Trying flights tutorial, problem with using renku update

RenkuLab
1

Hello, I am new to Renku and I am trying to go through the following tutorial: First Steps Tutorial — Renku documentation.

When I try the Build a reproducible workflow > Update your results, I get an error when I call

renku update -all
I get a very long log message, including
Writing failed [Errno 2] No such file or directory: ‘tmp/avirphlu/.cache/black/20.8b1/tmp1_8anv1r’
And I don’t see modifications in 01-CountFlights.ran.ipynb after updating 00-FilterFlights.py. I then tried to run 01-CountFlights.ipynb again, but now I get the same error and
data/output/2019-01-flights-count.txt is unchanged and
renku status
give me some Outdated outputs and Modified inputs no matter how many times (OK, once should have been enough) I try to call ranku update --all!

I am confused on how I should use renku update and on why it is not updating data and notebooks as expected. Could someone help?

Thanks!

2

Hi! I do not see anything obviously erroneous in what you posted. Is your project on RenkuLab? If so, could you post a link to it? (You may need to make it internal or public so I can see it.) I can take a look at the project to better understand what is going on.

3

Yes, the project is on RenkuLab and I just made it public. Looking at the output files (.csv and .txt in data/output/), the renku update –all command seems to have worked fine (I changed ‘AUS’ to ‘CLE’ in 00-FilterFlights.py to test a new change). However, I expected to see 01-CountFlights.ran.ipynb to be updated with a modified output after I change 00-FilterFlights.py and run renku update –all, but it does not.

In the very long output when I call renku update –all I see “Writing failed: [Errno 2] No such file or directory: “ which might or might not be related to my issue.

4

Thanks for making the project public. I’m looking into it and trying to reconstruct exactly what happened. Looking at the history for data/output/2019-01-flights-count.txt, I see the first version is in commit 805a557e and produced this content:

diff --git a/data/output/2019-01-flights-count.txt b/data/output/2019-01-flights-count.txt
new file mode 100644
index 0000000..2745edf
--- /dev/null
+++ b/data/output/2019-01-flights-count.txt
@@ -0,0 +1 @@
+There were 23078 flights to Austin, TX in Jan 2019.

The next version is in the commit d852d61d, which is the result of a renku update --all and produced this modification:

diff --git a/data/output/2019-01-flights-count.txt b/data/output/2019-01-flights-count.txt
index 2745edf..b3cc5e4 100644
--- a/data/output/2019-01-flights-count.txt
+++ b/data/output/2019-01-flights-count.txt
@@ -1 +1 @@
-There were 23078 flights to Austin, TX in Jan 2019.
+There were 4951 flights to Austin, TX in Jan 2019.

Then I see in commit 3e20d415, that you changed the search string to CLE (as you mention above). After that, you ran a renku update --all and that produced

diff --git a/data/output/2019-01-flights-count.txt b/data/output/2019-01-flights-count.txt
index b3cc5e4..8c2dcfa 100644
--- a/data/output/2019-01-flights-count.txt
+++ b/data/output/2019-01-flights-count.txt
@@ -1 +1 @@
-There were 4951 flights to Austin, TX in Jan 2019.
+There were 3907 flights to Austin, TX in Jan 2019.

So the history of changes all look good.

I’m not sure why you were encountering these errors. In any case, I do not see any trace of them in the history of the project.

As for why a renku update --all might not change the content of the file. I’m guessing that what is happening here is that you need to close and re-open the file to see the updates. I think JupyterLab is not noticing that the file has been updated and so the content is not refreshed. Closing an open view on the file and re-opening it should fix that.

I hope this helps!