Simplify

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.

[kennethshsu]

Summary of Changes

refactor to reuse the Development class code, PR into #932_triangle_weight, not main

Related GitHub Issue(s)

#932

Additional Context for Reviewers

X=X * X.w_ doesn't break it? The parameter name is factor?
Does this refactor work? Passing all tests.

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Weight logic now runs through DevelopmentBase, so any subtle differences from the removed local helpers could change exclusion behavior even though tests pass.

Overview
TriangleWeight now subclasses DevelopmentBase instead of sklearn mixins alone, and drops the large duplicated weight-building implementation in favor of the shared _set_weight_func pipeline on the development base class.

fit still applies existing triangle weights when present (X * X.w_), but passes the input as the factor keyword expected by the base helper rather than a positional X. _cascade_param is a thin wrapper around _param_array_helper.

Public API and docstring behavior (including the documented drop-operation order that differs from Development) are unchanged; this is primarily deduplication against chainladder/development/base.py.

^{Reviewed by Cursor Bugbot for commit 018920c. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.48%. Comparing base (9a9a796) to head (018920c).

Additional details and impacted files

@@ Coverage Diff @@## #932_triangle_weight #989 +/- ##
========================================================
- Coverage 88.53% 88.48% -0.05% 
========================================================
Files 89 89 Lines 5189 5081 -108 Branches 658 644 -14 ========================================================
- Hits 4594 4496 -98 + Misses 442 437 -5 + Partials 153 148 -5 

Flag	Coverage Δ
unittests	88.48% <100.00%> (-0.05%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kennethshsu]

@henrydingliu can you take a look at this?

[henrydingliu]

@kennethshsu i'll have to refer you to the "Additional Context" section of the #941. Keeping the two sets of weight-assigning functions that are currently in DevelopmentBase doesn't help us get to a simplier DevelopmentBase. We have to lift one of them out, so we can eventually clean up DevelopmentBase entirely.