[#835] allow key-agnostic iRODSMeta construction for AVU assignments by d-w-moore · Pull Request #836 · irods/python-irodsclient

d-w-moore · 2026-09-14T15:49:55Z

No description provided.

…assignments

d-w-moore · 2026-09-15T17:07:50Z

ready for review. Until review comments are seen, the changes from here on will be only for codacy/ruff and other minor modifications as needed.

MartinFlores751

Code is looking good to me, only have one suggestion. Learned about the := operator.

MartinFlores751 · 2026-10-19T15:24:11Z

+            # Set a number of metadata AVUs with the iRODSMeta builder invocation.  (Each of
+            # these AVU values follow a predictable relation defined by the test_mapping function.)
+            def test_mapping(name): return str(ord(name))
+            avu_names = [chr(_) for _ in range(ord('a'), ord('z')+1)]


Style question. Is it preferable to use _ here? I thought _ was equivalent to "discard this result" and was a bit surprised to see it work.

_ is a valid dummy variable name, so our tools like it better than having a meaningfule variable like loopVariable because, you know, namespace pollution. Of course it's also true as of Python 3.x that list comprehension temporaries do not leak outward, so ... really we can choose anything.

_ , in the interpreter, means "last result". Other than that , I don't think there's any special meaning....

Style question. Is it preferable to use _ here? I thought _ was equivalent to "discard this result" and was a bit surprised to see it work.

now that I look at it though, [ letter for letter in string.ascii_lowercase ] might be more self-documenting - ?

_ is a valid dummy variable name, so our tools like it better than having a meaningfule variable like loopVariable because, you know, namespace pollution. Of course it's also true as of Python 3.x that list comprehension temporaries do not leak outward, so ... really we can choose anything.

_ , in the interpreter, means "last result". Other than that , I don't think there's any special meaning....

Learned a lot about how _ is treated and scoping behavior here. Thanks for the info!

Style question. Is it preferable to use _ here? I thought _ was equivalent to "discard this result" and was a bit surprised to see it work.

now that I look at it though, [ letter for letter in string.ascii_lowercase ] might be more self-documenting - ?

Definitely like self-documenting code. If you think it'll help future readers, I'd change _ to letter.

korydraughn · 2026-10-20T13:48:39Z

            # data.metadata(admin = True) generates a cloned object but for the one change to "admin".
            data.metadata.admin = True

+    def test_iRODSMeta_builder__issue_835(self):


Is this test about exercising basic functionality of the new builder?

korydraughn · 2026-10-20T13:50:38Z

+            # Use both forms of the iRODSMeta builder invocation, testing that both attempts
+            # resulted in the AVU we expected. (For sets, s1 < s2 iff s1 is a proper subset of s2.)
+            data_obj.metadata['mile'] = iRODSMeta.builder(value=KM_PER_MILE, units='km')
+            data_obj.metadata['foot'] = iRODSMeta.builder(CM_PER_FOOT, 'cm')
+            self.assertLess(
+                {
+                    iRODSMeta('mile', KM_PER_MILE, 'km'),
+                    iRODSMeta('foot', CM_PER_FOOT, 'cm'),
+                },
+                set(data_obj.metadata.items())
+            )


If I'm understanding correctly, set(data_obj.metadata.items()) will be the superset because it has more AVUs?

korydraughn · 2026-10-20T13:57:19Z

-With only one AVU on the object with a name of "key2", *get_one*
-is assured of not throwing an exception:
+Alternatively, in the indexed AVU assignment, the following construction can
+be used in lieu of the direct call to the iRODSMeta constructor, thus avoiding redundant use of the


We need to determine if there's a pattern in how we style types and other important words. I suspect there are inconsistencies in the README.

Suggested change

be used in lieu of the direct call to the iRODSMeta constructor, thus avoiding redundant use of the

be used in lieu of the direct call to the `iRODSMeta` constructor, thus avoiding redundant use of the

We need to do a consistency sweep across the READMEs. That can happen in a separate PR.

korydraughn · 2026-10-20T13:58:50Z

+any other pre-existing AVUs with a name field of 'key2' before the requested assignment is
+actually made.
+
+Be aware that use of the indexing form on obj.metadata to retrieve AVUs


Suggested change

Be aware that use of the indexing form on obj.metadata to retrieve AVUs

Be aware that use of the indexing form on `obj.metadata` to retrieve AVUs

korydraughn · 2026-10-20T13:59:48Z

+```
+x = obj.metadata['key1']
+```


Missing python styling keyword on opening backticks.

korydraughn · 2026-10-20T14:00:00Z

+```
+>>> obj.metadata['key2'] = iRODSMeta.builder(value='value5', units='units2')
+```


Missing python styling keyword on opening backticks.

korydraughn · 2026-10-20T14:08:48Z

 ```

-In versions of iRODS 4.2.12 and later, we can also do:
+In iRODS 4.2.12 and after, a rodsadmin can apply the ADMIN_KW


Suggested change

In iRODS 4.2.12 and after, a rodsadmin can apply the ADMIN_KW

In iRODS 4.2.12 and after, a rodsadmin can apply the `ADMIN_KW`

korydraughn · 2026-10-20T14:12:28Z

+actually made.
+
+Be aware that use of the indexing form on obj.metadata to retrieve AVUs
+with code such as


Add colon after "as"? Noticed other places in this diff had a colon.

Suggested change

with code such as

with code such as:

korydraughn · 2026-10-20T14:12:46Z

+
+For these reasons, the indexed assignment may be better considered an artifact of convenience
+rather than reliable, straightforward coding practice.  For clear and unambiguous intent, the
+canonical iRODS API endpoints should be used, with calls such as


Add colon after "as"? Noticed other places in this diff had a colon.

Suggested change

canonical iRODS API endpoints should be used, with calls such as

canonical iRODS API endpoints should be used, with calls such as:

[_835] allow key-agnostic iRODSMeta construction for __setitem__ AVU …

f4dec6b

…assignments

d-w-moore marked this pull request as draft September 14, 2026 15:54

d-w-moore added 2 commits September 14, 2026 12:24

[_835] test

60889bb

corrections to the test

4367d28

d-w-moore changed the title ~~#835 allow key-agnostic iRODSMeta construction for AVU assignments~~ [#835] allow key-agnostic iRODSMeta construction for AVU assignments Sep 14, 2026

d-w-moore marked this pull request as ready for review September 14, 2026 19:04

README section on iRODSMeta.builder

457df1a

d-w-moore marked this pull request as draft September 15, 2026 12:29

d-w-moore force-pushed the 835.m branch from b4733e9 to 1db8d06 Compare September 15, 2026 12:34

d-w-moore added 5 commits September 15, 2026 08:38

avubuilder chgs and further mods for iRODSMeta subclass compatibility

642a175

test example

b320849

README refinements

c68c9ca

other chgs to meta_test

415e7f5

[_new_issue] touch and improve README

af75c13

d-w-moore force-pushed the 835.m branch from 1db8d06 to af75c13 Compare September 15, 2026 12:39

d-w-moore added 3 commits September 15, 2026 09:35

tests/comments

ac5de9e

no longer need custom class component

5d8c0da

whitespace

3fefcf0

d-w-moore marked this pull request as ready for review September 15, 2026 16:38

use temp dat obj in test

d61ab5d

trel reviewed Sep 15, 2026

View reviewed changes

Comment thread irods/test/meta_test.py

trel reviewed Sep 15, 2026

View reviewed changes

Comment thread irods/meta.py Outdated

d-w-moore added 2 commits September 16, 2026 04:34

add more comments for explanation in tests

017e031

delete old commented-out code

21d2f4f

d-w-moore force-pushed the 835.m branch from 04b27f4 to 21d2f4f Compare September 16, 2026 08:34

korydraughn requested a review from MartinFlores751 October 15, 2026 20:41

MartinFlores751 reviewed Oct 19, 2026

View reviewed changes

use string.ascii_lowercase to build vector in test

adc3470

korydraughn reviewed Oct 20, 2026

View reviewed changes

	be used in lieu of the direct call to the iRODSMeta constructor, thus avoiding redundant use of the
	be used in lieu of the direct call to the `iRODSMeta` constructor, thus avoiding redundant use of the

	Be aware that use of the indexing form on obj.metadata to retrieve AVUs
	Be aware that use of the indexing form on `obj.metadata` to retrieve AVUs

	In iRODS 4.2.12 and after, a rodsadmin can apply the ADMIN_KW
	In iRODS 4.2.12 and after, a rodsadmin can apply the `ADMIN_KW`

	canonical iRODS API endpoints should be used, with calls such as
	canonical iRODS API endpoints should be used, with calls such as:

Conversation

d-w-moore commented Sep 14, 2026

Uh oh!

d-w-moore commented Sep 15, 2026

Uh oh!

Uh oh!

Uh oh!

MartinFlores751 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Milestone

Development

Uh oh!

4 participants