Spaces:
Sleeping
Sleeping
File size: 17,000 Bytes
70ea7be | 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 | """Dataset validator for integrity checks.
Validates dataset integrity against metadata expectations: table presence,
row counts, null values, foreign key relationships, split columns, and
target label columns. Uses the collect-all-errors pattern β reports every
issue in one pass rather than failing fast.
"""
import logging
from dataclasses import dataclass, field
from datetime import datetime, timezone
from pathlib import Path
import pandas as pd
from app.core.exceptions import DatasetError
from app.data.loader import DatasetLoader
logger = logging.getLogger(__name__)
# Tables that must contain a train_split column with exactly {train, validation, test}
TRAINING_TABLES: list[str] = [
"training_lo_tagging",
"training_bloom_classification",
"training_risk_prediction",
"training_mastery_prediction",
"training_answer_scoring",
"training_recommendation_outcomes",
"learning_outcomes",
"questions",
"student_profiles",
"student_attempts",
"mastery_profiles",
"engagement_logs",
"risk_profiles",
"recommendations",
"content_catalog",
]
# Foreign key relationships: (child_table, child_column, parent_table, parent_column)
FOREIGN_KEY_RELATIONSHIPS: list[tuple[str, str, str, str]] = [
("student_attempts", "student_id", "student_profiles", "student_id"),
("student_attempts", "question_id", "questions", "question_id"),
("questions", "lo_id", "learning_outcomes", "lo_id"),
("lo_dependencies", "lo_id", "learning_outcomes", "lo_id"),
("lo_dependencies", "prerequisite_lo_id", "learning_outcomes", "lo_id"),
]
# Target columns that must exist in their respective tables
TARGET_COLUMNS: list[tuple[str, str]] = [
("training_lo_tagging", "lo_id"),
("training_bloom_classification", "bloom_level"),
("training_mastery_prediction", "mastery_label"),
("training_risk_prediction", "risk_label"),
("training_risk_prediction", "risk_level"),
("training_answer_scoring", "teacher_marks"),
("training_recommendation_outcomes", "clicked"),
("training_recommendation_outcomes", "is_completed"),
]
VALID_SPLITS: set[str] = {"train", "validation", "test"}
@dataclass
class ValidationIssue:
"""A single validation issue found during dataset checks."""
check: str
table: str
column: str | None
message: str
severity: str # "error" | "warning"
@dataclass
class ValidationReport:
"""Aggregated result of all validation checks."""
timestamp: datetime
passed: bool
issues: list[ValidationIssue] = field(default_factory=list)
checks_run: int = 0
checks_passed: int = 0
class DatasetValidator:
"""Validates dataset integrity against metadata expectations.
Uses the collect-all-errors pattern: every check runs to completion
and all issues are aggregated into a single report.
"""
def __init__(self, loader: DatasetLoader, metadata: dict) -> None:
self._loader = loader
self._metadata = metadata
def check_table_presence(self) -> list[ValidationIssue]:
"""Verify all expected CSV files from metadata are present.
Returns:
List of ValidationIssue for any missing tables.
"""
issues: list[ValidationIssue] = []
table_counts = self._metadata.get("table_counts", {})
for table_file in table_counts:
table_name = table_file.replace(".csv", "")
path = self._loader.get_table_path(table_name)
if not path.exists():
issues.append(
ValidationIssue(
check="table_presence",
table=table_name,
column=None,
message=f"Expected CSV file not found: {path}",
severity="error",
)
)
return issues
def check_row_counts(self) -> list[ValidationIssue]:
"""Compare actual row counts against metadata expected counts.
Returns:
List of ValidationIssue for any row count mismatches.
"""
issues: list[ValidationIssue] = []
table_counts = self._metadata.get("table_counts", {})
for table_file, expected_count in table_counts.items():
table_name = table_file.replace(".csv", "")
try:
df = self._loader.load_table(table_name)
except DatasetError:
# Table missing β already reported by check_table_presence
continue
actual_count = len(df)
if actual_count != expected_count:
issues.append(
ValidationIssue(
check="row_count",
table=table_name,
column=None,
message=(
f"Row count mismatch: expected {expected_count}, "
f"got {actual_count}"
),
severity="error",
)
)
return issues
def check_null_values(self) -> list[ValidationIssue]:
"""Confirm no CSV file contains any null values.
Returns:
List of ValidationIssue for any columns with null values.
"""
issues: list[ValidationIssue] = []
table_counts = self._metadata.get("table_counts", {})
for table_file in table_counts:
table_name = table_file.replace(".csv", "")
try:
df = self._loader.load_table(table_name)
except DatasetError:
continue
null_counts = df.isnull().sum()
for col, count in null_counts.items():
if count > 0:
issues.append(
ValidationIssue(
check="null_check",
table=table_name,
column=str(col),
message=f"Column '{col}' has {count} null value(s)",
severity="error",
)
)
return issues
def check_foreign_keys(self) -> list[ValidationIssue]:
"""Validate all defined foreign key relationships.
Checks that all values in child columns are present in the
corresponding parent column.
Returns:
List of ValidationIssue for any FK violations.
"""
issues: list[ValidationIssue] = []
for child_table, child_col, parent_table, parent_col in FOREIGN_KEY_RELATIONSHIPS:
try:
child_df = self._loader.load_table(child_table)
parent_df = self._loader.load_table(parent_table)
except DatasetError:
# Tables missing β already reported by check_table_presence
continue
if child_col not in child_df.columns:
issues.append(
ValidationIssue(
check="foreign_key",
table=child_table,
column=child_col,
message=(
f"FK column '{child_col}' not found in table '{child_table}'"
),
severity="error",
)
)
continue
if parent_col not in parent_df.columns:
issues.append(
ValidationIssue(
check="foreign_key",
table=parent_table,
column=parent_col,
message=(
f"Referenced column '{parent_col}' not found in "
f"table '{parent_table}'"
),
severity="error",
)
)
continue
child_values = set(child_df[child_col].dropna().unique())
parent_values = set(parent_df[parent_col].dropna().unique())
orphans = child_values - parent_values
if orphans:
sample = sorted(str(v) for v in list(orphans)[:5])
issues.append(
ValidationIssue(
check="foreign_key",
table=child_table,
column=child_col,
message=(
f"FK violation: {len(orphans)} value(s) in "
f"'{child_table}.{child_col}' not found in "
f"'{parent_table}.{parent_col}'. "
f"Sample: {sample}"
),
severity="error",
)
)
return issues
def check_split_presence(self) -> list[ValidationIssue]:
"""Verify training tables have train_split column with exactly {train, validation, test}.
Returns:
List of ValidationIssue for any split column problems.
"""
issues: list[ValidationIssue] = []
for table_name in TRAINING_TABLES:
try:
df = self._loader.load_table(table_name)
except DatasetError:
continue
if "train_split" not in df.columns:
issues.append(
ValidationIssue(
check="split_presence",
table=table_name,
column="train_split",
message=(
f"Training table '{table_name}' is missing "
f"'train_split' column"
),
severity="error",
)
)
continue
actual_splits = set(df["train_split"].dropna().unique())
if actual_splits != VALID_SPLITS:
missing = VALID_SPLITS - actual_splits
extra = actual_splits - VALID_SPLITS
parts = []
if missing:
parts.append(f"missing splits: {sorted(missing)}")
if extra:
parts.append(f"unexpected splits: {sorted(extra)}")
issues.append(
ValidationIssue(
check="split_presence",
table=table_name,
column="train_split",
message=(
f"Split values mismatch in '{table_name}': "
f"{'; '.join(parts)}. "
f"Expected exactly {{train, validation, test}}, "
f"got {sorted(actual_splits)}"
),
severity="error",
)
)
return issues
def check_target_labels(self) -> list[ValidationIssue]:
"""Verify target columns exist in their respective training tables.
Returns:
List of ValidationIssue for any missing target columns.
"""
issues: list[ValidationIssue] = []
for table_name, column_name in TARGET_COLUMNS:
try:
df = self._loader.load_table(table_name)
except DatasetError:
continue
if column_name not in df.columns:
issues.append(
ValidationIssue(
check="target_labels",
table=table_name,
column=column_name,
message=(
f"Target column '{column_name}' not found in "
f"table '{table_name}'"
),
severity="error",
)
)
return issues
def run_all(self) -> ValidationReport:
"""Execute all validation checks and aggregate results.
Uses the collect-all-errors pattern: every check runs regardless
of whether previous checks found issues.
Returns:
A ValidationReport with all issues and pass/fail status.
"""
all_issues: list[ValidationIssue] = []
checks_run = 0
checks_passed = 0
checks = [
("table_presence", self.check_table_presence),
("row_counts", self.check_row_counts),
("null_values", self.check_null_values),
("foreign_keys", self.check_foreign_keys),
("split_presence", self.check_split_presence),
("target_labels", self.check_target_labels),
]
for check_name, check_fn in checks:
checks_run += 1
try:
issues = check_fn()
all_issues.extend(issues)
if not issues:
checks_passed += 1
logger.info(
"Check '%s': %s (%d issue(s))",
check_name,
"PASS" if not issues else "FAIL",
len(issues),
)
except Exception as exc:
all_issues.append(
ValidationIssue(
check=check_name,
table="",
column=None,
message=f"Check raised unexpected error: {exc}",
severity="error",
)
)
logger.error("Check '%s' raised an exception: %s", check_name, exc)
report = ValidationReport(
timestamp=datetime.now(timezone.utc),
passed=len(all_issues) == 0,
issues=all_issues,
checks_run=checks_run,
checks_passed=checks_passed,
)
logger.info(
"Validation complete: %d/%d checks passed, %d total issue(s)",
checks_passed,
checks_run,
len(all_issues),
)
return report
def write_report(self, report: ValidationReport, output_path: Path) -> None:
"""Write a markdown validation report to the specified path.
Creates parent directories if they don't exist. The report includes
a timestamp, overall pass/fail status, and details for each issue.
Args:
report: The ValidationReport to write.
output_path: Path where the markdown report will be written.
"""
output_path = Path(output_path)
output_path.parent.mkdir(parents=True, exist_ok=True)
lines: list[str] = []
lines.append("# Dataset Validation Report")
lines.append("")
lines.append(f"**Timestamp:** {report.timestamp.isoformat()}")
lines.append(f"**Status:** {'PASSED β' if report.passed else 'FAILED β'}")
lines.append(f"**Checks Run:** {report.checks_run}")
lines.append(f"**Checks Passed:** {report.checks_passed}")
lines.append(f"**Total Issues:** {len(report.issues)}")
lines.append("")
# Summary table of checks
lines.append("## Check Summary")
lines.append("")
lines.append("| Check | Status |")
lines.append("|-------|--------|")
check_names = [
"table_presence",
"row_count",
"null_check",
"foreign_key",
"split_presence",
"target_labels",
]
check_display = {
"table_presence": "Table Presence",
"row_count": "Row Counts",
"null_check": "Null Values",
"foreign_key": "Foreign Keys",
"split_presence": "Split Presence",
"target_labels": "Target Labels",
}
failed_checks = {issue.check for issue in report.issues}
for check in check_names:
status = "β FAIL" if check in failed_checks else "β PASS"
display = check_display.get(check, check)
lines.append(f"| {display} | {status} |")
lines.append("")
if report.issues:
lines.append("## Issues")
lines.append("")
for i, issue in enumerate(report.issues, 1):
col_info = f" (column: `{issue.column}`)" if issue.column else ""
lines.append(
f"{i}. **[{issue.severity.upper()}]** `{issue.table}`{col_info}: "
f"{issue.message}"
)
lines.append("")
else:
lines.append("## Result")
lines.append("")
lines.append("All validation checks passed. Dataset is ready for use.")
lines.append("")
content = "\n".join(lines)
output_path.write_text(content, encoding="utf-8")
logger.info("Validation report written to %s", output_path)
|