent ---
class ModularGAIAAgent:
    """GAIA-compliant agent with robust reasoning, tool chaining, RAG, and output normalization."""
    def __init__(self, api_url=DEFAULT_API_URL, tool_registry=None, context_files=None):
        self.api_url = api_url
        self.tools = ToolRegistry(tool_registry or TOOL_REGISTRY)
        self.reasoning_trace = []
        self.file_cache = set(os.listdir('.'))
        self.context_files = context_files or []

    def fetch_questions(self, from_api=True, questions_path="Hugging Face Questions"):
        """Fetch questions from API or local file."""
        try:
            if from_api:
                r = requests.get(f"{self.api_url}/questions")
                r.raise_for_status()
                return r.json()
            else:
                with open(questions_path) as f:
                    data = f.read()
                start = data.find("[")
                end = data.rfind("]") + 1
                questions = json.loads(data[start:end])
                return questions
        except Exception as e:
            logger.error(f"fetch_questions error: {e}")
            return []

    def cached_download_file(self, file_id, file_name):
        """Download file from GAIA API with caching to avoid redundant downloads."""
        cache_file = "file_download_cache.json"
        cache = load_cache(cache_file)
        if file_id in cache:
            local_path = cache[file_id]
            if os.path.exists(local_path):
                logger.info(f"Using cached file for {file_id}: {local_path}")
                return local_path
        local_path = self.download_file(file_id, file_name)
        if local_path:
            cache[file_id] = local_path
            save_cache(cache_file, cache)
        return local_path

    def download_file(self, file_id, file_name):
        return self.cached_download_file(file_id, file_name)

    def detect_file_type(self, file_name):
        """Detect file type using magic and extension as fallback."""
        file_type = detect_file_type_magic(file_name)
        if file_type == 'unknown':
            ext = os.path.splitext(file_name)[-1].lower()
            if ext in ['.mp3', '.wav', '.flac']:
                return 'audio'
            elif ext in ['.png', '.jpg', '.jpeg', '.bmp']:
                return 'image'
            elif ext in ['.py']:
                return 'code'
            elif ext in ['.xlsx']:
                return 'excel'
            elif ext in ['.csv']:
                return 'csv'
            elif ext in ['.json']:
                return 'json'
            elif ext in ['.txt', '.md']:
                return 'text'
            else:
                return 'unknown'
        return file_type

    def analyze_file(self, file_name, file_type):
        """Analyze file and return context for the question."""
        try:
            if file_type == 'audio':
                transcript = self.tools.get('asr_transcribe')(file_name)
                self.reasoning_trace.append(f"Transcribed audio: {transcript[:100]}...")
                return transcript
            elif file_type == 'image':
                caption = self.tools.get('image_caption')(file_name)
                self.reasoning_trace.append(f"Image caption: {caption}")
                return caption
            elif file_type == 'code':
                result = self.tools.get('code_analysis')(file_name)
                self.reasoning_trace.append(f"Code analysis result: {result}")
                return result
            elif file_type == 'excel':
                wb = openpyxl.load_workbook(file_name)
                ws = wb.active
                data = list(ws.values)
                headers = data[0]
                table = [dict(zip(headers, row)) for row in data[1:]]
                self.reasoning_trace.append(f"Excel table loaded: {table[:2]}...")
                return table
            elif file_type == 'csv':
                df = pd.read_csv(file_name)
                table = df.to_dict(orient='records')
                self.reasoning_trace.append(f"CSV table loaded: {table[:2]}...")
                return table
            elif file_type == 'json':
                with open(file_name) as f:
                    data = json.load(f)
                self.reasoning_trace.append(f"JSON loaded: {str(data)[:100]}...")
                return data
            elif file_type == 'text':
                with open(file_name) as f:
                    text = f.read()
                self.reasoning_trace.append(f"Text loaded: {text[:100]}...")
                return text
            else:
                self.reasoning_trace.append(f"Unknown file type: {file_name}")
                logger.warning(f"Unknown file type: {file_name}")
                return None
        except Exception as e:
            logger.error(f"analyze_file error: {e}")
            self.reasoning_trace.append(f"Analyze file error: {e}")
            return None

    def answer_question(self, question_obj):
        self.reasoning_trace = []
        q = question_obj["question"]
        file_name = question_obj.get("file_name", "")
        file_content = None
        file_type = None
        if file_name:
            file_id = file_name.split('.')[0]
            local_file = self.download_file(file_id, file_name)
            if local_file:
                file_type = self.detect_file_type(local_file)
                file_content = self.analyze_file(local_file, file_type)
            else:
                self.reasoning_trace.append(f"Failed to download file {file_name}, proceeding without file content.")
                logger.warning(f"File download failed for {file_id}, proceeding without file content.")
        # RAG: retrieve context if needed
        rag_context = ''
        if self.context_files:
            try:
                rag_context = retrieve_context(q, self.context_files)
                self.reasoning_trace.append(f"Retrieved context: {rag_context[:100]}...")
            except Exception as e:
                logger.error(f"RAG context retrieval error: {e}")
                self.reasoning_trace.append(f"Context retrieval error: {e}, proceeding without context.")
        # Plan tools using enhanced reasoning planner
        try:
            tool_names = reasoning_planner(q, file_type if file_type else '', self.tools)
        except Exception as e:
            logger.error(f"Reasoning planner error: {e}")
            self.reasoning_trace.append(f"Planning error: {e}, falling back to default tool.")
            tool_names = ['llama3_chat']
        context = rag_context
        answer = ''
        max_retries = 2  # Retry mechanism for tool failures
        # Iterative Thought-Action-Observation cycle (up to 5 iterations for Level 1)
        for i, tool_name in enumerate(tool_names):
            tool = self.tools.get(tool_name)
            if not tool:
                self.reasoning_trace.append(f"Tool {tool_name} not found, skipping.")
                continue
            retries = 0
            while retries < max_retries:
                try:
                    logger.info(f"Step {i+1}/{len(tool_names)}: Using tool: {tool_name} | Question: {q[:50]}... | Context: {str(context)[:100]}... | Attempt {retries+1}/{max_retries}")
                    self.reasoning_trace.append(f"Step {i+1}: Using tool {tool_name} (Attempt {retries+1})")
                    if tool_name == 'web_search_duckduckgo':
                        context = tool(q)
                        self.reasoning_trace.append(f"Web search results: {context[:100]}...")
                    elif tool_name == 'table_qa' and file_content:
                        answer = tool(q, file_content)
                        self.reasoning_trace.append(f"Table QA result: {answer}")
                    elif tool_name in ['asr_transcribe', 'image_caption', 'code_analysis'] and file_name:
                        context = tool(file_name)
                        self.reasoning_trace.append(f"File analysis ({tool_name}): {context[:100]}...")
                    elif tool_name == 'youtube_video_qa':
                        answer = tool(q, q)
                        self.reasoning_trace.append(f"YouTube QA result: {answer}")
                    elif tool_name in ['chess_move_analysis'] and file_name:
                        answer = tool(file_name, q)
                        self.reasoning_trace.append(f"Chess move analysis result: {answer}")
                    elif tool_name in ['botanical_classification']:
                        answer = tool(q)
                        self.reasoning_trace.append(f"Botanical classification result: {answer}")
                    else:  # LLM like llama3_chat
                        if context:
                            prompt = build_prompt(context, q)
                            answer = tool(prompt)
                            self.reasoning_trace.append(f"LLM response with context: {answer[:100]}...")
                        else:
                            answer = tool(q)
                            self.reasoning_trace.append(f"LLM direct response: {answer[:100]}...")
                    # Observation: Check if answer seems complete or needs further steps
                    if answer and len(answer.split()) > 2:  # Basic check for meaningful answer
                        self.reasoning_trace.append(f"Answer seems meaningful after step {i+1}, stopping iteration.")
                        break
                    elif i < len(tool_names) - 1:
                        self.reasoning_trace.append(f"Answer incomplete after step {i+1}, proceeding to next tool.")
                    break  # Exit retry loop on success
                except Exception as e:
                    logger.error(f"Tool {tool_name} error on attempt {retries+1}: {e}")
                    self.reasoning_trace.append(f"Tool {tool_name} error on attempt {retries+1}: {e}")
                    retries += 1
                    if retries >= max_retries:
                        self.reasoning_trace.append(f"Max retries reached for {tool_name}, skipping to next tool or defaulting.")
                        if i == len(tool_names) - 1:  # Last tool failed
                            answer = "Unable to answer due to tool failures."
                        break
                    time.sleep(1)  # Brief delay before retry
        self.reasoning_trace.append(f"Tools used: {tool_names}")
        self.reasoning_trace.append(f"Final answer: {answer}")
        return gaia_normalize_answer(answer), self.reasoning_trace

    def answer_question_manual(self, question, file_upload, context_files):
        """Answer a manually input question with optional file and context."""
        try:
            # Handle file upload if provided
            file_name = None
            if file_upload:
                file_name = file_upload.name
                # Simulate GAIA file handling
                file_id = os.path.basename(file_name).split('.')[0]
                local_file = self.download_file(file_id, file_name)
                if local_file:
                    file_type = self.detect_file_type(local_file)
                    file_content = self.analyze_file(local_file, file_type)
                else:
                    file_content = None
            else:
                file_content = None
            # Handle context files if provided
            self.context_files = [f.name for f in context_files] if context_files else []
            # Create a mock question object
            question_obj = {
                "question": question,
                "file_name": file_name if file_name else ""
            }
            answer, trace = self.answer_question(question_obj)
            return answer, "\n".join(trace)
        except Exception as e:
            logger.error(f"Manual question error: {e}")
            return f"Error: {e}", f"Error occurred: {e}"

    def process_batch(self, token):
        """Process a batch of questions with progress updates."""
        try:
            questions = self.fetch_questions(token)
            if not questions:
                return "0/0 questions processed - fetch failed", []
            total = len(questions)
            results = []
            for i, q in enumerate(questions):
                try:
                    answer, trace = self.answer_question(q)
                    results.append({
                        "task_id": q["task_id"],
                        "question": q["question"],
                        "answer": answer,
                        "trace": trace
                    })
                    logger.info(f"Batch progress: {i+1}/{total} questions processed")
                    yield f"{i+1}/{total} questions processed", results
                except Exception as e:
                    logger.error(f"Batch processing error for question {i+1}: {e}")
                    results.append({
                        "task_id": q.get("task_id", "unknown"),
                        "question": q.get("question", "unknown"),
                        "answer": "Error processing",
                        "trace": [str(e)]
                    })
                    yield f"{i+1}/{total} questions processed", results
            logger.info(f"Batch processing complete: {total}/{total} questions processed")
        except Exception as e:
            logger.error(f"Batch processing overall error: {e}")
            yield "Error in batch processing", []

# --- Build Gradio Interface using Blocks (Maintaining Original Architecture) ---
with gr.Blocks() as demo:
    gr.Markdown("# Smart Agent Evaluation Runner")
    gr.Markdown("""
        **Instructions:**
        1. Clone this space, define your agent logic, tools, packages, etc.
        2. Log in to Hugging Face.
        3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
    """)

    gr.LoginButton()
    run_button = gr.Button("Run Evaluation & Submit All Answers")
    status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
    results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)

    run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table])

if __name__ == "__main__":
    print("Launching Gradio Interface for Smart Agent Evaluation...")
    demo.launch(debug=True, share=False)

# Define a wrapper to ensure compatibility
def run_and_submit_all_wrapper(profile: gr.OAuthProfile | None):
    return run_and_submit_all(profile)

# Update run_and_submit_all to use the enhanced ModularGAIAAgent
def run_and_submit_all(profile: gr.OAuthProfile | None):
    space_id = os.getenv("SPACE_ID")
    if profile:
        username = profile.username
        print(f"User logged in: {username}")
    else:
        return "Please Login to Hugging Face with the button.", None

    api_url = DEFAULT_API_URL
    questions_url = f"{api_url}/questions"
    submit_url = f"{api_url}/submit"

    agent = ModularGAIAAgent(api_url=DEFAULT_API_URL)
    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"

    try:
        response = requests.get(questions_url, timeout=15)
        response.raise_for_status()
        questions_data = response.json()
    except Exception as e:
        return f"Error fetching questions: {e}", None

    results_log = []
    answers_payload = []
    correct_answers = 0

    for item in questions_data:
        task_id = item.get("task_id")
        question_text = item.get("question")
        if not task_id or not question_text:
            continue

        submitted_answer, trace = agent.answer_question(item)
        answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
        results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})

    if not answers_payload:
        return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)

    submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
    print(f"Submitting {len(answers_payload)} answers to: {submit_url}")

    try:
        response = requests.post(submit_url, json=submission_data, timeout=60)
        response.raise_for_status()
        result_data = response.json()
        final_status = (
            f"Submission Successful!\n"
            f"User: {result_data.get('username')}\n"
            f"Overall Score: {result_data.get('score', 'N/A')}% "
            f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
            f"Message: {result_data.get('message', 'No message received.')}"
        )
        results_df = pd.DataFrame(results_log)
        return final_status, results_df
    except Exception as e:
        return f"Submission Failed: {e}", pd.DataFrame(results_log)