Fix/small forecasting pipeline fixes by BelhsanHmida · Pull Request #2011 · FlexMeasures/flexmeasures

BelhsanHmida · 2026-03-09T10:04:52Z

Description

This PR includes several small fixes and improvements to the forecasting pipeline jobs:

Fix the TypeError: Job.__init__() missing 1 required argument: 'connection' raised by the wrap-up job.
Improve the returned job information by including the number of jobs created together with the wrap-up job ID.
Update tests to reflect the updated job return structure.
Add db session commit before the forecasting jobs are created to prevent data_source not in database error
add n_cycles in calculation of forecasts assertion in test_forecasting_pipeline.py to correctly test how many forecasts generated in case of multiple cycles

These changes ensure that forecasting jobs run correctly and that job metadata returned by the pipeline is clearer and easier to interpret.

Sign-off

I agree to contribute to the project under Apache 2 License.
To the best of my knowledge, the proposed patch is not based on code under GPL or other license that is incompatible with FlexMeasures

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

read-the-docs-community · 2026-03-09T10:09:58Z

Documentation build overview

📚 flexmeasures | 🛠️ Build #31857658 | 📁 Comparing 39351a3 against latest (f1707fc)

🔍 Preview build

Show files changed (4 files in total): 📝 4 modified | ➕ 0 added | ➖ 0 deleted

File	Status
changelog.html	📝 modified
genindex.html	📝 modified
_autosummary/flexmeasures.data.models.forecasting.pipelines.train_predict.html	📝 modified
api/v3_0.html	📝 modified

Signed-off-by: F.N. Claessen <claessen@seita.nl>

…dated Signed-off-by: F.N. Claessen <claessen@seita.nl>

…ters Signed-off-by: F.N. Claessen <claessen@seita.nl>

…ycle job Signed-off-by: F.N. Claessen <claessen@seita.nl>

Flix6x · 2026-03-09T13:04:05Z

Thanks. I also pushed some changes to the test myself now, which now covers the case where more than 1 cycle is asked, thereby resulting in a wrap-up job being returned. We did not cover that yet. There was also a mistake in an if-statement in the test that resulted in us checking each item in an empty list of cycle_job_ids. See f26c41b.

The new test case unexpectedly failed on creating 3 forecasts instead of the expected 2 (one per cycle). Could you please investigate why? From inspecting the forecasts, I believe the last one doesn't make sense. Also the assert comment needs to be updated. None of the test cases lead to 4 forecasts per horizon, I think.

        assert (
            len(forecasts) == m_viewpoints * n_hourly_horizons * n_events_per_horizon
        ), f"we expect 4 forecasts per horizon for each viewpoint within the prediction window, and {m_viewpoints} viewpoints with each {n_hourly_horizons} hourly horizons"

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

… pipeline Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

…ecasts Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

…rity Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

BelhsanHmida · 2026-03-11T00:09:20Z

Thanks. I also pushed some changes to the test myself now, which now covers the case where more than 1 cycle is asked, thereby resulting in a wrap-up job being returned. We did not cover that yet. There was also a mistake in an if-statement in the test that resulted in us checking each item in an empty list of cycle_job_ids. See f26c41b.

The new test case unexpectedly failed on creating 3 forecasts instead of the expected 2 (one per cycle). Could you please investigate why? From inspecting the forecasts, I believe the last one doesn't make sense. Also the assert comment needs to be updated. None of the test cases lead to 4 forecasts per horizon, I think.

I investigated this.

The extra forecast comes from viewpoints within each cycle. In this case:

start = 2025-01-08 00:00+02:00
end = 2025-01-09 00:00+02:00
retrain-frequency = 12h
forecast-frequency = 12h
max-forecast-horizon = 1h
So the pipeline runs 2 cycles, but each cycle also generates 2 viewpoints because the predict window is still 24h and the predict pipeline uses predict_period_in_hours / forecast_frequency = 24 / 12 = 2 viewpoints per cycle.

That gives these forecasted events:

cycle 1, viewpoint 1: event at 2025-01-08 00:00+02:00
cycle 1, viewpoint 2: event at 2025-01-08 12:00+02:00
cycle 2, viewpoint 1: event at 2025-01-08 12:00+02:00
cycle 2, viewpoint 2: event at 2025-01-09 00:00+02:00
So the “last” forecast does make sense: it is the second viewpoint of the second cycle. The overlap is at 2025-01-08 12:00+02:00, which is forecast once from the first cycle and then again from the second cycle with a newer belief time.

This is also why:

with most_recent_beliefs_only=True, you see 3 rows, because the two beliefs for 2025-01-08 12:00+02:00 collapse to the most recent one
with most_recent_beliefs_only=False, you see the full 4 saved forecasts

to more correctly test which forecasts are being saved into the sensor i set most_recent_beliefs_only=False in sensor.search_beliefs and introduced number of cycles in calculation of forecasts assertion

Flix6x · 2026-03-11T07:35:40Z

Thank you for the clear explanation. That almost clears it up for me completely. Two lingering doubts:

Why does the second cycle have a viewpoint exactly at the end? That feels odd to me. The only case where that would actually make sense to me is when forecasting an instantaneous sensor (for which the last event in the predict period is exactly at the end).

And, if the end is meant to be included, why does the first cycle not include the end? That looks inconsistent, doesn't it?

BelhsanHmida · 2026-03-11T09:35:19Z

Thank you for the clear explanation. That almost clears it up for me completely. Two lingering doubts:

Why does the second cycle have a viewpoint exactly at the end? That feels odd to me. The only case where that would actually make sense to me is when forecasting an instantaneous sensor (for which the last event in the predict period is exactly at the end).

And, if the end is meant to be included, why does the first cycle not include the end? That looks inconsistent, doesn't it?

yes true the last viewpoint we should see is end - sensor_resolution. i'm looking into this

BelhsanHmida · 2026-03-17T11:58:41Z

Thank you for the clear explanation. That almost clears it up for me completely. Two lingering doubts:
Why does the second cycle have a viewpoint exactly at the end? That feels odd to me. The only case where that would actually make sense to me is when forecasting an instantaneous sensor (for which the last event in the predict period is exactly at the end).
And, if the end is meant to be included, why does the first cycle not include the end? That looks inconsistent, doesn't it?

yes true the last viewpoint we should see is end - sensor_resolution. i'm looking into this

i identified the issue to be in base_pipeline when we assigned the end date when splitting the data-frames. i think this issue should be handled in a separate pr.

Flix6x · 2026-03-17T12:40:33Z

Okay, please create the issue for it.

…n.commit() some time to finish writing a new source to the db Signed-off-by: F.N. Claessen <claessen@seita.nl>

Signed-off-by: F.N. Claessen <claessen@seita.nl>

This reverts commit 888b980. Signed-off-by: F.N. Claessen <claessen@seita.nl>

Signed-off-by: F.N. Claessen <claessen@seita.nl>

…-pipeline-fixes

Signed-off-by: F.N. Claessen <claessen@seita.nl>

…-pipeline-fixes

Signed-off-by: F.N. Claessen <claessen@seita.nl>

… we put it back into the session we are committing Signed-off-by: F.N. Claessen <claessen@seita.nl>

…b.session.commit() some time to finish writing a new source to the db" This reverts commit 4ebaa97. Signed-off-by: F.N. Claessen <claessen@seita.nl>

Signed-off-by: F.N. Claessen <claessen@seita.nl>

- Fix 'viewpoint' -> 'viewpoints' in comment - Add ', and' before n_cycles in assertion message for clarity - Improve comment on db.session.merge() to explain why merge is needed before commit (get_or_create_source flushes but does not commit)

… method; cherry-picked from cbff3b4 Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com>

* fix: enhance job ID return structure in TrainPredictPipeline Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: update forecasting job return structure in SensorAPI Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: update job fetching logic in test_train_predict_pipeline Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * chore: remove commented-out breakpoint in test_forecasting.py Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: as_job is no longer in parameters Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: update job count retrieval in add_forecast function Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: add connection queue to fetch job Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * style: black Signed-off-by: F.N. Claessen <claessen@seita.nl> * docs: changelog entry Signed-off-by: F.N. Claessen <claessen@seita.nl> * feat: check if wrap-up job actually finished rather than failed Signed-off-by: F.N. Claessen <claessen@seita.nl> * feat: add test case for 2 cycles, yielding 2 jobs and a wrap-up job Signed-off-by: F.N. Claessen <claessen@seita.nl> * dev: comment out failing assert, which needs to be investgated and updated Signed-off-by: F.N. Claessen <claessen@seita.nl> * refactor: move checking the status of the wrap-up job to where it matters Signed-off-by: F.N. Claessen <claessen@seita.nl> * fix: use job ID itself in case the returned job is the one existing cycle job Signed-off-by: F.N. Claessen <claessen@seita.nl> * fix: add db.commit before forecasting jobs are created Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * dev: uncomment test assertion statement Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * Test(feat): search all beliefs forecasts saved into the sensor by the pipeline Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * test(feat): add n_cycles variable to use to account for length of forecasts Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * style: run pre-commit Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: improve assertion message in test_train_predict_pipeline for clarity Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> * fix: first create all jobs, then queue all jobs, giving the db.session.commit() some time to finish writing a new source to the db Signed-off-by: F.N. Claessen <claessen@seita.nl> * feat: enqueue job only after the transactional request Signed-off-by: F.N. Claessen <claessen@seita.nl> * Revert "feat: enqueue job only after the transactional request" This reverts commit 888b980. Signed-off-by: F.N. Claessen <claessen@seita.nl> * docs: resolve silent merge conflict in changelog Signed-off-by: F.N. Claessen <claessen@seita.nl> * docs: delete duplicate changelog entry Signed-off-by: F.N. Claessen <claessen@seita.nl> * docs: add release date for v0.31.2 Signed-off-by: F.N. Claessen <claessen@seita.nl> * docs: advance a different bugfix to v0.31.2 Signed-off-by: F.N. Claessen <claessen@seita.nl> * fix: self.data_source found itself in a different session somehow, so we put it back into the session we are committing Signed-off-by: F.N. Claessen <claessen@seita.nl> * Revert "fix: first create all jobs, then queue all jobs, giving the db.session.commit() some time to finish writing a new source to the db" This reverts commit 4ebaa97. Signed-off-by: F.N. Claessen <claessen@seita.nl> * refactor: move asserts to where they matter Signed-off-by: F.N. Claessen <claessen@seita.nl> * fix: cherry-pick from d40246f Signed-off-by: F.N. Claessen <claessen@seita.nl> * fix: improve comments and assertions from code review - Fix 'viewpoint' -> 'viewpoints' in comment - Add ', and' before n_cycles in assertion message for clarity - Improve comment on db.session.merge() to explain why merge is needed before commit (get_or_create_source flushes but does not commit) * refactor: move reattachment of sensor objects to session into a class method; cherry-picked from cbff3b4 Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com> --------- Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com> Signed-off-by: F.N. Claessen <claessen@seita.nl> Co-authored-by: F.N. Claessen <claessen@seita.nl> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com> (cherry picked from commit f3c22f3) Signed-off-by: F.N. Claessen <claessen@seita.nl>

BelhsanHmida added 7 commits March 9, 2026 10:37

fix: enhance job ID return structure in TrainPredictPipeline

921d958

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: update forecasting job return structure in SensorAPI

b7068fa

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: update job fetching logic in test_train_predict_pipeline

505c284

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

chore: remove commented-out breakpoint in test_forecasting.py

10b370a

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: as_job is no longer in parameters

6342867

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: update job count retrieval in add_forecast function

cc1bf1b

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: add connection queue to fetch job

a7ff4a9

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

BelhsanHmida self-assigned this Mar 9, 2026

BelhsanHmida added the Forecasting label Mar 9, 2026

nhoening added this to the 0.31.2 milestone Mar 9, 2026

Flix6x added 7 commits March 9, 2026 13:22

style: black

745e513

Signed-off-by: F.N. Claessen <claessen@seita.nl>

docs: changelog entry

47f160d

Signed-off-by: F.N. Claessen <claessen@seita.nl>

feat: check if wrap-up job actually finished rather than failed

3af2c62

Signed-off-by: F.N. Claessen <claessen@seita.nl>

feat: add test case for 2 cycles, yielding 2 jobs and a wrap-up job

474b860

Signed-off-by: F.N. Claessen <claessen@seita.nl>

dev: comment out failing assert, which needs to be investgated and up…

7f824a7

…dated Signed-off-by: F.N. Claessen <claessen@seita.nl>

refactor: move checking the status of the wrap-up job to where it mat…

29705a2

…ters Signed-off-by: F.N. Claessen <claessen@seita.nl>

fix: use job ID itself in case the returned job is the one existing c…

f26c41b

…ycle job Signed-off-by: F.N. Claessen <claessen@seita.nl>

BelhsanHmida added 6 commits March 10, 2026 02:56

fix: add db.commit before forecasting jobs are created

f326efc

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

dev: uncomment test assertion statement

67862ed

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

Test(feat): search all beliefs forecasts saved into the sensor by the…

64465e9

… pipeline Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

test(feat): add n_cycles variable to use to account for length of for…

4ac1846

…ecasts Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

style: run pre-commit

bdc5e28

Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

fix: improve assertion message in test_train_predict_pipeline for cla…

5340350

…rity Signed-off-by: Mohamed Belhsan Hmida <mohamedbelhsanhmida@gmail.com>

BelhsanHmida added 2 commits March 11, 2026 06:26

Merge branch 'main' into fix/small-forecasting-pipeline-fixes

8271e28

Merge branch 'main' into fix/small-forecasting-pipeline-fixes

60a85b3

Flix6x added 12 commits March 17, 2026 15:57

fix: first create all jobs, then queue all jobs, giving the db.sessio…

4ebaa97

…n.commit() some time to finish writing a new source to the db Signed-off-by: F.N. Claessen <claessen@seita.nl>

feat: enqueue job only after the transactional request

888b980

Signed-off-by: F.N. Claessen <claessen@seita.nl>

Revert "feat: enqueue job only after the transactional request"

56f825a

This reverts commit 888b980. Signed-off-by: F.N. Claessen <claessen@seita.nl>

docs: resolve silent merge conflict in changelog

e41c1a5

Signed-off-by: F.N. Claessen <claessen@seita.nl>

Merge remote-tracking branch 'origin/main' into fix/small-forecasting…

9085b9e

…-pipeline-fixes

docs: delete duplicate changelog entry

064f9cb

Signed-off-by: F.N. Claessen <claessen@seita.nl>

docs: add release date for v0.31.2

7ff82d1

Signed-off-by: F.N. Claessen <claessen@seita.nl>

Merge remote-tracking branch 'origin/main' into fix/small-forecasting…

657b0c0

…-pipeline-fixes

docs: advance a different bugfix to v0.31.2

ff78988

Signed-off-by: F.N. Claessen <claessen@seita.nl>

fix: self.data_source found itself in a different session somehow, so…

4ed641a

… we put it back into the session we are committing Signed-off-by: F.N. Claessen <claessen@seita.nl>

Revert "fix: first create all jobs, then queue all jobs, giving the d…

d276c8a

…b.session.commit() some time to finish writing a new source to the db" This reverts commit 4ebaa97. Signed-off-by: F.N. Claessen <claessen@seita.nl>

refactor: move asserts to where they matter

d710365

Signed-off-by: F.N. Claessen <claessen@seita.nl>

Copilot AI mentioned this pull request Mar 17, 2026

Fix detached DataSource and expired sensors when enqueuing forecast jobs #2034

Closed

Flix6x and others added 3 commits March 18, 2026 09:26

fix: cherry-pick from d40246f

a6d09bb

Signed-off-by: F.N. Claessen <claessen@seita.nl>

refactor: move reattachment of sensor objects to session into a class…

39351a3

… method; cherry-picked from cbff3b4 Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com>

Flix6x approved these changes Mar 18, 2026

View reviewed changes

Flix6x added the bug Something isn't working label Mar 18, 2026

Flix6x merged commit f3c22f3 into main Mar 18, 2026
11 checks passed

Flix6x deleted the fix/small-forecasting-pipeline-fixes branch March 18, 2026 08:59

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix/small forecasting pipeline fixes#2011

Fix/small forecasting pipeline fixes#2011
Flix6x merged 37 commits intomainfrom
fix/small-forecasting-pipeline-fixes

BelhsanHmida commented Mar 9, 2026 •

edited

Loading

Uh oh!

read-the-docs-community bot commented Mar 9, 2026 •

edited

Loading

Uh oh!

Flix6x commented Mar 9, 2026

Uh oh!

BelhsanHmida commented Mar 11, 2026

Uh oh!

Flix6x commented Mar 11, 2026

Uh oh!

BelhsanHmida commented Mar 11, 2026 •

edited by Flix6x

Loading

Uh oh!

BelhsanHmida commented Mar 17, 2026 •

edited

Loading

Uh oh!

Flix6x commented Mar 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

BelhsanHmida commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Sign-off

Uh oh!

read-the-docs-community bot commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Documentation build overview

Uh oh!

Flix6x commented Mar 9, 2026

Uh oh!

BelhsanHmida commented Mar 11, 2026

Uh oh!

Flix6x commented Mar 11, 2026

Uh oh!

BelhsanHmida commented Mar 11, 2026 • edited by Flix6x Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BelhsanHmida commented Mar 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Flix6x commented Mar 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

BelhsanHmida commented Mar 9, 2026 •

edited

Loading

read-the-docs-community bot commented Mar 9, 2026 •

edited

Loading

BelhsanHmida commented Mar 11, 2026 •

edited by Flix6x

Loading

BelhsanHmida commented Mar 17, 2026 •

edited

Loading