Run ferc_to_sqlite and pudl_etl independent of integration tests #2825

rousik · 2023-09-03T16:35:32Z

This reworks how tests (unit, integration) are done by github actions.

In particular, it modifies how ci-integration is executed by running ferc_to_sqlite and pudl_etl commands independently, making use of dagster parallelism and larger runners.

.github/workflows/tox-pytest.yml

For more information, see https://pre-commit.ci

…l into parallel-pytest

.github/workflows/tox-pytest.yml

zaneselvans · 2023-09-06T17:30:09Z

I think the call to pytest is failing because you need to install some or all of the extra dependencies when you install the catalystcoop.pudl package -- right now you're just installing the bare minimum requirements. so...

pip install ./[doc,test,dev,datasette]

rather than just

pip install .

rousik · 2023-12-13T21:34:47Z

I think I'm going to loose my mind here. I thought I finally figured the right configuration here only to find out that it still seems to be busted, so back to the rabbit hole of investigations 🤦

zaneselvans · 2023-12-15T04:06:44Z

@rousik I merged dev in to get the pre-commit version of the lockfile updates, which should involve much less churn.

It seems like it's mostly getting the coverage, with 90.1%. Are the tests themselves not being covered now maybe? It looks like somehow the test/unit and test/integration lines in the coverage config section of pyproject.toml got lost merging. I added them back in.

zaneselvans · 2024-01-17T20:17:47Z

@rousik Can you update the main description for this PR to reflect the current situation? It's been through so many changes and the code on main has changed so much that I'm no longer clear what the net effect is.

The integration test run time seems to be almost identical to the current tests (~1.5 hours).

Is the primary change that this PR is running the ETL outside the context of pytest? Where were we going with that? Is the next step to kick off the FERC to SQLite conversions as their own jobs on smaller runners which can run in parallel rather than serially, and have the main PUDL ETL start on a larger runner as soon as the FERC 1 DBs are available, hopefully shortening the overall integration test runtime and also reducing the running time for the paid runner?

rousik · 2024-01-18T05:16:57Z

@rousik Can you update the main description for this PR to reflect the current situation? It's been through so many changes and the code on main has changed so much that I'm no longer clear what the net effect is.

The integration test run time seems to be almost identical to the current tests (~1.5 hours).

Is the primary change that this PR is running the ETL outside the context of pytest? Where were we going with that? Is the next step to kick off the FERC to SQLite conversions as their own jobs on smaller runners which can run in parallel rather than serially, and have the main PUDL ETL start on a larger runner as soon as the FERC 1 DBs are available, hopefully shortening the overall integration test runtime and also reducing the running time for the paid runner?

I suppose I need to rerun some of these tests again, because we should be getting speed-ups from this even if nothing more is done. Isolating ferc_to_sqlite, pudl_etl and integration tests in separate steps should make it a little more easy to diagnose which parts failed, and should also (as pointed out) allow us to further speed-up/improve the individual steps as needed. We could beef-up the pudl_etl part, hopefully getting faster there, and then sharding ferc_to_sqlite which should be a good time improvement.

rousik · 2024-01-18T05:22:04Z

@rousik Can you update the main description for this PR to reflect the current situation? It's been through so many changes and the code on main has changed so much that I'm no longer clear what the net effect is.

The integration test run time seems to be almost identical to the current tests (~1.5 hours).

Is the primary change that this PR is running the ETL outside the context of pytest? Where were we going with that? Is the next step to kick off the FERC to SQLite conversions as their own jobs on smaller runners which can run in parallel rather than serially, and have the main PUDL ETL start on a larger runner as soon as the FERC 1 DBs are available, hopefully shortening the overall integration test runtime and also reducing the running time for the paid runner?

I have re-read contents of this PR and description and I think they are still in sync. This refactors ci-integration into smaller individual tasks.

zaneselvans

Mostly requests for clarification on why some things are changing.

.github/workflows/pytest.yml

zaneselvans · 2024-01-18T19:07:00Z

.github/workflows/pytest.yml

-    strategy:
-      fail-fast: false


Why do we want fail-fast enabled, given that we've already the dependency above specified? I think this would mean that if the docs build fails, the unit and integration tests would be canceled, right?

.github/workflows/pytest.yml

zaneselvans · 2024-01-18T19:12:39Z

.github/workflows/pytest.yml

@@ -198,37 +221,40 @@ jobs:
        with:
          path: coverage


Why do we use path: coverage here but when we're uploading the artifact above we use path: coverage.xml?

Perhaps this is a bit of confusing api. When uploading artifacts, path tells you what files/directories you want to upload, and optionally you can give them name to uniquely idenfiy them. When you're downloading artifacts, path tells you where to download things. You can either pick specific artifact names to download, or you can just download everything (which is happening here). What happens is that you'll end up with coverage/coverage-unit/coverage.xml, coverage/coverage-integration/coverage.xml and so on.

zaneselvans · 2024-01-18T19:20:46Z

pyproject.toml

 # See note above on need to specify separate sources for pytest-coverage and coverage.
-source = ["src/pudl/", "test/integration/", "test/unit/"]
+# source = ["src/pudl/", "test/integration/", "test/unit/"]
+include = [
+  "src/pudl/**",
+  "test/integration/**",
+  "test/unit/**",
+  "*/site-packages/pudl/**",
+]


What's going on with the change from source to include here? Should these be identical to the pudl_sources below? Should the commented out source line be removed? Is the comment referring above still relevant? Does switching from source to include mean we don't need the --cov=... arguments above?

If we are not installing as editable, pudl will be loaded from site-packages/pudl and I feel like source was misbehaving and did not quite work, while include does.

AFAIK in the new ci-integration that is part of this PR, the coverage commands should work w/o the need of --cov arguments above. I'm not even sure that was there when I was working on this PR initially.

pudl_sources below is a translation block that maps site-packages/pudl/foo.py onto src/pudl/foo.py, so that the coverage works even when the libraries are loaded form slightly different location.

I suppose all of this could be made to work if we simply install pudl as editable.

.github/workflows/pytest.yml

For more information, see https://pre-commit.ci

rousik

Huh, lots of my comments were left in limbo. Posting.

.github/workflows/pytest.yml

rousik · 2023-11-29T00:29:13Z

.github/workflows/pytest.yml

@@ -58,13 +64,13 @@ jobs:
          path: coverage.xml

  ci-unit:
+    name: Unit tests


Do you know where are those merge requirements specified? I think it would be nice to use human friendly names here, but we can do it in a separate step distinct from this one.

.github/workflows/pytest.yml

rousik · 2024-01-18T20:07:16Z

.github/workflows/pytest.yml

@@ -198,37 +221,40 @@ jobs:
        with:
          path: coverage


Perhaps this is a bit of confusing api. When uploading artifacts, path tells you what files/directories you want to upload, and optionally you can give them name to uniquely idenfiy them. When you're downloading artifacts, path tells you where to download things. You can either pick specific artifact names to download, or you can just download everything (which is happening here). What happens is that you'll end up with coverage/coverage-unit/coverage.xml, coverage/coverage-integration/coverage.xml and so on.

.github/workflows/pytest.yml

rousik · 2024-01-18T20:13:25Z

pyproject.toml

 # See note above on need to specify separate sources for pytest-coverage and coverage.
-source = ["src/pudl/", "test/integration/", "test/unit/"]
+# source = ["src/pudl/", "test/integration/", "test/unit/"]
+include = [
+  "src/pudl/**",
+  "test/integration/**",
+  "test/unit/**",
+  "*/site-packages/pudl/**",
+]


If we are not installing as editable, pudl will be loaded from site-packages/pudl and I feel like source was misbehaving and did not quite work, while include does.

AFAIK in the new ci-integration that is part of this PR, the coverage commands should work w/o the need of --cov arguments above. I'm not even sure that was there when I was working on this PR initially.

pudl_sources below is a translation block that maps site-packages/pudl/foo.py onto src/pudl/foo.py, so that the coverage works even when the libraries are loaded form slightly different location.

I suppose all of this could be made to work if we simply install pudl as editable.

zaneselvans

Comparing locally generated test coverage and the CodeCov output from CI.

zaneselvans · 2024-01-19T14:56:57Z

pyproject.toml

-source = ["src/pudl/", "test/integration/", "test/unit/"]
+include = [
+  "src/pudl/**",
+  "test/integration/**",
+  "test/unit/**",
+  "*/site-packages/pudl/**",
+]
 omit = [
    # Never hit by integration tests:
    "src/pudl/validate.py",
 ]
+sigterm = true
+concurrency=["multiprocessing"]
+debug = ["config", "trace"]
+
+[tool.coverage.paths]
+# When running pudl tools installed with pip, the sources are imported
+# from package-data/pudl directory. The following maps this to raw
+# source files.
+pudl_sources = [
+  "src/pudl/",
+  "*/site-packages/pudl/",
+  "test/unit",
+  "test/integration",
+]


I'm re-running the CodeCov tests now, but it seems like something has changed that's reducing the test coverage by a few hundred lines / a couple of percent, and I don't understand what it is.

Looking at the indirect coverage changes some differences make sense, like increased coverage for the ferc_to_sqlite and pudl_etl CLIs, but there's also substantial apparent losses of coverage in modules related to the FERC-EIA record linkage, and the pudltabl output module which don't make sense.

If I run

make pytest-coverage

locally I get overall coverage of 92.8% which is comparable to the coverage on main

Name Stmts Miss Cover ---------------------------------------------------------------------------------- src/pudl/__init__.py 14 0 100.0% src/pudl/analysis/__init__.py 1 0 100.0% src/pudl/analysis/fuel_by_plant.py 37 0 100.0% src/pudl/analysis/record_linkage/__init__.py 1 0 100.0% src/pudl/convert/__init__.py 1 0 100.0% src/pudl/etl/__init__.py 22 0 100.0% src/pudl/etl/eia_bulk_elec_assets.py 9 0 100.0% src/pudl/etl/epacems_assets.py 54 0 100.0% src/pudl/etl/static_assets.py 24 0 100.0% src/pudl/extract/__init__.py 1 0 100.0% src/pudl/extract/csv.py 29 0 100.0% src/pudl/extract/eia176.py 9 0 100.0% src/pudl/extract/eia860.py 40 0 100.0% src/pudl/extract/eia860m.py 30 0 100.0% src/pudl/extract/eia923.py 51 0 100.0% src/pudl/extract/eia_bulk_elec.py 46 0 100.0% src/pudl/extract/epacems.py 47 0 100.0% src/pudl/extract/ferc.py 14 0 100.0% src/pudl/extract/ferc2.py 21 0 100.0% src/pudl/extract/ferc6.py 15 0 100.0% src/pudl/extract/ferc60.py 15 0 100.0% src/pudl/extract/ferc714.py 21 0 100.0% src/pudl/ferc_to_sqlite/__init__.py 24 0 100.0% src/pudl/glue/__init__.py 1 0 100.0% src/pudl/metadata/__init__.py 1 0 100.0% src/pudl/metadata/codes.py 5 0 100.0% src/pudl/metadata/constants.py 21 0 100.0% src/pudl/metadata/dfs.py 12 0 100.0% src/pudl/metadata/enums.py 21 0 100.0% src/pudl/metadata/fields.py 25 0 100.0% src/pudl/metadata/labels.py 12 0 100.0% src/pudl/metadata/resources/__init__.py 12 0 100.0% src/pudl/metadata/resources/allocate_gen_fuel.py 4 0 100.0% src/pudl/metadata/resources/eia.py 4 0 100.0% src/pudl/metadata/resources/eia860.py 3 0 100.0% src/pudl/metadata/resources/eia861.py 2 0 100.0% src/pudl/metadata/resources/eia923.py 4 0 100.0% src/pudl/metadata/resources/eia_bulk_elec.py 2 0 100.0% src/pudl/metadata/resources/epacems.py 3 0 100.0% src/pudl/metadata/resources/ferc1.py 4 0 100.0% src/pudl/metadata/resources/ferc1_eia_record_linkage.py 3 0 100.0% src/pudl/metadata/resources/ferc714.py 3 0 100.0% src/pudl/metadata/resources/glue.py 3 0 100.0% src/pudl/metadata/resources/mcoe.py 4 0 100.0% src/pudl/metadata/resources/pudl.py 3 0 100.0% src/pudl/metadata/sources.py 5 0 100.0% src/pudl/output/__init__.py 1 0 100.0% src/pudl/output/eia860.py 19 0 100.0% src/pudl/output/eia_bulk_elec.py 14 0 100.0% src/pudl/output/sql/helpers.py 12 0 100.0% src/pudl/resources.py 17 0 100.0% src/pudl/transform/__init__.py 1 0 100.0% src/pudl/transform/eia_bulk_elec.py 26 0 100.0% src/pudl/transform/ferc714.py 76 0 100.0% src/pudl/transform/params/__init__.py 1 0 100.0% src/pudl/transform/params/ferc1.py 67 0 100.0% src/pudl/workspace/__init__.py 1 0 100.0% test/integration/console_scripts_test.py 23 0 100.0% test/integration/datasette_metadata_test.py 22 0 100.0% test/integration/etl_test.py 76 0 100.0% test/integration/ferc1_eia_train_test.py 30 0 100.0% test/integration/ferc_dbf_extract_test.py 38 0 100.0% test/integration/zenodo_datapackage_test.py 11 0 100.0% test/unit/analysis/plant_parts_eia_test.py 48 0 100.0% test/unit/analysis/spatial_test.py 67 0 100.0% test/unit/analysis/state_demand_test.py 8 0 100.0% test/unit/analysis/timeseries_cleaning_test.py 32 0 100.0% test/unit/console_scripts_test.py 10 0 100.0% test/unit/extract/csv_test.py 37 0 100.0% test/unit/extract/eia_bulk_elec_test.py 42 0 100.0% test/unit/extract/xbrl_test.py 74 0 100.0% test/unit/helpers_test.py 135 0 100.0% test/unit/output/epacems_test.py 7 0 100.0% test/unit/output/ferc1_test.py 9 0 100.0% test/unit/transform/classes_test.py 218 0 100.0% test/unit/transform/eia923_test.py 12 0 100.0% test/unit/transform/eia_bulk_elec_test.py 18 0 100.0% test/unit/transform/epacems_test.py 8 0 100.0% test/unit/transform/ferc1_test.py 133 0 100.0% test/unit/transform/glue_test.py 11 0 100.0% test/unit/workspace/datastore_test.py 82 0 100.0% test/unit/workspace/resource_cache_test.py 140 0 100.0% src/pudl/extract/eia861.py 36 1 97.2% src/pudl/extract/ferc1.py 113 1 99.1% test/integration/epacems_test.py 39 1 97.4% test/integration/output_test.py 69 1 98.6% test/unit/settings_test.py 145 1 99.3% src/pudl/__main__.py 2 2 0.0% src/pudl/analysis/mcoe.py 78 2 97.4% src/pudl/analysis/record_linkage/classify_plants_ferc1.py 31 2 93.5% src/pudl/etl/glue_assets.py 112 2 98.2% src/pudl/glue/ferc1_eia.py 135 2 98.5% src/pudl/output/censusdp1tract.py 28 2 92.9% src/pudl/output/eia923.py 93 2 97.8% src/pudl/workspace/setup.py 43 2 95.3% test/unit/analysis/allocate_gen_fuel_test.py 96 2 97.9% test/unit/extract/excel_test.py 49 2 95.9% test/unit/harvest_test.py 59 2 96.6% test/unit/metadata_test.py 56 2 96.4% src/pudl/extract/xbrl.py 52 3 94.2% src/pudl/logging_helpers.py 15 3 80.0% src/pudl/output/epacems.py 29 3 89.7% src/pudl/settings.py 253 3 98.8% src/pudl/transform/epacems.py 34 3 91.2% test/integration/record_linkage_test.py 72 3 95.8% src/pudl/convert/censusdp1tract_to_sqlite.py 35 4 88.6% src/pudl/metadata/helpers.py 187 4 97.9% test/integration/glue_test.py 42 4 90.5% test/unit/glue.py 7 4 42.9% src/pudl/output/ferc714.py 150 5 96.7% test/integration/jupyter_notebooks_test.py 13 5 61.5% src/pudl/analysis/record_linkage/embed_dataframe.py 124 6 95.2% src/pudl/analysis/record_linkage/eia_ferc1_record_linkage.py 218 7 96.8% src/pudl/convert/metadata_to_rst.py 21 7 66.7% src/pudl/analysis/spatial.py 112 8 92.9% src/pudl/etl/check_foreign_keys.py 20 8 60.0% src/pudl/extract/excel.py 167 8 95.2% src/pudl/transform/eia.py 300 8 97.3% src/pudl/analysis/plant_parts_eia.py 260 9 96.5% src/pudl/workspace/resource_cache.py 124 9 92.7% src/pudl/transform/eia861.py 398 10 97.5% test/unit/io_managers_test.py 206 10 95.1% src/pudl/output/pudltabl.py 98 11 88.8% src/pudl/analysis/epacamd_eia.py 18 12 33.3% src/pudl/analysis/record_linkage/link_cross_year.py 106 12 88.7% src/pudl/analysis/record_linkage/name_cleaner.py 90 12 86.7% src/pudl/analysis/allocate_gen_fuel.py 323 13 96.0% src/pudl/analysis/state_demand.py 146 15 89.7% src/pudl/ferc_to_sqlite/cli.py 44 15 65.9% src/pudl/transform/classes.py 394 21 94.7% src/pudl/etl/cli.py 48 22 54.2% src/pudl/transform/eia860.py 205 23 88.8% src/pudl/extract/dbf.py 258 25 90.3% src/pudl/workspace/datastore.py 298 25 91.6% src/pudl/io_managers.py 268 26 90.3% src/pudl/analysis/service_territory.py 123 31 74.8% src/pudl/transform/eia923.py 231 34 85.3% src/pudl/analysis/timeseries_cleaning.py 460 44 90.4% src/pudl/helpers.py 407 53 87.0% src/pudl/output/ferc1.py 632 66 89.6% src/pudl/transform/ferc1.py 1545 68 95.6% src/pudl/output/eia.py 176 72 59.1% src/pudl/analysis/record_linkage/eia_ferc1_train.py 172 79 54.1% src/pudl/metadata/classes.py 798 101 87.3% ---------------------------------------------------------------------------------- TOTAL 13072 943 92.8%

test/unit/glue.py needs to be renamed so pytest collects tests from it, but that's not a new thing.

test/integration/console_scripts_test.py doesn't get hit all in the CodeCov report but has 100% coverage if I run locally.

src/pudl/analysis/record_linkage/eia_ferc1_train.py coverage is dramatically lower in CodeCov than locally.

src/pudl/analysis/service_territory.py is also significantly lower. My guess is this is related to the scripts not getting tested?

src/pudl/output/pudltabl.py is basically not getting hit at all in CodeCov which makes no sense to me, since the PudlTabl objects are used all over the place in the tests and have 88.8% coverage when run locally.

Sorry for delays, life has been busy. I will try to follow up on these in the coming week and try to figure out what is going on here. The fact that coverage is so brittle and it is so hard to figure out why it fluctuates so much is very discouraging; it shouldn't be that hard :-/

rousik added 2 commits September 2, 2023 13:34

Try etl/integraiton separation in tests.

6bf1115

fix some formatting in the python thingies

d89e1f4

rousik changed the base branch from main to dev September 3, 2023 16:35

rousik added 5 commits September 4, 2023 11:41

Minor changes to the tox-pytest setup

3b43bef

Comment out most things to isolate ci-integration-etl

a8fb7dc

fix few things

8587dc7

fix condarc

a5b2867

install snappy before pudl

4916320

zaneselvans reviewed Sep 6, 2023

View reviewed changes

.github/workflows/tox-pytest.yml Outdated Show resolved Hide resolved

zaneselvans reviewed Sep 6, 2023

View reviewed changes

.github/workflows/tox-pytest.yml Outdated Show resolved Hide resolved

zaneselvans reviewed Sep 6, 2023

View reviewed changes

.github/workflows/tox-pytest.yml Outdated Show resolved Hide resolved

rousik added 5 commits September 6, 2023 12:58

Fix up integration tests, run on the large machine.

2b1a116

add alembic setup

bb623af

run conda info ahead of running the etl

da07819

Log conda env

7dc86aa

Disable large runners for now

162f62a

zaneselvans reviewed Sep 6, 2023

View reviewed changes

.github/workflows/tox-pytest.yml Outdated Show resolved Hide resolved

rousik and others added 7 commits September 6, 2023 13:14

fix exec shell for the action

68c6dee

fix exec shell for the action

09fb56d

[pre-commit.ci] auto fixes from pre-commit.com hooks

f05fcd3

For more information, see https://pre-commit.ci

Fix path to integration tests.

173ad8c

Add concurrency to cancel previous runs

c9b83ab

Merge branch 'parallel-pytest' of github.com:catalyst-cooperative/pud…

e35efc9

…l into parallel-pytest

Fix path to integration tests

a81ebf0

zaneselvans reviewed Sep 6, 2023

View reviewed changes

.github/workflows/tox-pytest.yml Outdated Show resolved Hide resolved

fix pytest paths

364b053

rousik changed the title ~~Parallel pytest - PR to test the new flow~~ Run pudl_etl prior to running integration tests Sep 6, 2023

rousik added 2 commits September 6, 2023 14:56

restructure etl steps

d4b46fc

fix typo in alembic setup

ffd56e0

rousik and others added 2 commits December 13, 2023 14:58

Blablabla.

5c9a194

Update conda-lock.yml and rendered conda environment files.

72ce2f2

rousik and others added 6 commits December 14, 2023 11:36

Try to tune coverage parameters once more.

6d032f9

Update conda-lock.yml and rendered conda environment files.

892adac

Merge remote-tracking branch 'origin/dev' into parallel-pytest

7f8619a

Reset lock files to match dev branch.

e94e277

Merge branch 'dev' into parallel-pytest

da3b055

Merge branch 'dev' into parallel-pytest

5d49446

Record coverage of unit and integration tests themselves

6696f2f

zaneselvans added the testing Writing tests, creating test data, automating testing, etc. label Dec 15, 2023

rousik added 2 commits December 29, 2023 11:59

Merge remote-tracking branch 'origin/dev' into parallel-pytest

fec6061

Merge branch 'dev' into parallel-pytest

c201637

Base automatically changed from dev to main January 5, 2024 04:14

Merge branch 'main' into parallel-pytest

8ec32bb

Merge remote-tracking branch 'origin/main' into parallel-pytest

e52c349

zaneselvans self-requested a review January 18, 2024 18:59

zaneselvans assigned rousik Jan 18, 2024

zaneselvans requested changes Jan 18, 2024

View reviewed changes

rousik and others added 2 commits January 18, 2024 13:14

Make some changes as per feedback.

131b9fa

[pre-commit.ci] auto fixes from pre-commit.com hooks

d17fa54

For more information, see https://pre-commit.ci

rousik commented Jan 18, 2024

View reviewed changes

zaneselvans changed the title ~~Rework integration tests (more parallelism, nicer things)~~ Run ferc_to_sqlite and pudl_etl independent of integration tests Jan 19, 2024

zaneselvans added 2 commits January 18, 2024 19:13

Merge branch 'main' into parallel-pytest

01a5efc

Rename glue.py unit tests so pytest picks them up.

c51d8e2

zaneselvans reviewed Jan 19, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Run ferc_to_sqlite and pudl_etl independent of integration tests #2825

Run ferc_to_sqlite and pudl_etl independent of integration tests #2825

rousik commented Sep 3, 2023 •

edited by zaneselvans

Loading

zaneselvans commented Sep 6, 2023

rousik commented Dec 13, 2023

zaneselvans commented Dec 15, 2023 •

edited

Loading

zaneselvans commented Jan 17, 2024

rousik commented Jan 18, 2024

rousik commented Jan 18, 2024

zaneselvans left a comment

zaneselvans Jan 18, 2024

zaneselvans Jan 18, 2024

rousik Jan 18, 2024

zaneselvans Jan 18, 2024

rousik Jan 18, 2024

rousik left a comment

rousik Nov 29, 2023

rousik Jan 18, 2024

rousik Jan 18, 2024

zaneselvans left a comment

zaneselvans Jan 19, 2024

rousik Feb 3, 2024

Run ferc_to_sqlite and pudl_etl independent of integration tests #2825

Are you sure you want to change the base?

Run ferc_to_sqlite and pudl_etl independent of integration tests #2825

Conversation

rousik commented Sep 3, 2023 • edited by zaneselvans Loading

zaneselvans commented Sep 6, 2023

rousik commented Dec 13, 2023

zaneselvans commented Dec 15, 2023 • edited Loading

zaneselvans commented Jan 17, 2024

rousik commented Jan 18, 2024

rousik commented Jan 18, 2024

zaneselvans left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rousik left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

zaneselvans left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rousik commented Sep 3, 2023 •

edited by zaneselvans

Loading

zaneselvans commented Dec 15, 2023 •

edited

Loading