dataflow-core 2.0.5__py3-none-any.whl → 2.0.6__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of dataflow-core might be problematic. Click here for more details.
- dataflow/dataflow.py +43 -4
- dataflow/scripts/create_environment.sh +34 -0
- {dataflow_core-2.0.5.dist-info → dataflow_core-2.0.6.dist-info}/METADATA +1 -1
- {dataflow_core-2.0.5.dist-info → dataflow_core-2.0.6.dist-info}/RECORD +7 -6
- {dataflow_core-2.0.5.dist-info → dataflow_core-2.0.6.dist-info}/WHEEL +0 -0
- {dataflow_core-2.0.5.dist-info → dataflow_core-2.0.6.dist-info}/entry_points.txt +0 -0
- {dataflow_core-2.0.5.dist-info → dataflow_core-2.0.6.dist-info}/top_level.txt +0 -0
dataflow/dataflow.py
CHANGED
|
@@ -2,21 +2,21 @@ import os, requests
|
|
|
2
2
|
from .models.database import DatabaseManager
|
|
3
3
|
from sqlalchemy.inspection import inspect
|
|
4
4
|
from .utils.aws_secrets_manager import SecretsManagerClient
|
|
5
|
-
import json
|
|
5
|
+
import json, asyncio, pkg_resources
|
|
6
6
|
from authenticator.package.configuration import ConfigurationManager
|
|
7
7
|
|
|
8
8
|
|
|
9
9
|
class Dataflow:
|
|
10
10
|
def __init__(self):
|
|
11
|
-
self.dataflow_config = ConfigurationManager('/dataflow/app/auth_config/dataflow_auth.cfg')
|
|
12
|
-
self.auth_api = self.dataflow_config.get_config_value('auth', 'ui_auth_api')
|
|
13
11
|
self.secrets_manager = SecretsManagerClient('us-east-1')
|
|
14
12
|
|
|
15
13
|
def auth(self, session_id: str):
|
|
16
14
|
"""Retrieve user information from the auth API."""
|
|
17
15
|
try:
|
|
16
|
+
dataflow_config = ConfigurationManager('/dataflow/app/auth_config/dataflow_auth.cfg')
|
|
17
|
+
auth_api = dataflow_config.get_config_value('auth', 'ui_auth_api')
|
|
18
18
|
response = requests.get(
|
|
19
|
-
|
|
19
|
+
auth_api,
|
|
20
20
|
cookies={"dataflow_session": session_id, "jupyterhub-hub-login": ""}
|
|
21
21
|
)
|
|
22
22
|
|
|
@@ -86,3 +86,42 @@ class Dataflow:
|
|
|
86
86
|
|
|
87
87
|
except Exception as e:
|
|
88
88
|
return None
|
|
89
|
+
|
|
90
|
+
async def create_env(self, env_name, py_version, py_requirements, status, env_version=None):
|
|
91
|
+
"""
|
|
92
|
+
Creates a conda environment at the specified path and installs libraries in one command.
|
|
93
|
+
"""
|
|
94
|
+
config = ConfigurationManager('/dataflow/app/config/dataflow.cfg')
|
|
95
|
+
status = status.lower()
|
|
96
|
+
if status == "published":
|
|
97
|
+
env_base_path = config.get_config_value('paths', 'published_env_path')
|
|
98
|
+
conda_env_path = os.path.join(env_base_path, env_name)
|
|
99
|
+
else:
|
|
100
|
+
env_base_path = config.get_config_value('paths', 'drafts_env_path')
|
|
101
|
+
conda_env_path = os.path.join(env_base_path, env_name, f"{env_name}_v{env_version}")
|
|
102
|
+
try:
|
|
103
|
+
if not os.path.exists(conda_env_path):
|
|
104
|
+
os.makedirs(conda_env_path, exist_ok=True)
|
|
105
|
+
|
|
106
|
+
py_requirements = ",".join(py_requirements)
|
|
107
|
+
|
|
108
|
+
script_path = pkg_resources.resource_filename('dataflow', 'scripts/create_environment.sh')
|
|
109
|
+
|
|
110
|
+
# Make the script executable
|
|
111
|
+
os.chmod(script_path, 0o755)
|
|
112
|
+
|
|
113
|
+
# Prepare command with arguments
|
|
114
|
+
command = ["bash", script_path, py_requirements, conda_env_path, py_version]
|
|
115
|
+
|
|
116
|
+
process = await asyncio.create_subprocess_exec(
|
|
117
|
+
*command,
|
|
118
|
+
stdout=asyncio.subprocess.PIPE,
|
|
119
|
+
stderr=asyncio.subprocess.PIPE
|
|
120
|
+
)
|
|
121
|
+
|
|
122
|
+
return process
|
|
123
|
+
except OSError as e:
|
|
124
|
+
print(f"OS error while creating {conda_env_path}: {e}")
|
|
125
|
+
except Exception as e:
|
|
126
|
+
print(f"Unexpected error while creating {conda_env_path}: {e}")
|
|
127
|
+
return {"error": str(e)}
|
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
#!/bin/bash
|
|
2
|
+
set -e
|
|
3
|
+
|
|
4
|
+
IFS=',' read -r -a libraries <<< $1
|
|
5
|
+
conda_env_path=$2
|
|
6
|
+
py_version=$3
|
|
7
|
+
|
|
8
|
+
# Use an isolated conda package cache to avoid concurrency issues
|
|
9
|
+
export CONDA_PKGS_DIRS=$(mktemp -d)
|
|
10
|
+
# to delete conda package cache after script finishes
|
|
11
|
+
trap 'rm -rf "$CONDA_PKGS_DIRS"' EXIT
|
|
12
|
+
|
|
13
|
+
# 1. Creating conda environment
|
|
14
|
+
conda create --prefix ${conda_env_path} --yes python=${py_version}
|
|
15
|
+
${conda_env_path}/bin/pip install --root-user-action ignore ${libraries[@]}
|
|
16
|
+
|
|
17
|
+
# 3. Install Dataflow Airflow to a separate path in environment
|
|
18
|
+
${conda_env_path}/bin/pip install \
|
|
19
|
+
--force-reinstall --root-user-action ignore \
|
|
20
|
+
--no-warn-conflicts dataflow-airflow==2.10.3 \
|
|
21
|
+
--target ${conda_env_path}/bin/airflow-libraries/ \
|
|
22
|
+
--constraint https://raw.githubusercontent.com//apache/airflow/constraints-2.10.3/constraints-${py_version}.txt || true
|
|
23
|
+
|
|
24
|
+
files=(
|
|
25
|
+
${conda_env_path}/lib/python${py_version}/site-packages/dbt/config/profile.py
|
|
26
|
+
${conda_env_path}/lib/python${py_version}/site-packages/dbt/task/debug.py
|
|
27
|
+
)
|
|
28
|
+
for file in ${files[@]}
|
|
29
|
+
do
|
|
30
|
+
awk '{gsub("from dbt.clients.yaml_helper import load_yaml_text", "from dbt.dataflow_config.secrets_manager import load_yaml_text"); print}' $file > temp
|
|
31
|
+
mv temp $file
|
|
32
|
+
done
|
|
33
|
+
|
|
34
|
+
echo "Environment Creation Successful"
|
|
@@ -10,15 +10,16 @@ authenticator/package/models/session.py,sha256=j6PhbrTMJxEkeDT4Vf5SqGtM_LI_vZy9O
|
|
|
10
10
|
authenticator/package/models/user.py,sha256=IYogp_vt0yDBG5i936uNPjgTis77VYPzITn9XpQUIyw,788
|
|
11
11
|
dataflow/__init__.py,sha256=WTRg8HMpMWSgxYJ9ZGVldx4k07fAbta3mBmZ1hG9mWE,30
|
|
12
12
|
dataflow/configuration.py,sha256=7To6XwH1eESiYp39eqPcswXWwrdBUdPF6xN6WnazOF0,663
|
|
13
|
-
dataflow/dataflow.py,sha256=
|
|
13
|
+
dataflow/dataflow.py,sha256=viEKxu1bVFtenbPx8etCb7Gf-Hbwv0KvkE7occVbX8I,5136
|
|
14
14
|
dataflow/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
15
15
|
dataflow/models/database.py,sha256=y09pqnglsBBtDZlyhvqDAlpUSFovwAzBAi6jOYl_XNk,896
|
|
16
16
|
dataflow/models/session.py,sha256=C9crPh6ZDFuL27hZ_zhUXDZZ0ZiIDE8ZD19O_4WPw-I,488
|
|
17
17
|
dataflow/models/user.py,sha256=IYogp_vt0yDBG5i936uNPjgTis77VYPzITn9XpQUIyw,788
|
|
18
|
+
dataflow/scripts/create_environment.sh,sha256=r4cKRTMf1G9pXx1fqMpzkvd-ROzUH0Qp16f9t84Nh5c,1300
|
|
18
19
|
dataflow/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
19
20
|
dataflow/utils/aws_secrets_manager.py,sha256=FqHm3YRynv580FpFsS0RfI1MSGY5aq-V7t4blpiYsS4,2588
|
|
20
|
-
dataflow_core-2.0.
|
|
21
|
-
dataflow_core-2.0.
|
|
22
|
-
dataflow_core-2.0.
|
|
23
|
-
dataflow_core-2.0.
|
|
24
|
-
dataflow_core-2.0.
|
|
21
|
+
dataflow_core-2.0.6.dist-info/METADATA,sha256=BCXu55SZS71k7ZQ2ilfFn_nimogjDv0SDKtrDyje5iw,317
|
|
22
|
+
dataflow_core-2.0.6.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
|
|
23
|
+
dataflow_core-2.0.6.dist-info/entry_points.txt,sha256=lDLG2MMWlKfkqsVWFghF7sx-yEvM2xqMmHE7rMTysE4,118
|
|
24
|
+
dataflow_core-2.0.6.dist-info/top_level.txt,sha256=SZsUOpSCK9ntUy-3Tusxzf5A2e8ebwD8vouPb1dPt_8,23
|
|
25
|
+
dataflow_core-2.0.6.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|
|
File without changes
|