move dataset load to build-time

This commit is contained in:
Fraser
2023-03-31 21:02:53 -04:00
parent 88efcbc709
commit 42b016cae5
3 changed files with 29 additions and 20 deletions
+21
View File
@@ -0,0 +1,21 @@
import pickle
import requests
import os
print('Downloading dataset...')
url = os.environ.get('DATASET_URL')
if url is None:
print('No dataset url provided.')
exit()
dataset_dict_bytes = requests.get(url).content
print('Unpacking dataset...')
dataset_dict = pickle.loads(dataset_dict_bytes)
print('Writing dataset to disk...')
with open('dataset.pkl', 'wb') as f:
pickle.dump(dataset_dict, f)
print('Done!')
+3 -20
View File
@@ -6,7 +6,6 @@ import dataclasses
import os
import openai
import pickle
import requests
# ---------------------------------- env setup ---------------------------------
@@ -20,25 +19,9 @@ openai.api_key = OPENAI_API_KEY
# -------------------------------- load dataset --------------------------------
if os.path.exists('dataset.pkl'):
print('Found dataset.pkl')
print('Loading dataset...')
with open('dataset.pkl', 'rb') as f:
dataset_dict = pickle.load(f)
else:
print('No dataset.pkl found on disk.')
print('Downloading dataset...')
url = os.environ.get('DATASET_URL')
if url is None:
print('No dataset url provided.')
exit()
dataset_dict_bytes = requests.get(url).content
print('Unpacking dataset...')
dataset_dict = pickle.loads(dataset_dict_bytes)
print('Loading dataset...')
with open('dataset.pkl', 'rb') as f:
dataset_dict = pickle.load(f)
print('Done!')
+5
View File
@@ -0,0 +1,5 @@
[phases.build]
cmds = ['python3 dataset_dl.py']
[start]
cmd = 'gunicorn --timeout 300 "main:app"'