Spaces:
Running
Running
File size: 32,594 Bytes
1d8e193 |
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 |
import json
import gzip
import gradio as gr
from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
import pandas as pd
from apscheduler.schedulers.background import BackgroundScheduler
from huggingface_hub import snapshot_download
from io import StringIO
from dataclasses import dataclass, field
from src.about import (
CITATION_BUTTON_LABEL,
CITATION_BUTTON_TEXT,
EVALUATION_QUEUE_TEXT,
INTRODUCTION_TEXT,
LLM_BENCHMARKS_TEXT,
TITLE,
)
from src.display.css_html_js import custom_css
from src.display.utils import (
BENCHMARK_COLS,
BENCHMARK_COLS_MULTIMODAL,
BENCHMARK_COLS_MIB_SUBGRAPH,
BENCHMARK_COLS_MIB_CAUSALGRAPH,
COLS,
COLS_MIB_SUBGRAPH,
COLS_MIB_CAUSALGRAPH,
COLS_MULTIMODAL,
EVAL_COLS,
EVAL_TYPES,
AutoEvalColumn,
AutoEvalColumn_mib_subgraph,
AutoEvalColumn_mib_causalgraph,
fields,
)
from src.envs import API, EVAL_REQUESTS_PATH, QUEUE_REPO, REPO_ID, TOKEN, RESULTS_REPO_MIB_SUBGRAPH, EVAL_RESULTS_MIB_SUBGRAPH_PATH, RESULTS_REPO_MIB_CAUSALGRAPH, EVAL_RESULTS_MIB_CAUSALGRAPH_PATH
from src.populate import get_evaluation_queue_df, get_leaderboard_df, get_leaderboard_df_mib_subgraph, get_leaderboard_df_mib_causalgraph
from src.submission.submit import add_new_eval
from src.about import TasksMib_Subgraph
# class SmartSelectColumns(SelectColumns):
# """
# Enhanced SelectColumns component with basic filtering functionality.
# """
# def __init__(
# self,
# benchmark_keywords: Optional[List[str]] = None,
# model_keywords: Optional[List[str]] = None,
# initial_selected: Optional[List[str]] = None,
# **kwargs
# ):
# """
# Initialize SmartSelectColumns with minimal configuration.
# Args:
# benchmark_keywords: List of benchmark names to filter by
# model_keywords: List of model names to filter by
# initial_selected: List of columns to show initially
# """
# super().__init__(**kwargs)
# self.benchmark_keywords = benchmark_keywords or []
# self.model_keywords = model_keywords or []
# self.initial_selected = initial_selected or []
# def get_filtered_groups(self, df: pd.DataFrame) -> Dict[str, List[str]]:
# """
# Create column groups based on simple substring matching.
# """
# filtered_groups = {}
# # Create benchmark groups
# for benchmark in self.benchmark_keywords:
# matching_cols = [
# col for col in df.columns
# if benchmark in col.lower()
# ]
# if matching_cols:
# group_name = f"Benchmark group for {benchmark}"
# filtered_groups[group_name] = matching_cols
# # Create model groups
# for model in self.model_keywords:
# matching_cols = [
# col for col in df.columns
# if model in col.lower()
# ]
# if matching_cols:
# group_name = f"Model group for {model}"
# filtered_groups[group_name] = matching_cols
# return filtered_groups
# def update(
# self,
# value: Union[pd.DataFrame, Dict[str, List[str]], Any]
# ) -> Dict:
# """Update component with new values."""
# if isinstance(value, pd.DataFrame):
# choices = list(value.columns)
# selected = self.initial_selected if self.initial_selected else choices
# filtered_cols = self.get_filtered_groups(value)
# return {
# "choices": choices,
# "value": selected,
# "filtered_cols": filtered_cols
# }
# if hasattr(value, '__dataclass_fields__'):
# field_names = [field.name for field in fields(value)]
# return {
# "choices": field_names,
# "value": self.initial_selected if self.initial_selected else field_names
# }
# return super().update(value)
from gradio.events import Dependency
class ModifiedLeaderboard(Leaderboard):
"""Extends Leaderboard to support substring-based column filtering"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
# Process substring groups if they exist
if (isinstance(self.select_columns_config, SelectColumns) and
self.select_columns_config.substring_groups):
self.process_substring_groups()
def process_substring_groups(self):
"""Processes substring groups to add them to the selectable columns"""
groups = self.select_columns_config.substring_groups
if not groups:
return
# Create a mapping of group name to matching columns
group_to_columns = {}
for group_name, patterns in groups.groups.items():
matching_cols = set()
for pattern in patterns:
regex = re.compile(pattern.replace('*', '.*'))
matching_cols.update(
col for col in self.headers
if regex.search(col)
)
if matching_cols:
group_to_columns[group_name] = list(matching_cols)
# Add groups to the headers and update column selection logic
self.group_to_columns = group_to_columns
self.original_headers = self.headers.copy()
# Add group names to the start of headers
self.headers = list(group_to_columns.keys()) + self.original_headers
# Update default selection to include groups
if self.select_columns_config.default_selection:
self.select_columns_config.default_selection = (
list(group_to_columns.keys()) +
self.select_columns_config.default_selection
)
def preprocess(self, payload):
"""Override preprocess to handle group selection"""
df = super().preprocess(payload)
# If we don't have substring groups, return normally
if not hasattr(self, 'group_to_columns'):
return df
# Process group selections
selected_columns = set()
for column in payload.headers:
if column in self.group_to_columns:
# If a group is selected, add all its columns
selected_columns.update(self.group_to_columns[column])
elif column in self.original_headers:
# Add individually selected columns
selected_columns.add(column)
# Return DataFrame with only selected columns
return df[list(selected_columns)]
from typing import Callable, Literal, Sequence, Any, TYPE_CHECKING
from gradio.blocks import Block
if TYPE_CHECKING:
from gradio.components import Timer
from gradio_leaderboard import SelectColumns, Leaderboard
import pandas as pd
from typing import List, Dict, Optional
from dataclasses import fields
class SmartSelectColumns(SelectColumns):
"""
Enhanced SelectColumns component matching exact original parameters.
"""
def __init__(
self,
benchmark_keywords: Optional[List[str]] = None,
model_keywords: Optional[List[str]] = None,
initial_selected: Optional[List[str]] = None,
label: Optional[str] = None,
show_label: bool = True,
info: Optional[str] = None,
allow: bool = True
):
# Match exact parameters from working SelectColumns
super().__init__(
default_selection=initial_selected or [],
cant_deselect=[],
allow=allow,
label=label,
show_label=show_label,
info=info
)
self.benchmark_keywords = benchmark_keywords or []
self.model_keywords = model_keywords or []
# Store groups for later use
self._groups = {}
def get_filtered_groups(self, columns: List[str]) -> Dict[str, List[str]]:
"""Get column groups based on keywords."""
filtered_groups = {}
# Add benchmark groups
for benchmark in self.benchmark_keywords:
matching_cols = [
col for col in columns
if benchmark in col.lower()
]
if matching_cols:
filtered_groups[f"Benchmark group for {benchmark}"] = matching_cols
# Add model groups
for model in self.model_keywords:
matching_cols = [
col for col in columns
if model in col.lower()
]
if matching_cols:
filtered_groups[f"Model group for {model}"] = matching_cols
self._groups = filtered_groups
return filtered_groups
import re
@dataclass
class SubstringSelectColumns(SelectColumns):
"""
Extends SelectColumns to support filtering columns by predefined substrings.
When a substring is selected, all columns containing that substring will be selected.
"""
substring_groups: Dict[str, List[str]] = field(default_factory=dict)
selected_substrings: List[str] = field(default_factory=list)
def __post_init__(self):
# Ensure default_selection is a list
if self.default_selection is None:
self.default_selection = []
# Build reverse mapping of column to substrings
self.column_to_substrings = {}
for substring, patterns in self.substring_groups.items():
for pattern in patterns:
# Convert glob-style patterns to regex
regex = re.compile(pattern.replace('*', '.*'))
# Find matching columns in default_selection
for col in self.default_selection:
if regex.search(col):
if col not in self.column_to_substrings:
self.column_to_substrings[col] = []
self.column_to_substrings[col].append(substring)
# Apply initial substring selections
if self.selected_substrings:
self.update_selection_from_substrings()
def update_selection_from_substrings(self) -> List[str]:
"""
Updates the column selection based on selected substrings.
Returns the new list of selected columns.
"""
selected_columns = self.cant_deselect.copy()
# If no substrings selected, show all columns
if not self.selected_substrings:
selected_columns.extend([
col for col in self.default_selection
if col not in self.cant_deselect
])
return selected_columns
# Add columns that match any selected substring
for col, substrings in self.column_to_substrings.items():
if any(s in self.selected_substrings for s in substrings):
if col not in selected_columns:
selected_columns.append(col)
return selected_columns
def restart_space():
API.restart_space(repo_id=REPO_ID)
### Space initialisation
try:
# print(EVAL_REQUESTS_PATH)
snapshot_download(
repo_id=QUEUE_REPO, local_dir=EVAL_REQUESTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
)
except Exception:
restart_space()
try:
# print(RESULTS_REPO_MIB_SUBGRAPH)
snapshot_download(
repo_id=RESULTS_REPO_MIB_SUBGRAPH, local_dir=EVAL_RESULTS_MIB_SUBGRAPH_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
)
except Exception:
restart_space()
try:
# print(RESULTS_REPO_MIB_CAUSALGRAPH)
snapshot_download(
repo_id=RESULTS_REPO_MIB_CAUSALGRAPH, local_dir=EVAL_RESULTS_MIB_CAUSALGRAPH_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
)
except Exception:
restart_space()
LEADERBOARD_DF_MIB_SUBGRAPH = get_leaderboard_df_mib_subgraph(EVAL_RESULTS_MIB_SUBGRAPH_PATH, EVAL_REQUESTS_PATH, COLS_MIB_SUBGRAPH, BENCHMARK_COLS_MIB_SUBGRAPH)
# LEADERBOARD_DF_MIB_CAUSALGRAPH = get_leaderboard_df_mib_causalgraph(EVAL_RESULTS_MIB_CAUSALGRAPH_PATH, EVAL_REQUESTS_PATH, COLS_MIB_CAUSALGRAPH, BENCHMARK_COLS_MIB_CAUSALGRAPH)
# In app.py, modify the LEADERBOARD initialization
LEADERBOARD_DF_MIB_CAUSALGRAPH_DETAILED, LEADERBOARD_DF_MIB_CAUSALGRAPH_AGGREGATED, LEADERBOARD_DF_MIB_CAUSALGRAPH_AVERAGED = get_leaderboard_df_mib_causalgraph(
EVAL_RESULTS_MIB_CAUSALGRAPH_PATH,
EVAL_REQUESTS_PATH,
COLS_MIB_CAUSALGRAPH,
BENCHMARK_COLS_MIB_CAUSALGRAPH
)
# LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS)
# LEADERBOARD_DF_MULTIMODAL = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS_MULTIMODAL, BENCHMARK_COLS_MULTIMODAL)
(
finished_eval_queue_df,
running_eval_queue_df,
pending_eval_queue_df,
) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
# def init_leaderboard_mib_subgraph(dataframe, track):
# # print(f"init_leaderboard_mib: dataframe head before loc is {dataframe.head()}\n")
# if dataframe is None or dataframe.empty:
# raise ValueError("Leaderboard DataFrame is empty or None.")
# # filter for correct track
# # dataframe = dataframe.loc[dataframe["Track"] == track]
# # print(f"init_leaderboard_mib: dataframe head after loc is {dataframe.head()}\n")
# return Leaderboard(
# value=dataframe,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
# select_columns=SelectColumns(
# default_selection=[c.name for c in fields(AutoEvalColumn_mib_subgraph) if c.displayed_by_default],
# cant_deselect=[c.name for c in fields(AutoEvalColumn_mib_subgraph) if c.never_hidden],
# label="Select Columns to Display:",
# ),
# search_columns=["Method"], # Changed from AutoEvalColumn_mib_subgraph.model.name to "Method"
# hide_columns=[c.name for c in fields(AutoEvalColumn_mib_subgraph) if c.hidden],
# bool_checkboxgroup_label="Hide models",
# interactive=False,
# )
# def init_leaderboard_mib_subgraph(dataframe, track):
# """Initialize the subgraph leaderboard with grouped column selection by benchmark."""
# if dataframe is None or dataframe.empty:
# raise ValueError("Leaderboard DataFrame is empty or None.")
# print("\nDebugging DataFrame columns:", dataframe.columns.tolist())
# # Create groups of columns by benchmark
# benchmark_groups = []
# # For each benchmark in our TasksMib_Subgraph enum...
# for task in TasksMib_Subgraph:
# benchmark = task.value.benchmark
# # Get all valid columns for this benchmark's models
# benchmark_cols = [
# f"{benchmark}_{model}"
# for model in task.value.models
# if f"{benchmark}_{model}" in dataframe.columns
# ]
# if benchmark_cols: # Only add if we have valid columns
# benchmark_groups.append(benchmark_cols)
# print(f"\nBenchmark group for {benchmark}:", benchmark_cols)
# # Create model groups as well
# model_groups = []
# all_models = list(set(model for task in TasksMib_Subgraph for model in task.value.models))
# # For each unique model...
# for model in all_models:
# # Get all valid columns for this model across benchmarks
# model_cols = [
# f"{task.value.benchmark}_{model}"
# for task in TasksMib_Subgraph
# if model in task.value.models
# and f"{task.value.benchmark}_{model}" in dataframe.columns
# ]
# if model_cols: # Only add if we have valid columns
# model_groups.append(model_cols)
# print(f"\nModel group for {model}:", model_cols)
# # Combine all groups
# all_groups = benchmark_groups + model_groups
# # Flatten groups for default selection (show everything initially)
# all_columns = [col for group in all_groups for col in group]
# print("\nAll available columns:", all_columns)
# return Leaderboard(
# value=dataframe,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
# select_columns=SelectColumns(
# default_selection=all_columns, # Show all columns initially
# label="Select Results:"
# ),
# search_columns=["Method"],
# hide_columns=[],
# interactive=False,
# )
def init_leaderboard_mib_subgraph(dataframe, track):
"""Initialize the subgraph leaderboard with display names for better readability."""
if dataframe is None or dataframe.empty:
raise ValueError("Leaderboard DataFrame is empty or None.")
print("\nDebugging DataFrame columns:", dataframe.columns.tolist())
# First, create our display name mapping
# This is like creating a translation dictionary between internal names and display names
model_name_mapping = {
"qwen2_5": "Qwen-2.5",
"gpt2": "GPT-2",
"gemma2": "Gemma-2",
"llama3": "Llama-3.1"
}
benchmark_mapping = {
"ioi": "IOI",
"mcqa": "MCQA",
"arithmetic_addition": "Arithmetic (+)",
"arithmetic_subtraction": "Arithmetic (-)",
"arc_easy": "ARC (Easy)",
"arc_challenge": "ARC (Challenge)"
}
display_mapping = {}
for task in TasksMib_Subgraph:
for model in task.value.models:
field_name = f"{task.value.benchmark}_{model}"
display_name = f"{benchmark_mapping[task.value.benchmark]} - {model_name_mapping[model]}"
display_mapping[field_name] = display_name
# Now when creating benchmark groups, we'll use display names
benchmark_groups = []
for task in TasksMib_Subgraph:
benchmark = task.value.benchmark
benchmark_cols = [
display_mapping[f"{benchmark}_{model}"] # Use display name from our mapping
for model in task.value.models
if f"{benchmark}_{model}" in dataframe.columns
]
if benchmark_cols:
benchmark_groups.append(benchmark_cols)
print(f"\nBenchmark group for {benchmark}:", benchmark_cols)
# Similarly for model groups
model_groups = []
all_models = list(set(model for task in TasksMib_Subgraph for model in task.value.models))
for model in all_models:
model_cols = [
display_mapping[f"{task.value.benchmark}_{model}"] # Use display name
for task in TasksMib_Subgraph
if model in task.value.models
and f"{task.value.benchmark}_{model}" in dataframe.columns
]
if model_cols:
model_groups.append(model_cols)
print(f"\nModel group for {model}:", model_cols)
# Combine all groups using display names
all_groups = benchmark_groups + model_groups
all_columns = [col for group in all_groups for col in group]
# Important: We need to rename our DataFrame columns to match display names
renamed_df = dataframe.rename(columns=display_mapping)
# all_columns = [c.name for c in fields(AutoEvalColumn_mib_subgraph) if c.displayed_by_default]
# all_columns = [c.name for c in fields(AutoEvalColumn_mib_subgraph)]
all_columns = renamed_df.columns.tolist()
print(benchmark_groups)
print(model_groups)
filter_groups = {"ioi": "*IOI*",
"llama": "*Llama*"}
# Original code
return ModifiedLeaderboard(
value=renamed_df, # Use DataFrame with display names
datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
select_columns=SubstringSelectColumns(
substring_groups=filter_groups,
default_selection=all_columns, # Now contains display names
label="Filter Results:",
allow=True
),
search_columns=["Method"],
hide_columns=[],
interactive=False,
)
# # Complete column groups for both benchmarks and models
# # Define keywords for filtering
# benchmark_keywords = ["ioi", "mcqa", "arithmetic_addition", "arithmetic_subtraction", "arc_easy", "arc_challenge"]
# model_keywords = ["qwen2_5", "gpt2", "gemma2", "llama3"]
# # Optional: Define display names
# mappings = {
# "ioi_llama3": "IOI (LLaMA-3)",
# "ioi_qwen2_5": "IOI (Qwen-2.5)",
# "ioi_gpt2": "IOI (GPT-2)",
# "ioi_gemma2": "IOI (Gemma-2)",
# "mcqa_llama3": "MCQA (LLaMA-3)",
# "mcqa_qwen2_5": "MCQA (Qwen-2.5)",
# "mcqa_gemma2": "MCQA (Gemma-2)",
# "arithmetic_addition_llama3": "Arithmetic Addition (LLaMA-3)",
# "arithmetic_subtraction_llama3": "Arithmetic Subtraction (LLaMA-3)",
# "arc_easy_llama3": "ARC Easy (LLaMA-3)",
# "arc_easy_gemma2": "ARC Easy (Gemma-2)",
# "arc_challenge_llama3": "ARC Challenge (LLaMA-3)",
# "eval_name": "Evaluation Name",
# "Method": "Method",
# "Average": "Average Score"
# }
# # mappings = {}
# # Create SmartSelectColumns instance
# smart_columns = SmartSelectColumns(
# benchmark_keywords=benchmark_keywords,
# model_keywords=model_keywords,
# column_mapping=mappings,
# initial_selected=["Method", "Average"]
# )
# print("\nDebugging DataFrame columns:", renamed_df.columns.tolist())
# # Create Leaderboard
# leaderboard = Leaderboard(
# value=renamed_df,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
# select_columns=smart_columns,
# search_columns=["Method"],
# hide_columns=[],
# interactive=False
# )
# print(f"Successfully created leaderboard.")
# return leaderboard
# print("\nDebugging DataFrame columns:", dataframe.columns.tolist())
# # Define simple keywords for filtering
# benchmark_keywords = ["ioi", "mcqa", "arithmetic", "arc"]
# model_keywords = ["qwen2_5", "gpt2", "gemma2", "llama3"]
# # Create SmartSelectColumns instance with exact same parameters as working version
# smart_columns = SmartSelectColumns(
# benchmark_keywords=benchmark_keywords,
# model_keywords=model_keywords,
# initial_selected=["Method", "Average"],
# allow=True,
# label=None,
# show_label=True,
# info=None
# )
# try:
# print("\nCreating leaderboard...")
# # Get groups before creating leaderboard
# smart_columns.get_filtered_groups(dataframe.columns)
# leaderboard = Leaderboard(
# value=dataframe,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
# select_columns=smart_columns,
# search_columns=["Method"],
# hide_columns=[],
# interactive=False
# )
# print("Leaderboard created successfully")
# return leaderboard
# except Exception as e:
# print("Error creating leaderboard:", str(e))
# raise
# def init_leaderboard_mib_subgraph(dataframe, track):
# """Initialize the subgraph leaderboard with group-based column selection."""
# if dataframe is None or dataframe.empty:
# raise ValueError("Leaderboard DataFrame is empty or None.")
# print("\nDebugging DataFrame columns:", dataframe.columns.tolist())
# # Create selection mapping for benchmark groups
# selection_mapping = {}
# # Create benchmark groups with descriptive names
# for task in TasksMib_Subgraph:
# benchmark = task.value.benchmark
# # Get all columns for this benchmark's models
# benchmark_cols = [
# f"{benchmark}_{model}"
# for model in task.value.models
# if f"{benchmark}_{model}" in dataframe.columns
# ]
# if benchmark_cols:
# # Use a descriptive group name as the key
# group_name = f"Benchmark: {benchmark.upper()}"
# selection_mapping[group_name] = benchmark_cols
# print(f"\n{group_name} maps to:", benchmark_cols)
# # Create model groups with descriptive names
# all_models = list(set(model for task in TasksMib_Subgraph for model in task.value.models))
# for model in all_models:
# # Get all columns for this model across benchmarks
# model_cols = [
# f"{task.value.benchmark}_{model}"
# for task in TasksMib_Subgraph
# if model in task.value.models
# and f"{task.value.benchmark}_{model}" in dataframe.columns
# ]
# if model_cols:
# # Use a descriptive group name as the key
# group_name = f"Model: {model}"
# selection_mapping[group_name] = model_cols
# print(f"\n{group_name} maps to:", model_cols)
# # The selection options are the group names
# selection_options = list(selection_mapping.keys())
# print("\nSelection options:", selection_options)
# return Leaderboard(
# value=dataframe,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_subgraph)],
# select_columns=SelectColumns(
# default_selection=selection_options, # Show all groups by default
# label="Select Benchmark or Model Groups:"
# ),
# search_columns=["Method"],
# hide_columns=[],
# interactive=False,
# )
# def init_leaderboard_mib_causalgraph(dataframe, track):
# # print("Debugging column issues:")
# # print("\nActual DataFrame columns:")
# # print(dataframe.columns.tolist())
# # print("\nExpected columns for Leaderboard:")
# expected_cols = [c.name for c in fields(AutoEvalColumn_mib_causalgraph)]
# # print(expected_cols)
# # print("\nMissing columns:")
# missing_cols = [col for col in expected_cols if col not in dataframe.columns]
# # print(missing_cols)
# # print("\nSample of DataFrame content:")
# # print(dataframe.head().to_string())
# return Leaderboard(
# value=dataframe,
# datatype=[c.type for c in fields(AutoEvalColumn_mib_causalgraph)],
# select_columns=SelectColumns(
# default_selection=[c.name for c in fields(AutoEvalColumn_mib_causalgraph) if c.displayed_by_default],
# cant_deselect=[c.name for c in fields(AutoEvalColumn_mib_causalgraph) if c.never_hidden],
# label="Select Columns to Display:",
# ),
# search_columns=["Method"],
# hide_columns=[c.name for c in fields(AutoEvalColumn_mib_causalgraph) if c.hidden],
# bool_checkboxgroup_label="Hide models",
# interactive=False,
# )
def init_leaderboard_mib_causalgraph(dataframe, track):
# print("Debugging column issues:")
# print("\nActual DataFrame columns:")
# print(dataframe.columns.tolist())
# Create only necessary columns
return Leaderboard(
value=dataframe,
datatype=[c.type for c in fields(AutoEvalColumn_mib_causalgraph)],
select_columns=SelectColumns(
default_selection=["Method"], # Start with just Method column
cant_deselect=["Method"], # Method column should always be visible
label="Select Columns to Display:",
),
search_columns=["Method"],
hide_columns=[],
bool_checkboxgroup_label="Hide models",
interactive=False,
)
def init_leaderboard(dataframe, track):
if dataframe is None or dataframe.empty:
raise ValueError("Leaderboard DataFrame is empty or None.")
# filter for correct track
dataframe = dataframe.loc[dataframe["Track"] == track]
# print(f"\n\n\n dataframe is {dataframe}\n\n\n")
return Leaderboard(
value=dataframe,
datatype=[c.type for c in fields(AutoEvalColumn)],
select_columns=SelectColumns(
default_selection=[c.name for c in fields(AutoEvalColumn) if c.displayed_by_default],
cant_deselect=[c.name for c in fields(AutoEvalColumn) if c.never_hidden],
label="Select Columns to Display:",
),
search_columns=[AutoEvalColumn.model.name],
hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden],
bool_checkboxgroup_label="Hide models",
interactive=False,
)
def process_json(temp_file):
if temp_file is None:
return {}
# Handle file upload
try:
file_path = temp_file.name
if file_path.endswith('.gz'):
with gzip.open(file_path, 'rt') as f:
data = json.load(f)
else:
with open(file_path, 'r') as f:
data = json.load(f)
except Exception as e:
raise gr.Error(f"Error processing file: {str(e)}")
gr.Markdown("Upload successful!")
return data
demo = gr.Blocks(css=custom_css)
with demo:
gr.HTML(TITLE)
gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text")
with gr.Tabs(elem_classes="tab-buttons") as tabs:
# with gr.TabItem("Strict", elem_id="strict-benchmark-tab-table", id=0):
# leaderboard = init_leaderboard(LEADERBOARD_DF, "strict")
# with gr.TabItem("Strict-small", elem_id="strict-small-benchmark-tab-table", id=1):
# leaderboard = init_leaderboard(LEADERBOARD_DF, "strict-small")
# with gr.TabItem("Multimodal", elem_id="multimodal-benchmark-tab-table", id=2):
# leaderboard = init_leaderboard(LEADERBOARD_DF_MULTIMODAL, "multimodal")
# with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=4):
# gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
# with gr.TabItem("👶 Submit", elem_id="llm-benchmark-tab-table", id=5):
# with gr.Column():
# with gr.Row():
# gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text")
# with gr.TabItem("Subgraph", elem_id="subgraph", id=0):
# leaderboard = init_leaderboard_mib_subgraph(LEADERBOARD_DF_MIB_SUBGRAPH, "Subgraph")
with gr.TabItem("Subgraph", elem_id="subgraph", id=0):
# Add description for filters
gr.Markdown("""
### Filtering Options
Use the dropdown menus below to filter results by specific tasks or models.
You can combine filters to see specific task-model combinations.
""")
leaderboard = init_leaderboard_mib_subgraph(LEADERBOARD_DF_MIB_SUBGRAPH, "Subgraph")
print(f"Leaderboard is {leaderboard}")
# Then modify the Causal Graph tab section
with gr.TabItem("Causal Graph", elem_id="causalgraph", id=1):
with gr.Tabs() as causalgraph_tabs:
with gr.TabItem("Detailed View", id=0):
leaderboard_detailed = init_leaderboard_mib_causalgraph(
LEADERBOARD_DF_MIB_CAUSALGRAPH_DETAILED,
"Causal Graph"
)
with gr.TabItem("Aggregated View", id=1):
leaderboard_aggregated = init_leaderboard_mib_causalgraph(
LEADERBOARD_DF_MIB_CAUSALGRAPH_AGGREGATED,
"Causal Graph"
)
with gr.TabItem("Intervention Averaged", id=2):
leaderboard_averaged = init_leaderboard_mib_causalgraph(
LEADERBOARD_DF_MIB_CAUSALGRAPH_AVERAGED,
"Causal Graph"
)
# with gr.Row():
# with gr.Accordion("📙 Citation", open=False):
# citation_button = gr.Textbox(
# value=CITATION_BUTTON_TEXT,
# label=CITATION_BUTTON_LABEL,
# lines=20,
# elem_id="citation-button",
# show_copy_button=True,
# )
scheduler = BackgroundScheduler()
scheduler.add_job(restart_space, "interval", seconds=1800)
scheduler.start()
demo.launch(share=True, ssr_mode=False) |