Upgraded models to SQLAlchemy 2

[swainn]

Summary

Migrates tethysext-atcore to SQLAlchemy 2.0, including the modern typed-mapping syntax (Mapped[T] + mapped_column) and select()
query API. All tests pass on the new stack.

Primary changes

1. SQLAlchemy 2.0 API compliance

• engine.execute(sql_string) → with engine.connect() as conn: conn.execute(text(sql)) (or engine.begin() for DDL).
• All raw SQL strings now wrapped in text().
• GUID.process_result_value handles native uuid.UUID returns from psycopg2 in 2.0.
• URL repr changes (password masking) accommodated in tests.

2. Modern declarative syntax

• declarative_base() → class AppUsersBase(DeclarativeBase): pass.
• All model attributes use Mapped[T] annotations with mapped_column(...) — covers app_users, file_database,
  controller_metadata, and the resource_workflow hierarchy.
• backref= replaced with explicit back_populates= pairs across every relationship. New reverse-side relationships added on
  Resource.workflows/parents, AppUser.workflows, Organization.consultant, ControllerMetadata.step/result,
  ResourceWorkflowStep.workflow/parents, ResourceWorkflowResult.workflow/steps, and ResultsResourceWorkflowStep.source.

3. Query API migration

• session.query(M).filter(...).X() → session.execute(select(M).where(...)).scalars().X() across all production code
  (controllers, services, mixins, decorators, handlers, job_scripts) and integrated tests.
• session.query(M).get(id) → session.get(M, id).
• count() → select(func.count()).select_from(M).

4. Test fixtures

• MockEngine updated to support with engine.connect() context-manager pattern.
• Mocks for controllers patch select in the controller's namespace (or use session.get) so MagicMock model classes don't trip
  SQLA 2's ORM validation.

5. Bug fixes (previously hidden behind 1.4 behavior)

• table_input_wv.py: NODATA-filled optional columns now keep float64 dtype instead of being cast back to the template's object
  dtype, fixing a real bug exposed by pandas 3.0's stringify-on-object-cast behavior.
• condor_workflow_manager_tests: removed brittle hardcoded sequential workflow-id assertions; tests now check that all jobs in
  a manager share the same dynamically-assigned id.

6. Dev environment

• install.yml and README.md switched to sqlalchemy>=2, geoalchemy2>=0.13. README rewritten with venv-based install and Docker
  PostGIS test-DB workflow that mirrors CI.
• condorpy added as a dependency (was previously implicit; tests for the condor workflow manager require it).

Please review the checklist before submitting the Pull Request:

• Pull request has a meaning full name (not just the commit message from of your last commit)
• Code has been linted using flake8
• All methods have accurate Google-Style Docstrings
• 100% test coverage for new content
• Tests for each common use case (please don't write one test that covers all use cases)
• Bonus: Use TypeHints

Explain deviations from original design if applicable:

[Copilot]

Pull request overview

Migrates ATCore to SQLAlchemy 2.0 by updating ORM models to modern typed declarative mappings and refactoring query/engine usage across services, controllers, job scripts, and tests.

Changes:

• Updated models to SQLAlchemy 2.0 typed mappings (Mapped[], mapped_column) and explicit back_populates.
• Refactored ORM/query patterns to Session.get() and session.execute(select(...)) style.
• Updated tests/fixtures/mocks for SQLAlchemy 2.0 execution patterns and improved a few brittle test assertions.

Reviewed changes

Copilot reviewed 75 out of 76 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
tethysext/atcore/tests/unit_tests/services/model_db_spatial_manager.py	Updates engine mocks to use connect().__enter__().execute() call path.
tethysext/atcore/tests/integrated_tests/services/spatial_reference.py	Uses engine.begin() + text() for PostGIS extension DDL.
tethysext/atcore/tests/integrated_tests/services/resource_condor_workflow_tests.py	Updates mocked ORM usage from query().get() to get().
tethysext/atcore/tests/integrated_tests/services/model_database.py	Updates MockEngine to support connect() context manager and URL handling via make_url.
tethysext/atcore/tests/integrated_tests/services/file_database/file_database_client_tests.py	Migrates test queries/counts to SQLA 2.0 select()/execute() APIs.
tethysext/atcore/tests/integrated_tests/services/file_database/file_collection_client_tests.py	Migrates test queries/counts to SQLA 2.0 select()/execute() APIs.
tethysext/atcore/tests/integrated_tests/services/app_users/permissions_manager.py	Replaces ORM query with select() + scalar_one().
tethysext/atcore/tests/integrated_tests/services/app_users/condor_workflow_manager_tests.py	Makes workflow-id assertions robust to sequence state; updates helper signature.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/table_input_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/spatial_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/spatial_input_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/spatial_dataset_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/spatial_condor_job_rws_test.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/spatial_attributes_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/resource_workflow_steps/results_rws_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/initializer.py	Migrates query to select() + scalar_one_or_none().
tethysext/atcore/tests/integrated_tests/models/files_database/file_database_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/files_database/file_collection_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/app_users/spatial_resource_tests.py	Updates geometry and ORM queries to SQLA 2.0 select()/execute().
tethysext/atcore/tests/integrated_tests/models/app_users/resource_workflow_step_tests.py	Switches to Session.get() for load-by-id.
tethysext/atcore/tests/integrated_tests/models/app_users/resource_tests.py	Migrates counts/listing to SQLA 2.0 select()/execute().
tethysext/atcore/tests/integrated_tests/models/app_users/organization_tests.py	Migrates queries to select() + scalar_one_or_none().
tethysext/atcore/tests/integrated_tests/models/app_users/app_user_tests.py	Migrates many queries/counts to SQLA 2.0 select()/execute() APIs.
tethysext/atcore/tests/integrated_tests/mixins/file_collection_mixin_tests.py	Migrates count queries to select(func.count()).
tethysext/atcore/tests/integrated_tests/controllers/rest/spatial_reference.py	Uses engine.begin() + text() for PostGIS extension DDL.
tethysext/atcore/tests/integrated_tests/controllers/resources/tabs/workflows_tab_tests.py	Migrates workflow lookup to select() and Session.get().
tethysext/atcore/tests/integrated_tests/controllers/resource_workflows/workflow_views/xms_tool_wv_tests.py	Migrates query to select() + scalar_one().
tethysext/atcore/tests/integrated_tests/controllers/resource_workflows/workflow_views/set_status_wv_tests.py	Migrates query to select() + scalar_one().
tethysext/atcore/tests/integrated_tests/controllers/resource_workflows/workflow_view_mixins_tests.py	Updates mocks/expectations from query().filter().one() to execute().scalar_one().
tethysext/atcore/tests/integrated_tests/controllers/resource_workflows/base.py	Migrates list queries and resource lookup mocking to SQLA 2 patterns.
tethysext/atcore/tests/integrated_tests/controllers/app_users/modify_user.py	Updates tests to patch select in controller namespace and mock execute().scalar_one().
tethysext/atcore/tests/integrated_tests/controllers/app_users/mixins.py	Updates mocks from query().get() to get() and exception paths accordingly.
tethysext/atcore/tests/integrated_tests/controllers/app_users/manage_users_tests.py	Updates exception mock target from query to get.
tethysext/atcore/tests/integrated_tests/controllers/app_users/manage_resources.py	Updates mocks to use Session.get() path.
tethysext/atcore/tests/integrated_tests/controllers/app_users/manage_organizations.py	Updates mocks/patching for select() and execute().scalars().all().
tethysext/atcore/tests/integrated_tests/controllers/app_users/manage_organization_members.py	Updates mocks to use Session.get() path.
tethysext/atcore/tests/integrated_tests/controllers/app_users/add_existing_user.py	Updates tests to patch select and align with execute(select(...)) usage.
tethysext/atcore/services/spatial_reference.py	Refactors engine execution to SQLAlchemy 2 connection context + text().
tethysext/atcore/services/resource_workflows/decorators.py	Replaces deprecated .query().get() calls with Session.get().
tethysext/atcore/services/resource_condor_workflow.py	Replaces deprecated .query().get() with Session.get().
tethysext/atcore/services/model_db_spatial_manager.py	Refactors raw SQL execution to connection context + text().
tethysext/atcore/services/model_database.py	Refactors raw SQL execution to connection context + text() and ensures disposal in finally.
tethysext/atcore/services/file_database.py	Migrates .query().get() and .count() patterns to Session.get() and select(func.count()).
tethysext/atcore/services/app_users/decorators.py	Migrates user lookup query to select() + scalar_one_or_none().
tethysext/atcore/models/types/guid.py	Handles native uuid.UUID values returned by drivers in process_result_value.
tethysext/atcore/models/resource_workflow_steps/results_rws.py	Updates relationships to typed mapping and explicit back_populates; adds source relation.
tethysext/atcore/models/file_database/file_database.py	Converts model to typed mapping; adds explicit relationship typing.
tethysext/atcore/models/file_database/file_collection.py	Converts model to typed mapping; adds explicit relationship typing.
tethysext/atcore/models/controller_metadata.py	Converts model to typed mapping; replaces backrefs with explicit reverse relationships.
tethysext/atcore/models/app_users/user_setting.py	Converts model to typed mapping.
tethysext/atcore/models/app_users/spatial_resource.py	Converts extent to typed mapping and migrates spatial SQL usage to select() + execute().
tethysext/atcore/models/app_users/resource_workflow_step.py	Converts model to typed mapping; replaces backrefs with explicit parents/workflow/result relationships.
tethysext/atcore/models/app_users/resource_workflow_result.py	Converts model to typed mapping; adds explicit reverse relationships (workflow, steps).
tethysext/atcore/models/app_users/resource_workflow.py	Converts model to typed mapping; replaces backrefs with explicit back_populates.
tethysext/atcore/models/app_users/resource.py	Converts model to typed mapping; adds explicit parents/workflows relationships.
tethysext/atcore/models/app_users/organization.py	Converts model to typed mapping; replaces consultant backref with explicit relationships; wraps raw SQL in text().
tethysext/atcore/models/app_users/initializer.py	Migrates staff-user lookup to select() + scalar_one_or_none().
tethysext/atcore/models/app_users/base.py	Switches declarative base to SQLAlchemy 2 DeclarativeBase.
tethysext/atcore/models/app_users/app_user.py	Converts model to typed mapping; migrates multiple queries to select() + execute().
tethysext/atcore/job_scripts/update_resource_status.py	Replaces deprecated .query().get() with Session.get().
tethysext/atcore/handlers.py	Replaces deprecated .query().get() with Session.get().
tethysext/atcore/controllers/resources/tabs/workflows_tab.py	Migrates query building to select() and result consumption via execute().scalars().
tethysext/atcore/controllers/resource_workflows/workflow_views/table_input_wv.py	Fixes dtype coercion for NODATA-filled optional columns to preserve float dtype.
tethysext/atcore/controllers/resource_workflows/mixins.py	Migrates lookups to select() + scalar_one().
tethysext/atcore/controllers/app_users/modify_user.py	Migrates lookups to select() + scalar_one() and Session.get().
tethysext/atcore/controllers/app_users/modify_resource.py	Migrates resource/org lookups to Session.get() and list queries to select() + scalars().
tethysext/atcore/controllers/app_users/modify_organization.py	Migrates org/resource lookups to select() + scalar_one() and Session.get().
tethysext/atcore/controllers/app_users/mixins.py	Migrates resource lookups to Session.get().
tethysext/atcore/controllers/app_users/manage_users.py	Migrates list/delete operations to select() and Session.get().
tethysext/atcore/controllers/app_users/manage_resources.py	Migrates resource lookups to Session.get().
tethysext/atcore/controllers/app_users/manage_organizations.py	Migrates list/delete operations to select() and Session.get().
tethysext/atcore/controllers/app_users/manage_organization_members.py	Migrates org/user lookups to Session.get().
tethysext/atcore/controllers/app_users/add_existing_user.py	Migrates org/user listing to Session.get() and select() + scalars().
install.yml	Bumps SQLAlchemy requirement to >=2, adds GeoAlchemy2 >=0.13 and condorpy.
README.md	Updates dev/test setup docs (venv-based install and Docker PostGIS workflow).
.gitignore	Ignores local venv and .remember artifacts.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The SQL built for SRID lookup is invalid for PostgreSQL/SQLAlchemy and bypasses parameter binding: it formats srid directly into the string and compares it to @srid (which is not a valid bind parameter in SQLAlchemy/Postgres). This will produce a syntax error (or at best a query that can’t use indexes) and also creates SQL injection risk. Use a text() statement with a :srid bind param (e.g., ... WHERE srid = :srid) and pass {"srid": srid} to execute(); ideally coerce/validate srid to int first.

[Copilot]

Same issue as above in get_wkt_by_srid(): the query interpolates srid into the SQL string and still uses @srid, which isn’t a valid SQLAlchemy bind parameter on Postgres. Switch to a text() query with :srid and pass the parameter dict to execute() (and validate/coerce srid to an int).

[Copilot]

get_spatial_reference_system_by_query_string() builds a raw SQL string via .format(sql_query_input) and then executes it as text(...). This is SQL-injection prone (query terms come from user input) and also prevents the DB from caching query plans. Use bind parameters for the tsquery input (e.g., to_tsquery('english', :q)) and pass the value via execute(..., {"q": sql_query_input}); if you need operators like &, build the query string separately but still bind it.

[Copilot]

The SRID is interpolated directly into the SQL string (... WHERE srid = {}) and then executed via text(sql). Even if srid is typically an int, this bypasses bind parameters and is injection-prone if srid ever comes from request/user input. Prefer text('... WHERE srid = :srid') and pass {"srid": srid} to execute() (or build the statement with SQLAlchemy Core).

[Copilot]

Same parameter-binding issue in get_projection_string(): SRID is formatted into the SQL string before executing. Use a bind param (:srid) rather than string formatting so the DB can safely parameterize/cache the statement and avoid injection risk.

[Copilot]

These columns use mutable objects as default values (default={} and default=['get', ...]). With SQLAlchemy this can lead to shared mutable state between instances/rows and surprising mutations. Prefer callables for per-row defaults (e.g., default=dict and default=lambda: ['get', 'post', 'delete']).

Suggested change

	kwargs: Mapped[Optional[dict]] = mapped_column(PickleType, default={})
	http_methods: Mapped[Optional[list]] = mapped_column(PickleType, default=['get', 'post', 'delete'])
	kwargs: Mapped[Optional[dict]] = mapped_column(PickleType, default=dict)
	http_methods: Mapped[Optional[list]] = mapped_column(
	PickleType,
	default=lambda: ['get', 'post', 'delete'],
	)

[Copilot]

mapped_column(..., default={}) / default=[] uses mutable objects as defaults for _options, _parameters, and _active_roles. This can cause shared state across instances and hard-to-debug behavior when these dicts/lists are mutated. Use callable defaults instead (e.g., default=dict and default=list).

Suggested change

	_options: Mapped[Optional[dict]] = mapped_column(PickleType, default={})
	_attributes: Mapped[Optional[str]] = mapped_column(String)
	_parameters: Mapped[Optional[dict]] = mapped_column(PickleType, default={})
	_active_roles: Mapped[Optional[list]] = mapped_column(PickleType, default=[])
	_options: Mapped[Optional[dict]] = mapped_column(PickleType, default=dict)
	_attributes: Mapped[Optional[str]] = mapped_column(String)
	_parameters: Mapped[Optional[dict]] = mapped_column(PickleType, default=dict)
	_active_roles: Mapped[Optional[list]] = mapped_column(PickleType, default=list)

[Copilot]

_data and _options are declared with default={}. Using a mutable object as a column default can lead to shared state across instances when mutated. Prefer default=dict (callable) so each row gets a fresh dict.

Suggested change

	_data: Mapped[Optional[dict]] = mapped_column(PickleType, default={})
	_options: Mapped[Optional[dict]] = mapped_column(PickleType, default={})
	_data: Mapped[Optional[dict]] = mapped_column(PickleType, default=dict)
	_options: Mapped[Optional[dict]] = mapped_column(PickleType, default=dict)

[swainn]

@copilot rebase with master and resolve the merge conflicts in this pull request

[Copilot]

@copilot rebase with master and resolve the merge conflicts in this pull request

Merged master into the branch in commit ada074b. The only conflict was in .gitignore (both branches added .venv/ and .remember/ in different orders, and master also added apps/) — resolved by combining all three entries.