Spaces:
Sleeping
Sleeping
Update src/test_set.py
Browse files- src/test_set.py +93 -516
src/test_set.py
CHANGED
@@ -12,31 +12,22 @@ from config import (
|
|
12 |
ALL_UG40_LANGUAGES,
|
13 |
GOOGLE_SUPPORTED_LANGUAGES,
|
14 |
EVALUATION_TRACKS,
|
15 |
-
|
16 |
-
STATISTICAL_CONFIG,
|
17 |
)
|
18 |
import salt.dataset
|
19 |
-
from src.utils import get_all_language_pairs
|
20 |
from typing import Dict, List, Optional, Tuple
|
21 |
|
22 |
|
23 |
-
|
24 |
# Local CSV filenames for persistence
|
25 |
-
LOCAL_PUBLIC_CSV = "
|
26 |
-
LOCAL_COMPLETE_CSV = "
|
27 |
-
LOCAL_TRACK_CSVS = {
|
28 |
-
track: f"salt_test_set_{track}.csv" for track in EVALUATION_TRACKS.keys()
|
29 |
-
}
|
30 |
|
31 |
|
32 |
-
def
|
33 |
-
|
34 |
-
stratified_sampling: bool = True,
|
35 |
-
balance_tracks: bool = True,
|
36 |
-
) -> pd.DataFrame:
|
37 |
-
"""Generate scientifically rigorous test set with stratified sampling."""
|
38 |
|
39 |
-
print("🔬 Generating
|
40 |
|
41 |
try:
|
42 |
# Build SALT dataset config
|
@@ -63,22 +54,12 @@ def generate_scientific_test_set(
|
|
63 |
test_samples = []
|
64 |
sample_id_counter = 1
|
65 |
|
66 |
-
#
|
67 |
-
track_targets = calculate_track_sampling_targets(balance_tracks)
|
68 |
-
|
69 |
-
# Generate samples for each language pair with stratified sampling
|
70 |
for src_lang in ALL_UG40_LANGUAGES:
|
71 |
for tgt_lang in ALL_UG40_LANGUAGES:
|
72 |
if src_lang == tgt_lang:
|
73 |
continue
|
74 |
|
75 |
-
# Determine target sample size for this pair
|
76 |
-
pair_targets = calculate_pair_sampling_targets(
|
77 |
-
src_lang, tgt_lang, track_targets, max_samples_per_pair
|
78 |
-
)
|
79 |
-
|
80 |
-
target_samples = max(pair_targets.values()) if pair_targets else max_samples_per_pair
|
81 |
-
|
82 |
# Filter for this language pair
|
83 |
pair_data = full_data[
|
84 |
(full_data["source.language"] == src_lang) &
|
@@ -89,24 +70,13 @@ def generate_scientific_test_set(
|
|
89 |
print(f"⚠️ No data found for {src_lang} → {tgt_lang}")
|
90 |
continue
|
91 |
|
92 |
-
#
|
93 |
-
|
94 |
-
|
95 |
-
else:
|
96 |
-
# Simple random sampling
|
97 |
-
n_samples = min(len(pair_data), target_samples)
|
98 |
-
sampled = pair_data.sample(n=n_samples, random_state=42)
|
99 |
|
100 |
print(f"✅ {src_lang} → {tgt_lang}: {len(sampled)} samples")
|
101 |
|
102 |
for _, row in sampled.iterrows():
|
103 |
-
# Determine which tracks include this pair
|
104 |
-
tracks_included = []
|
105 |
-
for track_name, track_config in EVALUATION_TRACKS.items():
|
106 |
-
if (src_lang in track_config["languages"] and
|
107 |
-
tgt_lang in track_config["languages"]):
|
108 |
-
tracks_included.append(track_name)
|
109 |
-
|
110 |
test_samples.append({
|
111 |
"sample_id": f"salt_{sample_id_counter:06d}",
|
112 |
"source_text": row["source"],
|
@@ -118,10 +88,6 @@ def generate_scientific_test_set(
|
|
118 |
src_lang in GOOGLE_SUPPORTED_LANGUAGES and
|
119 |
tgt_lang in GOOGLE_SUPPORTED_LANGUAGES
|
120 |
),
|
121 |
-
"tracks_included": ",".join(tracks_included),
|
122 |
-
"statistical_weight": calculate_statistical_weight(
|
123 |
-
src_lang, tgt_lang, tracks_included
|
124 |
-
),
|
125 |
})
|
126 |
sample_id_counter += 1
|
127 |
|
@@ -130,313 +96,70 @@ def generate_scientific_test_set(
|
|
130 |
if test_df.empty:
|
131 |
raise ValueError("No test samples generated - check SALT dataset availability")
|
132 |
|
133 |
-
|
134 |
-
adequacy_report = validate_test_set_scientific_adequacy(test_df)
|
135 |
-
|
136 |
-
print(f"✅ Generated scientific test set: {len(test_df):,} samples")
|
137 |
-
print(f"📈 Test set adequacy: {adequacy_report['overall_adequacy']}")
|
138 |
|
139 |
return test_df
|
140 |
|
141 |
except Exception as e:
|
142 |
-
print(f"❌ Error generating
|
143 |
return pd.DataFrame(columns=[
|
144 |
"sample_id", "source_text", "target_text", "source_language",
|
145 |
-
"target_language", "domain", "google_comparable"
|
146 |
-
"statistical_weight"
|
147 |
])
|
148 |
|
149 |
|
150 |
-
def
|
151 |
-
"""
|
152 |
-
|
153 |
-
track_targets = {}
|
154 |
-
|
155 |
-
for track_name, track_config in EVALUATION_TRACKS.items():
|
156 |
-
# Base requirement from config
|
157 |
-
min_per_pair = track_config["min_samples_per_pair"]
|
158 |
-
|
159 |
-
# Number of language pairs in this track
|
160 |
-
n_pairs = len(track_config["languages"]) * (len(track_config["languages"]) - 1)
|
161 |
-
|
162 |
-
# Calculate total samples needed for statistical adequacy
|
163 |
-
if balance_tracks:
|
164 |
-
# Use publication-quality recommendation
|
165 |
-
target_per_pair = max(
|
166 |
-
min_per_pair,
|
167 |
-
SAMPLE_SIZE_RECOMMENDATIONS["publication_quality"] // n_pairs
|
168 |
-
)
|
169 |
-
else:
|
170 |
-
target_per_pair = min_per_pair
|
171 |
-
|
172 |
-
track_targets[track_name] = target_per_pair * n_pairs
|
173 |
-
|
174 |
-
print(f"📊 {track_name}: targeting {target_per_pair} samples/pair × {n_pairs} pairs = {track_targets[track_name]} total")
|
175 |
-
|
176 |
-
return track_targets
|
177 |
-
|
178 |
-
|
179 |
-
def calculate_pair_sampling_targets(
|
180 |
-
src_lang: str, tgt_lang: str, track_targets: Dict[str, int], max_samples: int
|
181 |
-
) -> Dict[str, int]:
|
182 |
-
"""Calculate sampling targets for a specific language pair across tracks."""
|
183 |
-
|
184 |
-
pair_targets = {}
|
185 |
-
|
186 |
-
for track_name, track_config in EVALUATION_TRACKS.items():
|
187 |
-
if (src_lang in track_config["languages"] and
|
188 |
-
tgt_lang in track_config["languages"]):
|
189 |
-
|
190 |
-
n_pairs_in_track = len(track_config["languages"]) * (len(track_config["languages"]) - 1)
|
191 |
-
target_per_pair = track_targets[track_name] // n_pairs_in_track
|
192 |
-
|
193 |
-
pair_targets[track_name] = min(target_per_pair, max_samples)
|
194 |
-
|
195 |
-
return pair_targets
|
196 |
-
|
197 |
-
|
198 |
-
def stratified_sample_pair_data(pair_data: pd.DataFrame, target_samples: int) -> pd.DataFrame:
|
199 |
-
"""Perform stratified sampling on pair data to ensure representativeness."""
|
200 |
-
|
201 |
-
# Try to stratify by domain if available
|
202 |
-
if "domain" in pair_data.columns and pair_data["domain"].nunique() > 1:
|
203 |
-
# Sample proportionally from each domain
|
204 |
-
domain_counts = pair_data["domain"].value_counts()
|
205 |
-
sampled_parts = []
|
206 |
-
|
207 |
-
for domain, count in domain_counts.items():
|
208 |
-
domain_data = pair_data[pair_data["domain"] == domain]
|
209 |
-
|
210 |
-
# Calculate proportional sample size
|
211 |
-
proportion = count / len(pair_data)
|
212 |
-
domain_target = max(1, int(target_samples * proportion))
|
213 |
-
domain_target = min(domain_target, len(domain_data))
|
214 |
-
|
215 |
-
if len(domain_data) >= domain_target:
|
216 |
-
domain_sample = domain_data.sample(n=domain_target, random_state=42)
|
217 |
-
sampled_parts.append(domain_sample)
|
218 |
-
|
219 |
-
if sampled_parts:
|
220 |
-
stratified_sample = pd.concat(sampled_parts, ignore_index=True)
|
221 |
-
|
222 |
-
# If we didn't get enough samples, fill with random sampling
|
223 |
-
if len(stratified_sample) < target_samples:
|
224 |
-
remaining_data = pair_data[~pair_data.index.isin(stratified_sample.index)]
|
225 |
-
additional_needed = target_samples - len(stratified_sample)
|
226 |
-
|
227 |
-
if len(remaining_data) >= additional_needed:
|
228 |
-
additional_sample = remaining_data.sample(n=additional_needed, random_state=42)
|
229 |
-
stratified_sample = pd.concat([stratified_sample, additional_sample], ignore_index=True)
|
230 |
-
|
231 |
-
return stratified_sample.head(target_samples)
|
232 |
-
|
233 |
-
# Fallback to simple random sampling
|
234 |
-
return pair_data.sample(n=min(target_samples, len(pair_data)), random_state=42)
|
235 |
-
|
236 |
-
|
237 |
-
def calculate_statistical_weight(
|
238 |
-
src_lang: str, tgt_lang: str, tracks_included: List[str]
|
239 |
-
) -> float:
|
240 |
-
"""Calculate statistical weight for a sample based on track inclusion."""
|
241 |
|
242 |
-
|
243 |
-
weight = 1.0
|
244 |
|
245 |
-
|
246 |
-
weight *= len(tracks_included)
|
247 |
-
|
248 |
-
# Higher weight for Google-comparable pairs (enable baseline comparison)
|
249 |
-
if (src_lang in GOOGLE_SUPPORTED_LANGUAGES and
|
250 |
-
tgt_lang in GOOGLE_SUPPORTED_LANGUAGES):
|
251 |
-
weight *= 1.5
|
252 |
-
|
253 |
-
# Normalize to reasonable range
|
254 |
-
return min(weight, 5.0)
|
255 |
-
|
256 |
-
|
257 |
-
def validate_test_set_scientific_adequacy(test_df: pd.DataFrame) -> Dict:
|
258 |
-
"""Validate that the test set meets scientific adequacy requirements."""
|
259 |
-
|
260 |
-
adequacy_report = {
|
261 |
-
"overall_adequacy": "insufficient",
|
262 |
-
"track_adequacy": {},
|
263 |
-
"issues": [],
|
264 |
-
"recommendations": [],
|
265 |
-
"statistics": {},
|
266 |
-
}
|
267 |
-
|
268 |
-
if test_df.empty:
|
269 |
-
adequacy_report["issues"].append("Test set is empty")
|
270 |
-
return adequacy_report
|
271 |
-
|
272 |
-
# Check each track
|
273 |
-
track_adequacies = []
|
274 |
-
|
275 |
-
for track_name, track_config in EVALUATION_TRACKS.items():
|
276 |
-
track_languages = track_config["languages"]
|
277 |
-
min_per_pair = track_config["min_samples_per_pair"]
|
278 |
-
|
279 |
-
# Filter to track data
|
280 |
-
track_data = test_df[
|
281 |
-
(test_df["source_language"].isin(track_languages)) &
|
282 |
-
(test_df["target_language"].isin(track_languages))
|
283 |
-
]
|
284 |
-
|
285 |
-
# Analyze pair coverage
|
286 |
-
pair_counts = {}
|
287 |
-
for src in track_languages:
|
288 |
-
for tgt in track_languages:
|
289 |
-
if src == tgt:
|
290 |
-
continue
|
291 |
-
|
292 |
-
pair_samples = track_data[
|
293 |
-
(track_data["source_language"] == src) &
|
294 |
-
(track_data["target_language"] == tgt)
|
295 |
-
]
|
296 |
-
pair_counts[f"{src}_{tgt}"] = len(pair_samples)
|
297 |
-
|
298 |
-
# Calculate adequacy metrics
|
299 |
-
total_pairs = len(pair_counts)
|
300 |
-
adequate_pairs = sum(1 for count in pair_counts.values() if count >= min_per_pair)
|
301 |
-
adequacy_rate = adequate_pairs / max(total_pairs, 1)
|
302 |
-
|
303 |
-
# Determine track adequacy level
|
304 |
-
if adequacy_rate >= 0.9:
|
305 |
-
track_adequacy = "excellent"
|
306 |
-
elif adequacy_rate >= 0.8:
|
307 |
-
track_adequacy = "good"
|
308 |
-
elif adequacy_rate >= 0.6:
|
309 |
-
track_adequacy = "fair"
|
310 |
-
else:
|
311 |
-
track_adequacy = "insufficient"
|
312 |
-
|
313 |
-
adequacy_report["track_adequacy"][track_name] = {
|
314 |
-
"adequacy": track_adequacy,
|
315 |
-
"adequacy_rate": adequacy_rate,
|
316 |
-
"total_samples": len(track_data),
|
317 |
-
"total_pairs": total_pairs,
|
318 |
-
"adequate_pairs": adequate_pairs,
|
319 |
-
"min_samples_per_pair": min_per_pair,
|
320 |
-
"pair_counts": pair_counts,
|
321 |
-
}
|
322 |
-
|
323 |
-
track_adequacies.append(track_adequacy)
|
324 |
-
|
325 |
-
# Add specific issues
|
326 |
-
if track_adequacy == "insufficient":
|
327 |
-
inadequate_pairs = [k for k, v in pair_counts.items() if v < min_per_pair]
|
328 |
-
adequacy_report["issues"].append(
|
329 |
-
f"{track_name}: {len(inadequate_pairs)} pairs below minimum"
|
330 |
-
)
|
331 |
-
|
332 |
-
# Overall adequacy assessment
|
333 |
-
if all(adequacy in ["excellent", "good"] for adequacy in track_adequacies):
|
334 |
-
adequacy_report["overall_adequacy"] = "excellent"
|
335 |
-
elif all(adequacy in ["excellent", "good", "fair"] for adequacy in track_adequacies):
|
336 |
-
adequacy_report["overall_adequacy"] = "good"
|
337 |
-
elif any(adequacy in ["good", "fair"] for adequacy in track_adequacies):
|
338 |
-
adequacy_report["overall_adequacy"] = "fair"
|
339 |
-
else:
|
340 |
-
adequacy_report["overall_adequacy"] = "insufficient"
|
341 |
-
|
342 |
-
# Overall statistics
|
343 |
-
adequacy_report["statistics"] = {
|
344 |
-
"total_samples": len(test_df),
|
345 |
-
"total_language_pairs": len(test_df.groupby(["source_language", "target_language"])),
|
346 |
-
"google_comparable_samples": int(test_df["google_comparable"].sum()),
|
347 |
-
"domain_distribution": test_df["domain"].value_counts().to_dict(),
|
348 |
-
"track_sample_distribution": {
|
349 |
-
track: adequacy_report["track_adequacy"][track]["total_samples"]
|
350 |
-
for track in EVALUATION_TRACKS.keys()
|
351 |
-
},
|
352 |
-
}
|
353 |
-
|
354 |
-
# Generate recommendations
|
355 |
-
if adequacy_report["overall_adequacy"] in ["insufficient", "fair"]:
|
356 |
-
adequacy_report["recommendations"].append(
|
357 |
-
"Consider increasing sample size for better statistical power"
|
358 |
-
)
|
359 |
-
|
360 |
-
if adequacy_report["statistics"]["google_comparable_samples"] < 1000:
|
361 |
-
adequacy_report["recommendations"].append(
|
362 |
-
"More Google-comparable samples recommended for baseline comparison"
|
363 |
-
)
|
364 |
-
|
365 |
-
return adequacy_report
|
366 |
-
|
367 |
-
|
368 |
-
def _generate_and_save_scientific_test_set() -> Tuple[pd.DataFrame, pd.DataFrame]:
|
369 |
-
"""Generate and save both public and complete versions of the scientific test set."""
|
370 |
-
|
371 |
-
print("🔬 Generating and saving scientific test sets...")
|
372 |
-
|
373 |
-
full_df = generate_scientific_test_set()
|
374 |
|
375 |
if full_df.empty:
|
376 |
-
print("❌ Failed to generate
|
377 |
empty_public = pd.DataFrame(columns=[
|
378 |
"sample_id", "source_text", "source_language",
|
379 |
-
"target_language", "domain", "google_comparable"
|
380 |
-
"tracks_included", "statistical_weight"
|
381 |
])
|
382 |
empty_complete = pd.DataFrame(columns=[
|
383 |
"sample_id", "source_text", "target_text", "source_language",
|
384 |
-
"target_language", "domain", "google_comparable"
|
385 |
-
"tracks_included", "statistical_weight"
|
386 |
])
|
387 |
return empty_public, empty_complete
|
388 |
|
389 |
# Public version (no target_text)
|
390 |
public_df = full_df[[
|
391 |
"sample_id", "source_text", "source_language",
|
392 |
-
"target_language", "domain", "google_comparable"
|
393 |
-
"tracks_included", "statistical_weight"
|
394 |
]].copy()
|
395 |
|
396 |
-
# Save
|
397 |
try:
|
398 |
public_df.to_csv(LOCAL_PUBLIC_CSV, index=False)
|
399 |
full_df.to_csv(LOCAL_COMPLETE_CSV, index=False)
|
400 |
-
print(f"✅ Saved
|
401 |
except Exception as e:
|
402 |
-
print(f"⚠️ Error saving
|
403 |
-
|
404 |
-
# Save track-specific versions for easier analysis
|
405 |
-
for track_name, track_config in EVALUATION_TRACKS.items():
|
406 |
-
try:
|
407 |
-
track_languages = track_config["languages"]
|
408 |
-
track_public = public_df[
|
409 |
-
(public_df["source_language"].isin(track_languages)) &
|
410 |
-
(public_df["target_language"].isin(track_languages))
|
411 |
-
]
|
412 |
-
|
413 |
-
track_filename = LOCAL_TRACK_CSVS[track_name]
|
414 |
-
track_public.to_csv(track_filename, index=False)
|
415 |
-
print(f"✅ Saved {track_name} track: {track_filename} ({len(track_public):,} samples)")
|
416 |
-
|
417 |
-
except Exception as e:
|
418 |
-
print(f"⚠️ Error saving {track_name} track CSV: {e}")
|
419 |
|
420 |
return public_df, full_df
|
421 |
|
422 |
|
423 |
-
def
|
424 |
-
"""Load the
|
425 |
|
426 |
# 1) Try HF Hub
|
427 |
try:
|
428 |
-
print("📥 Attempting to load
|
429 |
-
ds = load_dataset(TEST_SET_DATASET
|
430 |
df = ds.to_pandas()
|
431 |
|
432 |
-
# Validate
|
433 |
-
required_cols = ["sample_id", "source_text", "source_language", "target_language"
|
434 |
-
"tracks_included", "statistical_weight"]
|
435 |
if all(col in df.columns for col in required_cols):
|
436 |
-
print(f"✅ Loaded
|
437 |
return df
|
438 |
else:
|
439 |
-
print("⚠️ HF Hub test set missing
|
440 |
|
441 |
except Exception as e:
|
442 |
print(f"⚠️ HF Hub load failed: {e}")
|
@@ -447,35 +170,34 @@ def get_public_test_set_scientific() -> pd.DataFrame:
|
|
447 |
df = pd.read_csv(LOCAL_PUBLIC_CSV)
|
448 |
required_cols = ["sample_id", "source_text", "source_language", "target_language"]
|
449 |
if all(col in df.columns for col in required_cols):
|
450 |
-
print(f"✅ Loaded
|
451 |
return df
|
452 |
else:
|
453 |
print("⚠️ Local CSV has invalid structure, regenerating...")
|
454 |
except Exception as e:
|
455 |
-
print(f"⚠️ Failed to read local
|
456 |
|
457 |
# 3) Regenerate & save
|
458 |
-
print("🔄 Generating new
|
459 |
-
public_df, _ =
|
460 |
return public_df
|
461 |
|
462 |
|
463 |
-
def
|
464 |
-
"""Load the complete
|
465 |
|
466 |
# 1) Try HF Hub private
|
467 |
try:
|
468 |
-
print("📥 Attempting to load complete
|
469 |
-
ds = load_dataset(TEST_SET_DATASET + "-
|
470 |
df = ds.to_pandas()
|
471 |
|
472 |
-
required_cols = ["sample_id", "source_text", "target_text", "source_language",
|
473 |
-
"target_language", "tracks_included", "statistical_weight"]
|
474 |
if all(col in df.columns for col in required_cols):
|
475 |
-
print(f"✅ Loaded complete
|
476 |
return df
|
477 |
else:
|
478 |
-
print("⚠️ HF Hub complete test set missing
|
479 |
|
480 |
except Exception as e:
|
481 |
print(f"⚠️ HF Hub private load failed: {e}")
|
@@ -486,63 +208,31 @@ def get_complete_test_set_scientific() -> pd.DataFrame:
|
|
486 |
df = pd.read_csv(LOCAL_COMPLETE_CSV)
|
487 |
required_cols = ["sample_id", "source_text", "target_text", "source_language", "target_language"]
|
488 |
if all(col in df.columns for col in required_cols):
|
489 |
-
print(f"✅ Loaded complete
|
490 |
return df
|
491 |
else:
|
492 |
print("⚠️ Local complete CSV has invalid structure, regenerating...")
|
493 |
except Exception as e:
|
494 |
-
print(f"⚠️ Failed to read local complete
|
495 |
|
496 |
# 3) Regenerate & save
|
497 |
-
print("🔄 Generating new complete
|
498 |
-
_, complete_df =
|
499 |
return complete_df
|
500 |
|
501 |
|
502 |
-
def
|
503 |
-
"""
|
504 |
-
|
505 |
-
if track not in EVALUATION_TRACKS:
|
506 |
-
print(f"❌ Unknown track: {track}")
|
507 |
-
return pd.DataFrame()
|
508 |
-
|
509 |
-
# Try track-specific CSV first
|
510 |
-
track_csv = LOCAL_TRACK_CSVS.get(track)
|
511 |
-
if track_csv and os.path.exists(track_csv):
|
512 |
-
try:
|
513 |
-
df = pd.read_csv(track_csv)
|
514 |
-
print(f"✅ Loaded {track} test set from track-specific CSV ({len(df):,} samples)")
|
515 |
-
return df
|
516 |
-
except Exception as e:
|
517 |
-
print(f"⚠️ Failed to read {track} CSV: {e}")
|
518 |
-
|
519 |
-
# Fallback to filtering main test set
|
520 |
-
public_df = get_public_test_set_scientific()
|
521 |
-
|
522 |
-
if public_df.empty:
|
523 |
-
return pd.DataFrame()
|
524 |
-
|
525 |
-
track_languages = EVALUATION_TRACKS[track]["languages"]
|
526 |
-
track_df = public_df[
|
527 |
-
(public_df["source_language"].isin(track_languages)) &
|
528 |
-
(public_df["target_language"].isin(track_languages))
|
529 |
-
]
|
530 |
-
|
531 |
-
print(f"✅ Filtered {track} test set from main set ({len(track_df):,} samples)")
|
532 |
-
return track_df
|
533 |
-
|
534 |
-
|
535 |
-
def create_test_set_download_scientific() -> Tuple[str, Dict]:
|
536 |
-
"""Create scientific test set download with comprehensive metadata."""
|
537 |
|
538 |
-
public_df =
|
539 |
|
540 |
if public_df.empty:
|
541 |
stats = {
|
542 |
"total_samples": 0,
|
543 |
"track_breakdown": {},
|
544 |
-
"
|
545 |
-
"
|
|
|
546 |
}
|
547 |
return LOCAL_PUBLIC_CSV, stats
|
548 |
|
@@ -552,7 +242,7 @@ def create_test_set_download_scientific() -> Tuple[str, Dict]:
|
|
552 |
try:
|
553 |
public_df.to_csv(download_path, index=False)
|
554 |
except Exception as e:
|
555 |
-
print(f"⚠️ Error updating
|
556 |
|
557 |
# Calculate comprehensive statistics
|
558 |
try:
|
@@ -560,7 +250,7 @@ def create_test_set_download_scientific() -> Tuple[str, Dict]:
|
|
560 |
stats = {
|
561 |
"total_samples": len(public_df),
|
562 |
"languages": sorted(list(set(public_df["source_language"]).union(public_df["target_language"]))),
|
563 |
-
"
|
564 |
}
|
565 |
|
566 |
# Track-specific breakdown
|
@@ -573,11 +263,9 @@ def create_test_set_download_scientific() -> Tuple[str, Dict]:
|
|
573 |
]
|
574 |
|
575 |
track_breakdown[track_name] = {
|
576 |
-
"name": track_config["name"],
|
577 |
"total_samples": len(track_data),
|
578 |
"language_pairs": len(track_data.groupby(["source_language", "target_language"])),
|
579 |
-
"
|
580 |
-
"statistical_adequacy": len(track_data) >= track_config["min_samples_per_pair"] * len(track_languages) * (len(track_languages) - 1),
|
581 |
}
|
582 |
|
583 |
stats["track_breakdown"] = track_breakdown
|
@@ -585,52 +273,56 @@ def create_test_set_download_scientific() -> Tuple[str, Dict]:
|
|
585 |
# Google-comparable statistics
|
586 |
if "google_comparable" in public_df.columns:
|
587 |
stats["google_comparable_samples"] = int(public_df["google_comparable"].sum())
|
588 |
-
stats["google_comparable_rate"] = float(public_df["google_comparable"].mean())
|
589 |
else:
|
590 |
stats["google_comparable_samples"] = 0
|
591 |
-
stats["google_comparable_rate"] = 0.0
|
592 |
-
|
593 |
-
# Scientific adequacy assessment
|
594 |
-
adequacy_report = validate_test_set_scientific_adequacy(public_df)
|
595 |
-
stats["adequacy_assessment"] = adequacy_report["overall_adequacy"]
|
596 |
-
stats["adequacy_details"] = adequacy_report
|
597 |
-
|
598 |
-
# Scientific metadata
|
599 |
-
stats["scientific_metadata"] = {
|
600 |
-
"stratified_sampling": True,
|
601 |
-
"statistical_weighting": "statistical_weight" in public_df.columns,
|
602 |
-
"track_balanced": True,
|
603 |
-
"confidence_level": STATISTICAL_CONFIG["confidence_level"],
|
604 |
-
"recommended_for": [
|
605 |
-
track for track, info in track_breakdown.items()
|
606 |
-
if info["statistical_adequacy"]
|
607 |
-
],
|
608 |
-
}
|
609 |
|
610 |
except Exception as e:
|
611 |
-
print(f"⚠️ Error calculating
|
612 |
stats = {
|
613 |
"total_samples": len(public_df),
|
614 |
"track_breakdown": {},
|
615 |
-
"
|
616 |
-
"
|
|
|
617 |
}
|
618 |
|
619 |
return download_path, stats
|
620 |
|
621 |
|
622 |
-
def
|
623 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
624 |
|
625 |
try:
|
626 |
-
public_df =
|
627 |
-
complete_df =
|
628 |
|
629 |
if public_df.empty or complete_df.empty:
|
630 |
return {
|
631 |
"alignment_check": False,
|
632 |
"total_samples": 0,
|
633 |
-
"scientific_adequacy": {},
|
634 |
"track_analysis": {},
|
635 |
"error": "Test sets are empty or could not be loaded",
|
636 |
}
|
@@ -642,7 +334,6 @@ def validate_test_set_integrity_scientific() -> Dict:
|
|
642 |
track_analysis = {}
|
643 |
for track_name, track_config in EVALUATION_TRACKS.items():
|
644 |
track_languages = track_config["languages"]
|
645 |
-
min_per_pair = track_config["min_samples_per_pair"]
|
646 |
|
647 |
# Analyze public set for this track
|
648 |
track_public = public_df[
|
@@ -656,140 +347,26 @@ def validate_test_set_integrity_scientific() -> Dict:
|
|
656 |
(complete_df["target_language"].isin(track_languages))
|
657 |
]
|
658 |
|
659 |
-
# Calculate coverage
|
660 |
-
pair_coverage = {}
|
661 |
-
for src in track_languages:
|
662 |
-
for tgt in track_languages:
|
663 |
-
if src == tgt:
|
664 |
-
continue
|
665 |
-
|
666 |
-
public_subset = track_public[
|
667 |
-
(track_public["source_language"] == src) &
|
668 |
-
(track_public["target_language"] == tgt)
|
669 |
-
]
|
670 |
-
|
671 |
-
complete_subset = track_complete[
|
672 |
-
(track_complete["source_language"] == src) &
|
673 |
-
(track_complete["target_language"] == tgt)
|
674 |
-
]
|
675 |
-
|
676 |
-
pair_coverage[f"{src}_{tgt}"] = {
|
677 |
-
"public_count": len(public_subset),
|
678 |
-
"complete_count": len(complete_subset),
|
679 |
-
"alignment": len(public_subset) == len(complete_subset),
|
680 |
-
"meets_minimum": len(public_subset) >= min_per_pair,
|
681 |
-
}
|
682 |
-
|
683 |
-
# Track summary
|
684 |
-
total_pairs = len(pair_coverage)
|
685 |
-
adequate_pairs = sum(1 for info in pair_coverage.values() if info["meets_minimum"])
|
686 |
-
aligned_pairs = sum(1 for info in pair_coverage.values() if info["alignment"])
|
687 |
-
|
688 |
track_analysis[track_name] = {
|
689 |
-
"
|
690 |
-
"
|
691 |
-
"
|
692 |
-
"
|
693 |
-
"alignment_rate": aligned_pairs / max(total_pairs, 1),
|
694 |
-
"pair_coverage": pair_coverage,
|
695 |
-
"statistical_power": calculate_track_statistical_power(track_public, track_config),
|
696 |
}
|
697 |
|
698 |
-
# Overall scientific adequacy
|
699 |
-
adequacy_report = validate_test_set_scientific_adequacy(public_df)
|
700 |
-
|
701 |
return {
|
702 |
"alignment_check": public_ids <= private_ids,
|
703 |
"total_samples": len(public_df),
|
704 |
"track_analysis": track_analysis,
|
705 |
-
"scientific_adequacy": adequacy_report,
|
706 |
"public_samples": len(public_df),
|
707 |
"private_samples": len(complete_df),
|
708 |
"id_alignment_rate": len(public_ids & private_ids) / len(public_ids) if public_ids else 0.0,
|
709 |
-
"integrity_score": calculate_integrity_score(track_analysis, adequacy_report),
|
710 |
}
|
711 |
|
712 |
except Exception as e:
|
713 |
return {
|
714 |
"alignment_check": False,
|
715 |
"total_samples": 0,
|
716 |
-
"scientific_adequacy": {},
|
717 |
"track_analysis": {},
|
718 |
"error": f"Validation failed: {str(e)}",
|
719 |
-
}
|
720 |
-
|
721 |
-
|
722 |
-
def calculate_track_statistical_power(track_data: pd.DataFrame, track_config: Dict) -> float:
|
723 |
-
"""Calculate statistical power estimate for a track."""
|
724 |
-
|
725 |
-
if track_data.empty:
|
726 |
-
return 0.0
|
727 |
-
|
728 |
-
# Simple power estimation based on sample size
|
729 |
-
min_required = track_config["min_samples_per_pair"]
|
730 |
-
languages = track_config["languages"]
|
731 |
-
total_pairs = len(languages) * (len(languages) - 1)
|
732 |
-
|
733 |
-
# Calculate average samples per pair
|
734 |
-
pair_counts = []
|
735 |
-
for src in languages:
|
736 |
-
for tgt in languages:
|
737 |
-
if src == tgt:
|
738 |
-
continue
|
739 |
-
|
740 |
-
pair_samples = track_data[
|
741 |
-
(track_data["source_language"] == src) &
|
742 |
-
(track_data["target_language"] == tgt)
|
743 |
-
]
|
744 |
-
pair_counts.append(len(pair_samples))
|
745 |
-
|
746 |
-
if not pair_counts:
|
747 |
-
return 0.0
|
748 |
-
|
749 |
-
avg_samples_per_pair = np.mean(pair_counts)
|
750 |
-
|
751 |
-
# Rough power estimation (0.8 power at 2x minimum, 0.95 at 4x minimum)
|
752 |
-
if avg_samples_per_pair >= min_required * 4:
|
753 |
-
return 0.95
|
754 |
-
elif avg_samples_per_pair >= min_required * 2:
|
755 |
-
return 0.8
|
756 |
-
elif avg_samples_per_pair >= min_required:
|
757 |
-
return 0.6
|
758 |
-
else:
|
759 |
-
return max(0.0, avg_samples_per_pair / min_required * 0.6)
|
760 |
-
|
761 |
-
|
762 |
-
def calculate_integrity_score(track_analysis: Dict, adequacy_report: Dict) -> float:
|
763 |
-
"""Calculate overall integrity score for the test set."""
|
764 |
-
|
765 |
-
if not track_analysis or not adequacy_report:
|
766 |
-
return 0.0
|
767 |
-
|
768 |
-
# Track adequacy scores
|
769 |
-
track_scores = []
|
770 |
-
for track_info in track_analysis.values():
|
771 |
-
adequacy_rate = track_info.get("adequacy_rate", 0.0)
|
772 |
-
alignment_rate = track_info.get("alignment_rate", 0.0)
|
773 |
-
track_score = (adequacy_rate + alignment_rate) / 2
|
774 |
-
track_scores.append(track_score)
|
775 |
-
|
776 |
-
# Overall adequacy mapping
|
777 |
-
adequacy_mapping = {
|
778 |
-
"excellent": 1.0,
|
779 |
-
"good": 0.8,
|
780 |
-
"fair": 0.6,
|
781 |
-
"insufficient": 0.2,
|
782 |
-
}
|
783 |
-
|
784 |
-
overall_adequacy_score = adequacy_mapping.get(
|
785 |
-
adequacy_report.get("overall_adequacy", "insufficient"), 0.2
|
786 |
-
)
|
787 |
-
|
788 |
-
# Combined score
|
789 |
-
if track_scores:
|
790 |
-
track_avg = np.mean(track_scores)
|
791 |
-
integrity_score = (track_avg + overall_adequacy_score) / 2
|
792 |
-
else:
|
793 |
-
integrity_score = overall_adequacy_score
|
794 |
-
|
795 |
-
return float(integrity_score)
|
|
|
12 |
ALL_UG40_LANGUAGES,
|
13 |
GOOGLE_SUPPORTED_LANGUAGES,
|
14 |
EVALUATION_TRACKS,
|
15 |
+
LANGUAGE_NAMES,
|
|
|
16 |
)
|
17 |
import salt.dataset
|
18 |
+
from src.utils import get_all_language_pairs
|
19 |
from typing import Dict, List, Optional, Tuple
|
20 |
|
21 |
|
|
|
22 |
# Local CSV filenames for persistence
|
23 |
+
LOCAL_PUBLIC_CSV = "salt_test_set.csv"
|
24 |
+
LOCAL_COMPLETE_CSV = "salt_complete_test_set.csv"
|
|
|
|
|
|
|
25 |
|
26 |
|
27 |
+
def generate_test_set(max_samples_per_pair: int = MAX_TEST_SAMPLES) -> pd.DataFrame:
|
28 |
+
"""Generate test set from SALT dataset."""
|
|
|
|
|
|
|
|
|
29 |
|
30 |
+
print("🔬 Generating SALT test set...")
|
31 |
|
32 |
try:
|
33 |
# Build SALT dataset config
|
|
|
54 |
test_samples = []
|
55 |
sample_id_counter = 1
|
56 |
|
57 |
+
# Generate samples for each language pair
|
|
|
|
|
|
|
58 |
for src_lang in ALL_UG40_LANGUAGES:
|
59 |
for tgt_lang in ALL_UG40_LANGUAGES:
|
60 |
if src_lang == tgt_lang:
|
61 |
continue
|
62 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
# Filter for this language pair
|
64 |
pair_data = full_data[
|
65 |
(full_data["source.language"] == src_lang) &
|
|
|
70 |
print(f"⚠️ No data found for {src_lang} → {tgt_lang}")
|
71 |
continue
|
72 |
|
73 |
+
# Sample data for this pair
|
74 |
+
n_samples = min(len(pair_data), max_samples_per_pair)
|
75 |
+
sampled = pair_data.sample(n=n_samples, random_state=42)
|
|
|
|
|
|
|
|
|
76 |
|
77 |
print(f"✅ {src_lang} → {tgt_lang}: {len(sampled)} samples")
|
78 |
|
79 |
for _, row in sampled.iterrows():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
80 |
test_samples.append({
|
81 |
"sample_id": f"salt_{sample_id_counter:06d}",
|
82 |
"source_text": row["source"],
|
|
|
88 |
src_lang in GOOGLE_SUPPORTED_LANGUAGES and
|
89 |
tgt_lang in GOOGLE_SUPPORTED_LANGUAGES
|
90 |
),
|
|
|
|
|
|
|
|
|
91 |
})
|
92 |
sample_id_counter += 1
|
93 |
|
|
|
96 |
if test_df.empty:
|
97 |
raise ValueError("No test samples generated - check SALT dataset availability")
|
98 |
|
99 |
+
print(f"✅ Generated test set: {len(test_df):,} samples")
|
|
|
|
|
|
|
|
|
100 |
|
101 |
return test_df
|
102 |
|
103 |
except Exception as e:
|
104 |
+
print(f"❌ Error generating test set: {e}")
|
105 |
return pd.DataFrame(columns=[
|
106 |
"sample_id", "source_text", "target_text", "source_language",
|
107 |
+
"target_language", "domain", "google_comparable"
|
|
|
108 |
])
|
109 |
|
110 |
|
111 |
+
def _generate_and_save_test_set() -> Tuple[pd.DataFrame, pd.DataFrame]:
|
112 |
+
"""Generate and save both public and complete versions of the test set."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
113 |
|
114 |
+
print("🔬 Generating and saving test sets...")
|
|
|
115 |
|
116 |
+
full_df = generate_test_set()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
117 |
|
118 |
if full_df.empty:
|
119 |
+
print("❌ Failed to generate test set")
|
120 |
empty_public = pd.DataFrame(columns=[
|
121 |
"sample_id", "source_text", "source_language",
|
122 |
+
"target_language", "domain", "google_comparable"
|
|
|
123 |
])
|
124 |
empty_complete = pd.DataFrame(columns=[
|
125 |
"sample_id", "source_text", "target_text", "source_language",
|
126 |
+
"target_language", "domain", "google_comparable"
|
|
|
127 |
])
|
128 |
return empty_public, empty_complete
|
129 |
|
130 |
# Public version (no target_text)
|
131 |
public_df = full_df[[
|
132 |
"sample_id", "source_text", "source_language",
|
133 |
+
"target_language", "domain", "google_comparable"
|
|
|
134 |
]].copy()
|
135 |
|
136 |
+
# Save versions
|
137 |
try:
|
138 |
public_df.to_csv(LOCAL_PUBLIC_CSV, index=False)
|
139 |
full_df.to_csv(LOCAL_COMPLETE_CSV, index=False)
|
140 |
+
print(f"✅ Saved test sets: {LOCAL_PUBLIC_CSV}, {LOCAL_COMPLETE_CSV}")
|
141 |
except Exception as e:
|
142 |
+
print(f"⚠️ Error saving CSVs: {e}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
143 |
|
144 |
return public_df, full_df
|
145 |
|
146 |
|
147 |
+
def get_public_test_set() -> pd.DataFrame:
|
148 |
+
"""Load the public test set with enhanced fallback logic."""
|
149 |
|
150 |
# 1) Try HF Hub
|
151 |
try:
|
152 |
+
print("📥 Attempting to load test set from HF Hub...")
|
153 |
+
ds = load_dataset(TEST_SET_DATASET, split="train", token=HF_TOKEN)
|
154 |
df = ds.to_pandas()
|
155 |
|
156 |
+
# Validate structure
|
157 |
+
required_cols = ["sample_id", "source_text", "source_language", "target_language"]
|
|
|
158 |
if all(col in df.columns for col in required_cols):
|
159 |
+
print(f"✅ Loaded test set from HF Hub ({len(df):,} samples)")
|
160 |
return df
|
161 |
else:
|
162 |
+
print("⚠️ HF Hub test set missing columns, regenerating...")
|
163 |
|
164 |
except Exception as e:
|
165 |
print(f"⚠️ HF Hub load failed: {e}")
|
|
|
170 |
df = pd.read_csv(LOCAL_PUBLIC_CSV)
|
171 |
required_cols = ["sample_id", "source_text", "source_language", "target_language"]
|
172 |
if all(col in df.columns for col in required_cols):
|
173 |
+
print(f"✅ Loaded test set from local CSV ({len(df):,} samples)")
|
174 |
return df
|
175 |
else:
|
176 |
print("⚠️ Local CSV has invalid structure, regenerating...")
|
177 |
except Exception as e:
|
178 |
+
print(f"⚠️ Failed to read local CSV: {e}")
|
179 |
|
180 |
# 3) Regenerate & save
|
181 |
+
print("🔄 Generating new test set...")
|
182 |
+
public_df, _ = _generate_and_save_test_set()
|
183 |
return public_df
|
184 |
|
185 |
|
186 |
+
def get_complete_test_set() -> pd.DataFrame:
|
187 |
+
"""Load the complete test set with targets."""
|
188 |
|
189 |
# 1) Try HF Hub private
|
190 |
try:
|
191 |
+
print("📥 Attempting to load complete test set from HF Hub...")
|
192 |
+
ds = load_dataset(TEST_SET_DATASET + "-private", split="train", token=HF_TOKEN)
|
193 |
df = ds.to_pandas()
|
194 |
|
195 |
+
required_cols = ["sample_id", "source_text", "target_text", "source_language", "target_language"]
|
|
|
196 |
if all(col in df.columns for col in required_cols):
|
197 |
+
print(f"✅ Loaded complete test set from HF Hub ({len(df):,} samples)")
|
198 |
return df
|
199 |
else:
|
200 |
+
print("⚠️ HF Hub complete test set missing columns, regenerating...")
|
201 |
|
202 |
except Exception as e:
|
203 |
print(f"⚠️ HF Hub private load failed: {e}")
|
|
|
208 |
df = pd.read_csv(LOCAL_COMPLETE_CSV)
|
209 |
required_cols = ["sample_id", "source_text", "target_text", "source_language", "target_language"]
|
210 |
if all(col in df.columns for col in required_cols):
|
211 |
+
print(f"✅ Loaded complete test set from local CSV ({len(df):,} samples)")
|
212 |
return df
|
213 |
else:
|
214 |
print("⚠️ Local complete CSV has invalid structure, regenerating...")
|
215 |
except Exception as e:
|
216 |
+
print(f"⚠️ Failed to read local complete CSV: {e}")
|
217 |
|
218 |
# 3) Regenerate & save
|
219 |
+
print("🔄 Generating new complete test set...")
|
220 |
+
_, complete_df = _generate_and_save_test_set()
|
221 |
return complete_df
|
222 |
|
223 |
|
224 |
+
def create_test_set_download() -> Tuple[str, Dict]:
|
225 |
+
"""Create test set download with comprehensive metadata."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
226 |
|
227 |
+
public_df = get_public_test_set()
|
228 |
|
229 |
if public_df.empty:
|
230 |
stats = {
|
231 |
"total_samples": 0,
|
232 |
"track_breakdown": {},
|
233 |
+
"languages": [],
|
234 |
+
"language_pairs": 0,
|
235 |
+
"google_comparable_samples": 0,
|
236 |
}
|
237 |
return LOCAL_PUBLIC_CSV, stats
|
238 |
|
|
|
242 |
try:
|
243 |
public_df.to_csv(download_path, index=False)
|
244 |
except Exception as e:
|
245 |
+
print(f"⚠️ Error updating CSV: {e}")
|
246 |
|
247 |
# Calculate comprehensive statistics
|
248 |
try:
|
|
|
250 |
stats = {
|
251 |
"total_samples": len(public_df),
|
252 |
"languages": sorted(list(set(public_df["source_language"]).union(public_df["target_language"]))),
|
253 |
+
"language_pairs": len(public_df.groupby(["source_language", "target_language"])),
|
254 |
}
|
255 |
|
256 |
# Track-specific breakdown
|
|
|
263 |
]
|
264 |
|
265 |
track_breakdown[track_name] = {
|
|
|
266 |
"total_samples": len(track_data),
|
267 |
"language_pairs": len(track_data.groupby(["source_language", "target_language"])),
|
268 |
+
"languages": track_languages,
|
|
|
269 |
}
|
270 |
|
271 |
stats["track_breakdown"] = track_breakdown
|
|
|
273 |
# Google-comparable statistics
|
274 |
if "google_comparable" in public_df.columns:
|
275 |
stats["google_comparable_samples"] = int(public_df["google_comparable"].sum())
|
|
|
276 |
else:
|
277 |
stats["google_comparable_samples"] = 0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
278 |
|
279 |
except Exception as e:
|
280 |
+
print(f"⚠️ Error calculating stats: {e}")
|
281 |
stats = {
|
282 |
"total_samples": len(public_df),
|
283 |
"track_breakdown": {},
|
284 |
+
"languages": [],
|
285 |
+
"language_pairs": 0,
|
286 |
+
"google_comparable_samples": 0,
|
287 |
}
|
288 |
|
289 |
return download_path, stats
|
290 |
|
291 |
|
292 |
+
def get_track_test_set(track: str) -> pd.DataFrame:
|
293 |
+
"""Get test set filtered for a specific track."""
|
294 |
+
|
295 |
+
if track not in EVALUATION_TRACKS:
|
296 |
+
print(f"❌ Unknown track: {track}")
|
297 |
+
return pd.DataFrame()
|
298 |
+
|
299 |
+
# Get main test set and filter
|
300 |
+
public_df = get_public_test_set()
|
301 |
+
|
302 |
+
if public_df.empty:
|
303 |
+
return pd.DataFrame()
|
304 |
+
|
305 |
+
track_languages = EVALUATION_TRACKS[track]["languages"]
|
306 |
+
track_df = public_df[
|
307 |
+
(public_df["source_language"].isin(track_languages)) &
|
308 |
+
(public_df["target_language"].isin(track_languages))
|
309 |
+
]
|
310 |
+
|
311 |
+
print(f"✅ Filtered {track} test set: {len(track_df):,} samples")
|
312 |
+
return track_df
|
313 |
+
|
314 |
+
|
315 |
+
def validate_test_set_integrity() -> Dict:
|
316 |
+
"""Validate test set integrity."""
|
317 |
|
318 |
try:
|
319 |
+
public_df = get_public_test_set()
|
320 |
+
complete_df = get_complete_test_set()
|
321 |
|
322 |
if public_df.empty or complete_df.empty:
|
323 |
return {
|
324 |
"alignment_check": False,
|
325 |
"total_samples": 0,
|
|
|
326 |
"track_analysis": {},
|
327 |
"error": "Test sets are empty or could not be loaded",
|
328 |
}
|
|
|
334 |
track_analysis = {}
|
335 |
for track_name, track_config in EVALUATION_TRACKS.items():
|
336 |
track_languages = track_config["languages"]
|
|
|
337 |
|
338 |
# Analyze public set for this track
|
339 |
track_public = public_df[
|
|
|
347 |
(complete_df["target_language"].isin(track_languages))
|
348 |
]
|
349 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
350 |
track_analysis[track_name] = {
|
351 |
+
"public_samples": len(track_public),
|
352 |
+
"complete_samples": len(track_complete),
|
353 |
+
"alignment": len(track_public) == len(track_complete),
|
354 |
+
"languages": track_languages,
|
|
|
|
|
|
|
355 |
}
|
356 |
|
|
|
|
|
|
|
357 |
return {
|
358 |
"alignment_check": public_ids <= private_ids,
|
359 |
"total_samples": len(public_df),
|
360 |
"track_analysis": track_analysis,
|
|
|
361 |
"public_samples": len(public_df),
|
362 |
"private_samples": len(complete_df),
|
363 |
"id_alignment_rate": len(public_ids & private_ids) / len(public_ids) if public_ids else 0.0,
|
|
|
364 |
}
|
365 |
|
366 |
except Exception as e:
|
367 |
return {
|
368 |
"alignment_check": False,
|
369 |
"total_samples": 0,
|
|
|
370 |
"track_analysis": {},
|
371 |
"error": f"Validation failed: {str(e)}",
|
372 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|