Upload .\scripts\quick_process.py with huggingface_hub
Browse files- .//scripts//quick_process.py +41 -0
.//scripts//quick_process.py
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Quick processing test."""
|
| 2 |
+
import sys
|
| 3 |
+
import json
|
| 4 |
+
import shutil
|
| 5 |
+
from pathlib import Path
|
| 6 |
+
|
| 7 |
+
PROJECT_ROOT = Path(__file__).resolve().parent.parent
|
| 8 |
+
sys.path.insert(0, str(PROJECT_ROOT / "src"))
|
| 9 |
+
from data_collection.data_processor import DataProcessor
|
| 10 |
+
|
| 11 |
+
raw_dir = str(PROJECT_ROOT / "data/raw_test")
|
| 12 |
+
processed_dir = str(PROJECT_ROOT / "data/processed_test")
|
| 13 |
+
Path(processed_dir).mkdir(parents=True, exist_ok=True)
|
| 14 |
+
|
| 15 |
+
seen_names = set()
|
| 16 |
+
for jsonl_file in Path(raw_dir).rglob("*.jsonl"):
|
| 17 |
+
if jsonl_file.parent == Path(raw_dir):
|
| 18 |
+
continue
|
| 19 |
+
parent_dir = jsonl_file.parent.name
|
| 20 |
+
prefixed_name = f"{parent_dir}_{jsonl_file.name}"
|
| 21 |
+
dest = Path(raw_dir) / prefixed_name
|
| 22 |
+
if prefixed_name not in seen_names:
|
| 23 |
+
shutil.copy2(jsonl_file, dest)
|
| 24 |
+
seen_names.add(prefixed_name)
|
| 25 |
+
print(f"Copied: {jsonl_file.name} -> {prefixed_name}")
|
| 26 |
+
|
| 27 |
+
processor = DataProcessor(raw_dir=raw_dir, processed_dir=processed_dir)
|
| 28 |
+
summary = processor.process_all()
|
| 29 |
+
|
| 30 |
+
print()
|
| 31 |
+
print("=== RESULTS ===")
|
| 32 |
+
for k, v in summary.items():
|
| 33 |
+
if k != "per_file_stats":
|
| 34 |
+
print(f" {k}: {v}")
|
| 35 |
+
print()
|
| 36 |
+
for fs in summary.get("per_file_stats", []):
|
| 37 |
+
f = fs["file"]
|
| 38 |
+
l = fs["loaded"]
|
| 39 |
+
k = fs["kept"]
|
| 40 |
+
t = fs["tokens"]
|
| 41 |
+
print(f" {f}: loaded={l}, kept={k}, tokens={t}")
|