[AIRFLOW-6162] Add back serialization as a module

[potiuk]

Make sure you have checked all steps below.

Jira

• My PR addresses the following Airflow Jira issues and references them in the PR title. For example, "[AIRFLOW-XXX] My Airflow PR"
  • https://issues.apache.org/jira/browse/AIRFLOW-6162

Description

• Here are some details about my PR, including screenshots of any UI changes:

Tests

• My PR adds the following unit tests OR does not need testing for this extremely good reason:

Commits

• My commits all reference Jira issues in their subject lines, and I have squashed multiple commits if they address the same issue. In addition, my commits follow the guidelines from "How to write a good git commit message":
  1. Subject is separated from body by a blank line
  2. Subject is limited to 50 characters (not including Jira issue reference)
  3. Subject does not end with a period
  4. Subject uses the imperative mood ("add", not "adding")
  5. Body wraps at 72 characters
  6. Body explains "what" and "why", not "how"

Documentation

• In case of new functionality, my PR adds documentation that describes how to use it.
  • All the public functions and the classes in the PR contain docstrings that explain what it does
  • If you implement backwards incompatible changes, please leave a note in the Updating.md so we can assign it to a appropriate release

[kaxil]

Is this required?

[potiuk]

Yes. MyPy complains if it's not there. Maybe it's worth looking whether there is a good reason for that @kaxil ?

[potiuk]

I will shortly

[kaxil]

Cool, other than that I am fine with this PR :)

Thanks @potiuk

[potiuk]

I think it's worth looking at it @kaxil -> there are still some parts of the code that needs some more typing to catch potential errors (DagBag especially) but this one looks a bit fishy to me.

[potiuk]

I think the root cause is that .subdag field is only set in SerializedBaseOperator and that's why the type is forced here.

[nemkin]

Tried this locally and it fixed the missing serialization module after running pip install . Thank you Jarek :)

[kaxil]

Yes unfortunately, subdag field is only set in 2 places. 1) In SubDagOperator and 2) In SerializedBaseOperator.

…

On Tue, Dec 3, 2019, 12:04 Jarek Potiuk ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In airflow/serialization/serialized_dag.py <#6718 (comment)>: > @@ -101,7 +102,7 @@ def deserialize_dag(cls, encoded_dag: dict) -> "SerializedDAG": setattr(dag, 'full_filepath', dag.fileloc) for task in dag.task_dict.values(): task.dag = dag - serializable_task: SerializedBaseOperator = task + serializable_task: SerializedBaseOperator = cast(SerializedBaseOperator, task) I think the root cause is that .subdag field is only set in SerializedBaseOperator and that's why the type is forced here. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#6718?email_source=notifications&email_token=ACDHIJRK2BUGT2I7RAVJOELQWZDOTA5CNFSM4JUWMCP2YY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCNX3QWQ#discussion_r353139748>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACDHIJUQCSU3UYLNHUZJQ63QWZDOVANCNFSM4JUWMCPQ> .

[potiuk]

@kaxil -> I have a fix in making for that (but the #6596 should be merge first because we have the errors fixed in this PR).

[potiuk]

In my upcoming fix I also merge serialized_dag and serialized_baseoperator into one module 'serialized_objects' - because there are cyclic dependencies (again detected by pylint) between those two classes as well and it makes perfect sense to group them in one module.

[potiuk]

@kaxil @ashb @mik-laj -> Just pushed fixup which removes the cast and fixes it "properly" I think. I moved subdag to BaseOperator and added few missing type annotations. This has let to more cyclic dependencies (between SerializedDag and SerializedBaseOperator). I believe those two classes are so closely coupled that it makes sense to put them in one module (basically that's what pylint is complaining about when telling about "cyclic imports" - that classes in separate modules are so closely coupled that they are referring each other).

BTW. I think the more type annotations we add the more couplings we detect and there will be another set of splits/merges between the core classes. We have still quite a few of not-annotated classes that are hiding such couplings.

[potiuk]

I will rebase it on top of #6596 to see if it works with hose fixes.

[codecov-io]

Codecov Report

Merging #6718 into master will decrease coverage by 0.33%.
The diff coverage is 96.09%.

@@            Coverage Diff             @@
##           master    #6718      +/-   ##
==========================================
- Coverage   83.84%   83.51%   -0.34%     
==========================================
  Files         669      668       -1     
  Lines       37692    37677      -15     
==========================================
- Hits        31603    31465     -138     
- Misses       6089     6212     +123

Impacted Files	Coverage Δ
airflow/models/dagbag.py	86.53% <0%> (ø)	⬆️
airflow/serialization/__init__.py	100% <100%> (ø)
airflow/models/baseoperator.py	96.3% <100%> (+0.01%)	⬆️
airflow/models/dag.py	90.95% <100%> (ø)	⬆️
airflow/models/serialized_dag.py	87.65% <100%> (ø)	⬆️
airflow/serialization/serialized_objects.py	90.94% <96.72%> (ø)
airflow/kubernetes/volume_mount.py	44.44% <0%> (-55.56%)	⬇️
airflow/kubernetes/volume.py	52.94% <0%> (-47.06%)	⬇️
airflow/kubernetes/pod_launcher.py	45.25% <0%> (-46.72%)	⬇️
airflow/kubernetes/refresh_config.py	50.98% <0%> (-23.53%)	⬇️
... and 5 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a36cfe0...256e216. Read the comment docs.

[potiuk]

Rebased now on top of #6596

[potiuk]

And some more pylint complaints (rightfully so) I also moved BaseSerialization in the same "serialization_objects.py". Now this module is really what it should be :).

[potiuk]

@ashb @kaxil -> pls take a look. It's changed since the approval a lot.