Tutorial for plugin creation : build failed

jlbellier

Hello everybody,

I am getting stuck while building the project "Pluging Dev (Tutorial)" defined in the DSS Tutorials.
The build fails in the Python recipe and returns the following message

I get a warninng telling me that a file does not exist

: File does not exist: /apps/dataiku/DSS_DATA/managed_datasets/DKU_TUT_PLUGINDEV.ecommerce_transactions_with_ip

What I do not understand is that is supposed to be the output, but I can see the schema of it while looking at the output parameters of the Python recipe.

I surely have misunderstood something, but I do not know what. I joined the list of packages I use (there are surely useless libraries for the project, but I use this environment for other projects).

I also send the full log I got for the build.

Could somebody give light to my lantern ?

Thank you in advance.

Best regards,

Jean-Luc.

---

Operating system used: Windows

Miguel Angel

Hi,

The actual error here is the following:

[15:37:58] [INFO] [dku.utils]  -   File "/apps/dataiku/DSS_DATA/code-envs/python/test_JLB/lib/python3.7/site-packages/pandas/io/common.py", line 3, in <module>
[15:37:58] [INFO] [dku.utils]  -     import bz2
[15:37:58] [INFO] [dku.utils]  -   File "/usr/local/lib/python3.7/bz2.py", line 19, in <module>
[15:37:58] [INFO] [dku.utils]  -     from _bz2 import BZ2Compressor, BZ2Decompressor
[15:37:58] [INFO] [dku.utils]  - ModuleNotFoundError: No module named '_bz2'

This implies that the Python 3.7 interpreter being used has not been compiled with bzip2 support. There are instructions online as to how to do that:https://stackoverflow.com/questions/22346269/bz2-is-module-not-available-when-installing-pandas-with-pip-in-python-virtual

After this, rebuild your 'test_JLB' code env. Make sure the 'Rebuild env' option has been checked.

Finally, try to run the job again.

jlbellier

Hello MiguelAngelC,

I discovered that there was a problem on Pandas, and I tried to install the bz2 package in my environement but it failed.
I surely missed something in the update of my environment.
Thank you for the explanation. I will try this again.

Have a good day.
Best regards,
Jean-Luc.

jlbellier

Hello everybody,

I tried to install bz2 but it fails : I get the message :

/apps/dataiku/DSS_DATA/tmp/pip-requirements-install/req8378613288349190582.txt (line 12)) (23.2.1)ERROR: Could not find a version that satisfies the requirement bz2 (from versions: none)ERROR: No matching distribution found for bz2

How can I solve this problem ? Is there an alternative way to install bz2 in my environment ? I used the standard way, and I selected in the dropdown box for pandas version : Pandas 1.0 (Py3 only)'.

Any help would be greatly appreciated.
Best regards,
Jean-Luc.