FINERACT-2649: batch migrate Tier 4 loan tests to feign

[DeathGun44]

What this PR does

fourth batch of the loan integration-test migration. It moves 54 test classes (655 test methods) off REST-assured onto the typed Feign client.

None of those 54 classes reference REST-assured any more. With this merged, 151 integration-test classes run on the typed client.

99 files changed:

• 54 test classes (655 test methods)
• 2 shared base classes (FeignLoanTestBase, BaseLoanIntegrationTest)
• 34 shared Feign helpers and builders (16 of them new)

The tests assert the same things as before. This changes how they talk to the server, not what they check. The few assertions that did change are listed under "Bugs found" and each one is a fix.

Commits

26 commits, each a group of related tests (loan charges, COB, external asset owner, delinquency, advanced payment allocation, and so on).

What's covered

• Client loan lifecycle - charges of every calculation type, credit balance refund, external IDs, backdated disbursement, overpaid status, modifying approved amounts, floating-rate interest recalculation, date validation.
• Close of Business (COB) - partitioning, catch-up, inline COB, account locks.
• External asset owner - transfer, cancel, search, initiate/undisbursed.
• Advanced Payment Allocation - repayment schedules, interest refunds, charges, reprocessing.
• Charges - specific due date, after maturity, instalment fees, penalties, charge-off accounting.
• Delinquency - buckets, business events, classification.
• Batch API, scheduler jobs, journal reversal, SQL-injection reporting.

Bugs found

The typed client is stricter than REST-assured, so a few real problems showed up:

• Wrong charge calculation type. ChargeRequestBuilders had CHARGE_CALCULATION_TYPE_PERCENTAGE_AMOUNT_AND_INTEREST = 4, but 4 is percent-of-interest. Percent-of-amount-and-interest is 3, which is what ChargesHelper uses and what the tests meant. Every test using that builder was charging on the wrong base and still passing, because the amounts were plausible. Now uses ChargeCalculationType.PERCENT_OF_AMOUNT_AND_INTEREST.
• A comparison that could never be true. In testLoanCharges_INSTALMENT_FEE the instalment numbers are Long (the request model requires it) but the schedule reports its period as an Integer. Long.equals(Integer) is always false, so the "this is the waived instalment" and "this is the paid instalment" branches never ran and the test compared against the wrong expected values. Now compared numerically.
• A crash hidden in a helper. modifyLoanApplication passed command through Map.of(...), which rejects nulls. Three charge tests in ClientLoanIntegrationTest pass null here and were throwing NullPointerException before reaching the server. Now uses the typed query parameter, which drops a null.
• A dropped product setting. LoanProductTestBuilder.withDisallowExpectedDisbursements(true) also quietly turned on a 100% over-applied disbursement allowance. The migrated builder set it to false, which broke chargeOff because it disburses two 1000 tranches against a 1000 approval. Restored.
• A COB step missing its prerequisite. CobPartitioningTest registered the external-asset-owner COB step but had lost the ASSET_TRANSFER GL mapping that step needs. The step list is tenant-wide and never restored, so this broke an unrelated later test.
• Weak assertions. The SQL-injection tests caught bare RuntimeException as "the server rejected this", so a client-side connection failure counted as a pass. They also looked for "404" in the exception message, which includes the request path, so a report whose ID contained 404 hid a real failure. Both now check the HTTP status.
• Interest calculation changed. One charges test had moved from SAME_AS_REPAYMENT_PERIOD to DAILY. Reverted.
• Locale changed. Account transfers were posted with en instead of en_GB. Restored.
• A reversal check that never fired. verifyTransactions compared manuallyReversed with Objects.equals(item.getManuallyReversed(), tr.reversed), but the API omits that field entirely for non-reversed transactions, so the comparison was Objects.equals(null, false) - always false, making the whole predicate unreachable. The check now treats null and false as the same state, and per review feedback it exists in both validators: LoanTestValidators (Feign) and BaseLoanIntegrationTest (REST-assured), where there was previously no check at all.

Shared test infrastructure

• LoanChargeCommandsApi and ClientChargeCommandsApi (new). Two hand-written Feign nterfaces, bound to the generated response models, for the three request shapes the generated request models cannot express: a bodyless ?command=waive POST (the generated client always serialises at least {}, which that endpoint answers with a 500), a locale-formatted decimal amount ("50,05"), and a client-charge amount with decimal places. No JSON is built by hand. Retyping the shared models instead was rejected: PostLoansLoanIdChargesRequest.amount has 118 call sites passing a number, and PostClientsClientIdChargesRequest.amount is an Integer whose correction is a breaking OpenAPI change (swagger-brake R010) that belongs in its own PR.
• LoanRequestBuilders - added a type-safe paymentAllocationOrder(PaymentAllocationType...) overload, replacing copies of the same private helper in individual tests.
• ChargeRequestBuilders - dropped its three local CHARGE_CALCULATION_TYPE_* constants. All call sites now use the ChargeCalculationType enum, so the name and the value can't drift apart again.

Checklist

Please make sure these boxes are checked before submitting your pull request - thanks!

• Write the commit message as per our guidelines
• Acknowledge that we will not review PRs that are not passing the build ("green") - it is your responsibility to get a proposed PR to pass the build, not primarily the project's maintainers.
• Create/update unit or integration tests for verifying the changes made.
• Follow our coding conventions.
• Add required Swagger annotation and update API documentation at fineract-provider/src/main/resources/static/legacy-docs/apiLive.htm with details of any API changes
• This PR must not be a "code dump". Large changes can be made in a branch, with assistance. Ask for help on the developer mailing list.
• If merging this PR resolves a JIRA issue, I will mark that issue as resolved and set "Fix Version/s" appropriately.

Your assigned reviewer(s) will follow our guidelines for code reviews.

[budaidev]

I don't see any larger issues, it looks good.

[DeathGun44]

Hi @adamsaghy can you please take a look at this one ,when you have a moment!

[DeathGun44]

Rebased ,also removed feignhttp calls as per the reviews on #6194 and updated pr description reflecting the current state

[DeathGun44]

test seems flaky, should pass on rerun

[DeathGun44]

@adamsaghy please take a look when you have a moment!

[budaidev]

please resolve the conflict

[adamsaghy]

@DeathGun44 Sorry for being slow. Let me review it now

[adamsaghy]

@DeathGun44 Ultimately it looks good. Great job (huge PR..).

Can you please review my concerns and advise on them?

[DeathGun44]

@adamsaghy please take a look when you have a moment!

[DeathGun44]

@adamsaghy Done!

[DeathGun44]

the failed test seems flaky

[adamsaghy]

LGTM