Enhance completions benchmark generator to extract words from a fallback Perl copyright file

This commit is contained in:
Mikhail Yevchenko
2026-05-21 19:33:41 +00:00
parent f77d943d79
commit 3285d9118f
+11 -5
View File
@@ -1,5 +1,6 @@
import random
import os
import re
from vastai import Worker, WorkerConfig, HandlerConfig, LogActionConfig, BenchmarkConfig
@@ -22,11 +23,16 @@ def completions_benchmark_generator() -> dict:
WORD_LIST = []
with open(__file__, 'r') as f:
# Use regex to extract words from the source code
import re
source_code = f.read()
WORD_LIST = re.findall(r'\b\w+\b', source_code)
# Try to load from perl copyright file first
try:
with open("/usr/share/doc/perl/copyright", 'r') as f:
source_code = f.read()
WORD_LIST = re.findall(r'\b\w+\b', source_code)
except (FileNotFoundError, IOError):
# Fallback to loading from python file
with open(__file__, 'r') as f:
source_code = f.read()
WORD_LIST = re.findall(r'\b\w+\b', source_code)
prompt = " ".join(random.choices(WORD_LIST, k=int(250)))
model = os.environ.get("MODEL_NAME")