acoustic-van-30942
05/29/2023, 7:24 PM@pyspark decorator. The documentation is a bit lacking (understandably so since it's experimental), but looks like it's a Metaflow extension.
Was also looking into awswrangler as well, which works quite well too.
I tried doing python setup.py install, but I get the following error:
RuntimeError: Conflicts in 'metaflow_extensions' files:
Packages 'metaflow-pyspark[pyspark]', and '/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_pyspark-0.0.1-py3.10.egg/metaflow_extensions[pyspark]' define the same configuration module 'metaflow_extensions.pyspark.plugins.__init__'dry-beach-38304
05/29/2023, 7:58 PMMETAFLOW_DEBUG_EXT and I may be able to glean additional information about why it is finding the extension twice.acoustic-van-30942
05/29/2023, 7:59 PMacoustic-van-30942
05/29/2023, 8:03 PMmetaflow_extensions: Found packages present at {'/home/sagemaker-user/spark-flow/metaflow_extensions', '/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_extensions'}
metaflow_extensions: Found extension package 'metaflow-pyspark'...
metaflow_extensions: Top-level 'pyspark' extends 'plugins' with config 'metaflow_extensions.pyspark.plugins.__init__'
metaflow_extensions: Ignoring duplicate package 'metaflow-pyspark' (duplicate paths in sys.path? (['/home/sagemaker-user/spark-flow', '/opt/conda/envs/pyspark/lib/python310.zip', '/opt/conda/envs/pyspark/lib/python3.10', '/opt/conda/envs/pyspark/lib/python3.10/lib-dynload', '/opt/conda/envs/pyspark/lib/python3.10/site-packages', '/opt/conda/envs/pyspark/lib/python3.10/site-packages/certifi-2023.5.7-py3.10.egg']))
metaflow_extensions: 'metaflow_extensions' distributions order is ['metaflow-pyspark']
metaflow_extensions: Non installed packages present at {'/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_extensions'}
metaflow_extensions: Walking path /opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_extensions (package name _pythonpath_0)
metaflow_extensions: Extends 'plugins' with config 'metaflow_extensions.pyspark.plugins.__init__'
Traceback (most recent call last):
File "/home/sagemaker-user/spark-flow/sparkflow.py", line 1, in <module>
from metaflow import FlowSpec, step, IncludeFile, Parameter
File "/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow/__init__.py", line 54, in <module>
from metaflow.extension_support import (
File "/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow/extension_support/__init__.py", line 794, in <module>
_all_packages, _pkgs_per_extension_point = _get_extension_packages()
File "/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow/extension_support/__init__.py", line 756, in _get_extension_packages
raise RuntimeError(
RuntimeError: Conflicts in 'metaflow_extensions' files:
Packages 'metaflow-pyspark[pyspark]', and '/opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_extensions[pyspark]' define the same configuration module 'metaflow_extensions.pyspark.plugins.__init__'dry-beach-38304
05/29/2023, 9:34 PMhome/sagemaker-user/spark-flow/metaflow_extensions and /opt/conda/envs/pyspark/lib/python3.10/site-packages/metaflow_extensions so metaflow gets confused. The messages are basically telling me:
• you have the package in those two places
• both places are in sys.path (that’s the duplicate message — it’s telling you that it is ignoring the second package it found)
I can actually make a small fix that will actually ignore the second package. Here it is ignoring it once when it does processing of sys.path but then still goes and looks for it.
I can have a small patch for you to do this. You can alternatively remove one of the two packages and it should work.acoustic-van-30942
05/29/2023, 9:37 PMdry-beach-38304
05/29/2023, 9:40 PMdry-beach-38304
05/29/2023, 9:40 PMacoustic-van-30942
05/29/2023, 9:45 PMdry-beach-38304
05/29/2023, 10:00 PM