Commit
·
4a87d90
1
Parent(s):
fe440e6
feat: renaming output files and fixing a bug
Browse files- crag_sampler/sampler.py +18 -18
- crag_sampler/utils.py +56 -64
- examples/basic_sampling.py +18 -16
crag_sampler/sampler.py
CHANGED
@@ -13,8 +13,8 @@ import subprocess
|
|
13 |
from .utils import (
|
14 |
read_jsonl_fields_fast,
|
15 |
process_answer_types,
|
16 |
-
|
17 |
-
|
18 |
)
|
19 |
|
20 |
|
@@ -51,23 +51,23 @@ class CragSampler:
|
|
51 |
)
|
52 |
return process_answer_types(df)
|
53 |
|
54 |
-
def
|
55 |
self,
|
56 |
-
|
57 |
stratify_columns: Optional[List[str]] = None,
|
58 |
output_path: Optional[str] = None,
|
59 |
force_compute: bool = False,
|
60 |
) -> Dict:
|
61 |
-
"""Create stratified
|
62 |
|
63 |
Args:
|
64 |
-
|
65 |
stratify_columns: Columns to use for stratification. If None, uses defaults
|
66 |
output_path: Path to save/load the JSON output
|
67 |
-
force_compute: If True, always compute
|
68 |
|
69 |
Returns:
|
70 |
-
Dictionary containing the
|
71 |
"""
|
72 |
if stratify_columns is None:
|
73 |
stratify_columns = [
|
@@ -80,37 +80,37 @@ class CragSampler:
|
|
80 |
if output_path is None:
|
81 |
output_path = os.path.join(
|
82 |
os.path.dirname(self.input_file),
|
83 |
-
f"{os.path.splitext(os.path.basename(self.input_file))[0]}
|
84 |
)
|
85 |
|
86 |
-
return
|
87 |
self.df,
|
88 |
-
|
89 |
stratify_columns=stratify_columns,
|
90 |
output_path=output_path,
|
91 |
force_compute=force_compute,
|
92 |
)
|
93 |
|
94 |
-
def
|
95 |
self,
|
96 |
-
|
97 |
output_dir: Optional[str] = None,
|
98 |
compress: bool = True,
|
99 |
n_processes: Optional[int] = None,
|
100 |
overwrite: bool = False,
|
101 |
) -> None:
|
102 |
-
"""Write
|
103 |
|
104 |
Args:
|
105 |
-
|
106 |
-
output_dir: Directory to save
|
107 |
compress: Whether to compress output files with bz2
|
108 |
n_processes: Number of processes to use
|
109 |
overwrite: If False, skip existing output files
|
110 |
"""
|
111 |
-
|
112 |
self.input_file,
|
113 |
-
|
114 |
output_dir=output_dir,
|
115 |
compress=compress,
|
116 |
n_processes=n_processes,
|
|
|
13 |
from .utils import (
|
14 |
read_jsonl_fields_fast,
|
15 |
process_answer_types,
|
16 |
+
create_stratified_subsets,
|
17 |
+
subset_jsonl_file,
|
18 |
)
|
19 |
|
20 |
|
|
|
51 |
)
|
52 |
return process_answer_types(df)
|
53 |
|
54 |
+
def create_subsets(
|
55 |
self,
|
56 |
+
n_subsets: int = 5,
|
57 |
stratify_columns: Optional[List[str]] = None,
|
58 |
output_path: Optional[str] = None,
|
59 |
force_compute: bool = False,
|
60 |
) -> Dict:
|
61 |
+
"""Create stratified subsets of the dataset.
|
62 |
|
63 |
Args:
|
64 |
+
n_subsets: Number of subsets to create
|
65 |
stratify_columns: Columns to use for stratification. If None, uses defaults
|
66 |
output_path: Path to save/load the JSON output
|
67 |
+
force_compute: If True, always compute subsets even if file exists
|
68 |
|
69 |
Returns:
|
70 |
+
Dictionary containing the subsets information
|
71 |
"""
|
72 |
if stratify_columns is None:
|
73 |
stratify_columns = [
|
|
|
80 |
if output_path is None:
|
81 |
output_path = os.path.join(
|
82 |
os.path.dirname(self.input_file),
|
83 |
+
f"{os.path.splitext(os.path.basename(self.input_file))[0]}_subsets.json",
|
84 |
)
|
85 |
|
86 |
+
return create_stratified_subsets(
|
87 |
self.df,
|
88 |
+
n_subsets=n_subsets,
|
89 |
stratify_columns=stratify_columns,
|
90 |
output_path=output_path,
|
91 |
force_compute=force_compute,
|
92 |
)
|
93 |
|
94 |
+
def write_subsets(
|
95 |
self,
|
96 |
+
subsets_file: str,
|
97 |
output_dir: Optional[str] = None,
|
98 |
compress: bool = True,
|
99 |
n_processes: Optional[int] = None,
|
100 |
overwrite: bool = False,
|
101 |
) -> None:
|
102 |
+
"""Write subsets to separate files.
|
103 |
|
104 |
Args:
|
105 |
+
subsets_file: Path to JSON file containing subset indices
|
106 |
+
output_dir: Directory to save subset files
|
107 |
compress: Whether to compress output files with bz2
|
108 |
n_processes: Number of processes to use
|
109 |
overwrite: If False, skip existing output files
|
110 |
"""
|
111 |
+
subset_jsonl_file(
|
112 |
self.input_file,
|
113 |
+
subsets_file,
|
114 |
output_dir=output_dir,
|
115 |
compress=compress,
|
116 |
n_processes=n_processes,
|
crag_sampler/utils.py
CHANGED
@@ -167,82 +167,80 @@ def process_answer_types(df: pd.DataFrame) -> pd.DataFrame:
|
|
167 |
return df
|
168 |
|
169 |
|
170 |
-
def
|
171 |
df: pd.DataFrame,
|
172 |
-
|
173 |
stratify_columns: List[str] = [
|
174 |
"domain",
|
175 |
"answer_type",
|
176 |
"question_type",
|
177 |
"static_or_dynamic",
|
178 |
],
|
179 |
-
output_path: str = "
|
180 |
force_compute: bool = False,
|
181 |
) -> Dict[str, Any]:
|
182 |
"""
|
183 |
-
Create stratified
|
184 |
-
Each
|
185 |
|
186 |
Args:
|
187 |
df: Input DataFrame
|
188 |
-
|
189 |
stratify_columns: Columns to use for stratification
|
190 |
output_path: Path to save/load the JSON output
|
191 |
-
force_compute: If True, always compute
|
192 |
|
193 |
Returns:
|
194 |
-
Dictionary containing the
|
195 |
"""
|
196 |
# Check if file exists and we can use it
|
197 |
if not force_compute and os.path.exists(output_path):
|
198 |
try:
|
199 |
with open(output_path, "r") as f:
|
200 |
-
|
201 |
|
202 |
# Validate the loaded data has the expected structure
|
203 |
if (
|
204 |
-
|
205 |
-
and
|
206 |
== stratify_columns
|
207 |
):
|
208 |
-
print(f"Loading existing
|
209 |
-
return
|
210 |
else:
|
211 |
-
print(
|
212 |
-
"Existing subsamples file has different parameters, recomputing..."
|
213 |
-
)
|
214 |
except Exception as e:
|
215 |
-
print(f"Error loading existing
|
216 |
|
217 |
# Create a combined category for stratification
|
218 |
df["strat_category"] = df[stratify_columns].astype(str).agg("_".join, axis=1)
|
219 |
|
220 |
-
# Initialize the
|
221 |
-
skf = StratifiedKFold(n_splits=
|
222 |
|
223 |
-
# Create
|
224 |
-
|
225 |
-
for
|
226 |
skf.split(df, df["strat_category"])
|
227 |
):
|
228 |
# Sort indices for consistent hashing
|
229 |
-
sorted_indices = sorted(
|
230 |
|
231 |
# Create a deterministic ID from the indices
|
232 |
-
|
233 |
|
234 |
-
# Calculate statistics for this
|
235 |
stats = {}
|
236 |
-
|
237 |
for col in stratify_columns:
|
238 |
-
stats[col] =
|
239 |
|
240 |
-
|
241 |
{
|
242 |
-
"
|
243 |
"statistics": stats,
|
244 |
"indices": sorted_indices,
|
245 |
-
"size": len(
|
246 |
}
|
247 |
)
|
248 |
|
@@ -253,12 +251,12 @@ def create_stratified_subsamples(
|
|
253 |
|
254 |
output_data = {
|
255 |
"metadata": {
|
256 |
-
"
|
257 |
"total_samples": len(df),
|
258 |
"stratify_columns": stratify_columns,
|
259 |
"global_statistics": global_stats,
|
260 |
},
|
261 |
-
"
|
262 |
}
|
263 |
|
264 |
# Save to JSON
|
@@ -268,11 +266,10 @@ def create_stratified_subsamples(
|
|
268 |
return output_data
|
269 |
|
270 |
|
271 |
-
def
|
272 |
input_file: str, indices: List[int], output_file: str, compress: bool = True
|
273 |
) -> None:
|
274 |
-
"""
|
275 |
-
Write a single subsample to a file using awk.
|
276 |
|
277 |
Args:
|
278 |
input_file: Path to input JSONL file
|
@@ -280,15 +277,10 @@ def write_subsample(
|
|
280 |
output_file: Path to output file
|
281 |
compress: Whether to compress output
|
282 |
"""
|
283 |
-
# Convert indices to
|
284 |
-
# NR is the current line number in awk
|
285 |
indices_set = set(i + 1 for i in indices) # Convert to 1-based indexing
|
286 |
-
|
287 |
-
|
288 |
-
# Create awk script with escaped curly braces
|
289 |
-
awk_script = (
|
290 |
-
f'BEGIN {{subsample("{indices_str}",a,","); for(i in a) n[a[i]];}} NR in n'
|
291 |
-
)
|
292 |
|
293 |
if input_file.endswith(".bz2"):
|
294 |
if compress:
|
@@ -301,7 +293,7 @@ def write_subsample(
|
|
301 |
else:
|
302 |
cmd = f"awk '{awk_script}' '{input_file}' > '{output_file}'"
|
303 |
|
304 |
-
print(f"Process {os.getpid()} - Starting
|
305 |
try:
|
306 |
result = subprocess.run(
|
307 |
cmd,
|
@@ -311,12 +303,12 @@ def write_subsample(
|
|
311 |
stdout=subprocess.PIPE,
|
312 |
text=True,
|
313 |
)
|
314 |
-
print(f"Process {os.getpid()} - Finished
|
315 |
|
316 |
-
# Verify the output file exists and has content
|
317 |
if os.path.exists(output_file) and os.path.getsize(output_file) > 0:
|
318 |
print(
|
319 |
-
f"Process {os.getpid()} - Successfully created {output_file}
|
|
|
320 |
)
|
321 |
else:
|
322 |
raise Exception(f"Output file {output_file} is empty or doesn't exist")
|
@@ -330,33 +322,33 @@ def write_subsample(
|
|
330 |
raise
|
331 |
|
332 |
|
333 |
-
def
|
334 |
input_file: str,
|
335 |
-
|
336 |
output_dir: Optional[str] = None,
|
337 |
compress: bool = True,
|
338 |
n_processes: Optional[int] = None,
|
339 |
overwrite: bool = False,
|
340 |
) -> None:
|
341 |
"""
|
342 |
-
|
343 |
|
344 |
Args:
|
345 |
input_file: Path to input JSONL file (can be bz2 compressed)
|
346 |
-
|
347 |
-
output_dir: Directory to save
|
348 |
compress: Whether to compress output files with bz2
|
349 |
-
n_processes: Number of processes to use (defaults to min(
|
350 |
overwrite: If False, skip existing output files (default: False)
|
351 |
"""
|
352 |
-
# Load
|
353 |
-
with open(
|
354 |
-
|
355 |
|
356 |
# Determine optimal number of processes
|
357 |
-
|
358 |
if n_processes is None:
|
359 |
-
n_processes = min(
|
360 |
|
361 |
if output_dir is None:
|
362 |
output_dir = os.path.dirname(input_file)
|
@@ -369,9 +361,9 @@ def subsample_jsonl_file(
|
|
369 |
# Prepare arguments for parallel processing
|
370 |
write_args = []
|
371 |
skipped_files = []
|
372 |
-
for
|
373 |
-
|
374 |
-
output_name = f"{base_name}
|
375 |
if compress:
|
376 |
output_name += ".bz2"
|
377 |
output_path = os.path.join(output_dir, output_name)
|
@@ -381,7 +373,7 @@ def subsample_jsonl_file(
|
|
381 |
skipped_files.append(output_path)
|
382 |
continue
|
383 |
|
384 |
-
write_args.append((input_file,
|
385 |
|
386 |
if skipped_files:
|
387 |
print(f"Skipping {len(skipped_files)} existing files:")
|
@@ -389,8 +381,8 @@ def subsample_jsonl_file(
|
|
389 |
print(f" - {file}")
|
390 |
|
391 |
if write_args:
|
392 |
-
print(f"Processing {len(write_args)}
|
393 |
with Pool(processes=n_processes) as pool:
|
394 |
-
pool.starmap(
|
395 |
else:
|
396 |
print("No files to process - all files exist and overwrite=False")
|
|
|
167 |
return df
|
168 |
|
169 |
|
170 |
+
def create_stratified_subsets(
|
171 |
df: pd.DataFrame,
|
172 |
+
n_subsets: int,
|
173 |
stratify_columns: List[str] = [
|
174 |
"domain",
|
175 |
"answer_type",
|
176 |
"question_type",
|
177 |
"static_or_dynamic",
|
178 |
],
|
179 |
+
output_path: str = "subsets.json",
|
180 |
force_compute: bool = False,
|
181 |
) -> Dict[str, Any]:
|
182 |
"""
|
183 |
+
Create stratified subsets of the dataset and save them to a JSON file.
|
184 |
+
Each subset gets a unique ID based on its indices.
|
185 |
|
186 |
Args:
|
187 |
df: Input DataFrame
|
188 |
+
n_subsets: Number of subsets to create
|
189 |
stratify_columns: Columns to use for stratification
|
190 |
output_path: Path to save/load the JSON output
|
191 |
+
force_compute: If True, always compute subsets even if file exists
|
192 |
|
193 |
Returns:
|
194 |
+
Dictionary containing the subsets information
|
195 |
"""
|
196 |
# Check if file exists and we can use it
|
197 |
if not force_compute and os.path.exists(output_path):
|
198 |
try:
|
199 |
with open(output_path, "r") as f:
|
200 |
+
subsets_data = json.load(f)
|
201 |
|
202 |
# Validate the loaded data has the expected structure
|
203 |
if (
|
204 |
+
subsets_data.get("metadata", {}).get("n_subsets") == n_subsets
|
205 |
+
and subsets_data.get("metadata", {}).get("stratify_columns")
|
206 |
== stratify_columns
|
207 |
):
|
208 |
+
print(f"Loading existing subsets from {output_path}")
|
209 |
+
return subsets_data
|
210 |
else:
|
211 |
+
print("Existing subsets file has different parameters, recomputing...")
|
|
|
|
|
212 |
except Exception as e:
|
213 |
+
print(f"Error loading existing subsets file: {e}, recomputing...")
|
214 |
|
215 |
# Create a combined category for stratification
|
216 |
df["strat_category"] = df[stratify_columns].astype(str).agg("_".join, axis=1)
|
217 |
|
218 |
+
# Initialize the subsetter
|
219 |
+
skf = StratifiedKFold(n_splits=n_subsets, shuffle=True, random_state=42)
|
220 |
|
221 |
+
# Create subsets
|
222 |
+
subsets_info = []
|
223 |
+
for subset_idx, (_, subset_indices) in enumerate(
|
224 |
skf.split(df, df["strat_category"])
|
225 |
):
|
226 |
# Sort indices for consistent hashing
|
227 |
+
sorted_indices = sorted(subset_indices.tolist())
|
228 |
|
229 |
# Create a deterministic ID from the indices
|
230 |
+
subset_id = hashlib.md5(str(sorted_indices).encode()).hexdigest()[:8]
|
231 |
|
232 |
+
# Calculate statistics for this subset
|
233 |
stats = {}
|
234 |
+
subset_df = df.iloc[subset_indices]
|
235 |
for col in stratify_columns:
|
236 |
+
stats[col] = subset_df[col].value_counts().to_dict()
|
237 |
|
238 |
+
subsets_info.append(
|
239 |
{
|
240 |
+
"index": subset_idx,
|
241 |
"statistics": stats,
|
242 |
"indices": sorted_indices,
|
243 |
+
"size": len(subset_indices),
|
244 |
}
|
245 |
)
|
246 |
|
|
|
251 |
|
252 |
output_data = {
|
253 |
"metadata": {
|
254 |
+
"n_subsets": n_subsets,
|
255 |
"total_samples": len(df),
|
256 |
"stratify_columns": stratify_columns,
|
257 |
"global_statistics": global_stats,
|
258 |
},
|
259 |
+
"subsets": subsets_info,
|
260 |
}
|
261 |
|
262 |
# Save to JSON
|
|
|
266 |
return output_data
|
267 |
|
268 |
|
269 |
+
def write_subset(
|
270 |
input_file: str, indices: List[int], output_file: str, compress: bool = True
|
271 |
) -> None:
|
272 |
+
"""Write a single subset to a file using awk.
|
|
|
273 |
|
274 |
Args:
|
275 |
input_file: Path to input JSONL file
|
|
|
277 |
output_file: Path to output file
|
278 |
compress: Whether to compress output
|
279 |
"""
|
280 |
+
# Convert indices to 1-based indexing and create NR condition
|
|
|
281 |
indices_set = set(i + 1 for i in indices) # Convert to 1-based indexing
|
282 |
+
nr_conditions = " || ".join(f"NR == {i}" for i in sorted(indices_set))
|
283 |
+
awk_script = f"{nr_conditions}"
|
|
|
|
|
|
|
|
|
284 |
|
285 |
if input_file.endswith(".bz2"):
|
286 |
if compress:
|
|
|
293 |
else:
|
294 |
cmd = f"awk '{awk_script}' '{input_file}' > '{output_file}'"
|
295 |
|
296 |
+
print(f"Process {os.getpid()} - Starting subset to {output_file}")
|
297 |
try:
|
298 |
result = subprocess.run(
|
299 |
cmd,
|
|
|
303 |
stdout=subprocess.PIPE,
|
304 |
text=True,
|
305 |
)
|
306 |
+
print(f"Process {os.getpid()} - Finished subset to {output_file}")
|
307 |
|
|
|
308 |
if os.path.exists(output_file) and os.path.getsize(output_file) > 0:
|
309 |
print(
|
310 |
+
f"Process {os.getpid()} - Successfully created {output_file} "
|
311 |
+
f"({os.path.getsize(output_file)} bytes)"
|
312 |
)
|
313 |
else:
|
314 |
raise Exception(f"Output file {output_file} is empty or doesn't exist")
|
|
|
322 |
raise
|
323 |
|
324 |
|
325 |
+
def subset_jsonl_file(
|
326 |
input_file: str,
|
327 |
+
subsets_file: str,
|
328 |
output_dir: Optional[str] = None,
|
329 |
compress: bool = True,
|
330 |
n_processes: Optional[int] = None,
|
331 |
overwrite: bool = False,
|
332 |
) -> None:
|
333 |
"""
|
334 |
+
subset a large JSONL file into multiple files using sed for maximum performance.
|
335 |
|
336 |
Args:
|
337 |
input_file: Path to input JSONL file (can be bz2 compressed)
|
338 |
+
subsets_file: Path to JSON file containing subset indices
|
339 |
+
output_dir: Directory to save subset files (defaults to input file directory)
|
340 |
compress: Whether to compress output files with bz2
|
341 |
+
n_processes: Number of processes to use (defaults to min(n_subsets, cpu_count))
|
342 |
overwrite: If False, skip existing output files (default: False)
|
343 |
"""
|
344 |
+
# Load subsets information
|
345 |
+
with open(subsets_file, "r") as f:
|
346 |
+
subsets_data = json.load(f)
|
347 |
|
348 |
# Determine optimal number of processes
|
349 |
+
n_subsets = len(subsets_data["subsets"])
|
350 |
if n_processes is None:
|
351 |
+
n_processes = min(n_subsets, cpu_count())
|
352 |
|
353 |
if output_dir is None:
|
354 |
output_dir = os.path.dirname(input_file)
|
|
|
361 |
# Prepare arguments for parallel processing
|
362 |
write_args = []
|
363 |
skipped_files = []
|
364 |
+
for subset in subsets_data["subsets"]:
|
365 |
+
subset_idx = subset["index"]
|
366 |
+
output_name = f"{base_name}_subset_{subset_idx+1}.jsonl"
|
367 |
if compress:
|
368 |
output_name += ".bz2"
|
369 |
output_path = os.path.join(output_dir, output_name)
|
|
|
373 |
skipped_files.append(output_path)
|
374 |
continue
|
375 |
|
376 |
+
write_args.append((input_file, subset["indices"], output_path, compress))
|
377 |
|
378 |
if skipped_files:
|
379 |
print(f"Skipping {len(skipped_files)} existing files:")
|
|
|
381 |
print(f" - {file}")
|
382 |
|
383 |
if write_args:
|
384 |
+
print(f"Processing {len(write_args)} subsets using {n_processes} processes")
|
385 |
with Pool(processes=n_processes) as pool:
|
386 |
+
pool.starmap(write_subset, write_args)
|
387 |
else:
|
388 |
print("No files to process - all files exist and overwrite=False")
|
examples/basic_sampling.py
CHANGED
@@ -7,7 +7,7 @@ import os
|
|
7 |
def run_crag_task_1_and_2(
|
8 |
file_path: str,
|
9 |
fields_to_extract: list[str] = None,
|
10 |
-
|
11 |
output_dir: str = None,
|
12 |
compress: bool = True,
|
13 |
n_processes: int = None,
|
@@ -18,7 +18,7 @@ def run_crag_task_1_and_2(
|
|
18 |
Args:
|
19 |
file_path: Path to input JSONL file
|
20 |
fields_to_extract: List of fields to extract from JSONL
|
21 |
-
|
22 |
output_dir: Directory for output files
|
23 |
compress: Whether to compress output files
|
24 |
n_processes: Number of processes for parallel processing
|
@@ -29,27 +29,25 @@ def run_crag_task_1_and_2(
|
|
29 |
input_file=file_path, required_fields=fields_to_extract, use_cache=True
|
30 |
)
|
31 |
|
32 |
-
# Create output path for
|
33 |
output_path = os.path.join(
|
34 |
os.path.dirname(file_path),
|
35 |
-
f"{os.path.splitext(os.path.basename(file_path))[0]}
|
36 |
)
|
37 |
|
38 |
-
# Create
|
39 |
-
|
40 |
-
n_subsamples=n_subsamples, output_path=output_path
|
41 |
-
)
|
42 |
|
43 |
# Print statistics
|
44 |
-
print(f"Created {
|
45 |
print("\nGlobal statistics:")
|
46 |
-
print(json.dumps(
|
47 |
-
print("\nFirst
|
48 |
-
print(json.dumps(
|
49 |
|
50 |
-
# Write
|
51 |
-
sampler.
|
52 |
-
|
53 |
output_dir=output_dir,
|
54 |
compress=compress,
|
55 |
n_processes=n_processes,
|
@@ -61,5 +59,9 @@ def run_crag_task_1_and_2(
|
|
61 |
if __name__ == "__main__":
|
62 |
file_path = "./local_data/crag_task_1_and_2_dev_v4.jsonl.bz2"
|
63 |
fields_to_extract = ["domain", "answer", "question_type", "static_or_dynamic"]
|
|
|
|
|
64 |
|
65 |
-
run_crag_task_1_and_2(
|
|
|
|
|
|
7 |
def run_crag_task_1_and_2(
|
8 |
file_path: str,
|
9 |
fields_to_extract: list[str] = None,
|
10 |
+
n_subsets: int = 5,
|
11 |
output_dir: str = None,
|
12 |
compress: bool = True,
|
13 |
n_processes: int = None,
|
|
|
18 |
Args:
|
19 |
file_path: Path to input JSONL file
|
20 |
fields_to_extract: List of fields to extract from JSONL
|
21 |
+
n_subsets: Number of subsets to create
|
22 |
output_dir: Directory for output files
|
23 |
compress: Whether to compress output files
|
24 |
n_processes: Number of processes for parallel processing
|
|
|
29 |
input_file=file_path, required_fields=fields_to_extract, use_cache=True
|
30 |
)
|
31 |
|
32 |
+
# Create output path for subsets
|
33 |
output_path = os.path.join(
|
34 |
os.path.dirname(file_path),
|
35 |
+
f"{os.path.splitext(os.path.basename(file_path))[0]}_subsets.json",
|
36 |
)
|
37 |
|
38 |
+
# Create subsets
|
39 |
+
subsets_data = sampler.create_subsets(n_subsets=n_subsets, output_path=output_path)
|
|
|
|
|
40 |
|
41 |
# Print statistics
|
42 |
+
print(f"Created {subsets_data['metadata']['n_subsets']} subsets")
|
43 |
print("\nGlobal statistics:")
|
44 |
+
print(json.dumps(subsets_data["metadata"]["global_statistics"], indent=2))
|
45 |
+
print("\nFirst subset statistics:")
|
46 |
+
print(json.dumps(subsets_data["subsets"][0]["statistics"], indent=2))
|
47 |
|
48 |
+
# Write subsets to files
|
49 |
+
sampler.write_subsets(
|
50 |
+
subsets_file=output_path,
|
51 |
output_dir=output_dir,
|
52 |
compress=compress,
|
53 |
n_processes=n_processes,
|
|
|
59 |
if __name__ == "__main__":
|
60 |
file_path = "./local_data/crag_task_1_and_2_dev_v4.jsonl.bz2"
|
61 |
fields_to_extract = ["domain", "answer", "question_type", "static_or_dynamic"]
|
62 |
+
n_subsets = 20
|
63 |
+
output_dir = "./subset/crag_task_1_and_2"
|
64 |
|
65 |
+
run_crag_task_1_and_2(
|
66 |
+
file_path, fields_to_extract, n_subsets=n_subsets, overwrite=True
|
67 |
+
)
|