Repository navigation
Qualify installed migration journeys and document the supported surface #112
Description
User workflow and ownership
Implement the accepted Engineer preparation/migration unit below. Preserve existing native defaults, dataframe contracts and independent use with supplied data. Data owns source adjustments/revision selection; Diagnostic owns the only Alphalens facade/cleaner and analytical workflows; Backtest owns simulation/accounting. Engineer produces native features, eligibility and outcomes, not competing engines.
Dependencies: #107 (precise returns), #108 (factor methods), #109 (dated screens),\n#110 (outcomes) and #111 (optional inputs), for the relevant routes. This is final acceptance across those coherent
submissions, not a requirement for another tiny documentation-only PR. Keep executable migration
examples and reference pages with the implementing changes. Final qualification may be an evidence
checkpoint without additional source changes. New public records are searched/reused at execution.
Deliver executable native and compatibility versions of these user journeys:
- Wide prices to precise grouped returns, preserving documented leading-row/container differences.
- Actual Zipline RSI top/bottom-three selection plus a momentum/volatility/liquidity/sector screen,
using dated membership, warmup, a split, rename and late data. Produce standard long signals. - A custom factor/filter/classifier calculation ported to registered features/Polars expressions,
including multi-output selection, masks and a declared unavailable input. - Factor keys and prices to cumulative/noncumulative multi-horizon outcomes, with an optional
MultiIndex adapter and the same output supplied directly to a receiver without Engineer. - Legacy/native market, sector and loading inputs to validated optional long tables, including
missing data and split-compatible units. No automatic NAV-change returns across unknown flows.
For each journey publish an old-to-new call/argument/container table, exact supported options,
definition differences, data requirements, unsupported framework APIs and a runnable before/after
example. Explain when an import change suffices and when users must replace framework callbacks
with prepared tables. Provide operation-specific errors that name the missing declaration or
unsupported method. No global compatibility flag, silent approximation, automatic provider access
or forced install of Reloaded dependencies. Link existing consumer migration guides instead of
writing parallel analytics or trading documentation in Engineer.
Record a versioned acceptance manifest connecting every relevant legacy operation and all 27
assigned rows to tests or an explicit external-owner handoff. Capture valid external inputs so
Engineer work is not stalled by other owners. Engineer standalone completion requires passing
native and adapter preparation with exact expected boundary artifacts. Actual cross-library
integration acceptance separately requires those real producer artifacts to pass the public
consumer boundary. Do not mark external integration passed from a stub or claim another library's
missing analytics as an Engineer implementation. An unresolved receiver contract blocks only that
specific integration claim and has a named exact input/output discrepancy.
Proposed focused files: tests/integration/test_reloaded_migration.py and executable examples
under examples/; update existing feature/labeling/installation guides and API navigation rather
than fragmenting docs. Verify with uv run pytest tests/integration/test_reloaded_migration.py -q,
then the full source quality commands and configured final gates described above. Build wheel and
sdist, retain exact commit/tree/dependency identities and SHA256 hashes, install into isolated
disk-backed environments, verify imported origins/bytes and run the recipes outside source shadow.
Test the actual base dependency environment and optional conversion environments. pandas is already a mandatory Engineer dependency; this unit must not claim or invent a no-pandas installation profile. A successful source import is not an
installed artifact check. Any failing revision is repaired and relevant gates rerun at final head.
Completion requires no unclassified relevant operation, passing positive and negative workflow
tests, current-head native/default compatibility and platform gates, strict executable docs, and
a clean/backed-up worktree. Distinguish planned, implemented, merged, installed-qualified and
published versions. Publication/deployment requires a separate decision.
Premises
Input: installed migration artifacts
| Question | Answer | Kind |
|---|---|---|
| record | An immutable package artifact plus versioned recipe inputs, method declarations and expected keyed outputs. | definitional |
| absent-record | A missing artifact/test/option disposition is missing evidence, not a passing migration or supported omission. | definitional |
| ordering-spacing | Recipes retain the declared time/grid units and compare intermediates before final values. | definitional |
| entity-identity | Exact artifact hashes and stable data keys identify the implementation and observations under test. | definitional |
| known-at | Causal examples retain information-set constraints; realized labels remain separate from tradable signals. | definitional |
| institution | Engineer's actual source support/configured gates and named reference profiles govern claims; publication needs separate authority. | definitional |
Claims
| Id | Claim | How it could be false |
|---|---|---|
| installed-migration | Every required Engineer-owned journey runs from the built wheel without Reloaded runtime dependencies. | An example imports source or legacy code, loses metadata or requires an unsupported option. |
| coverage-accountable | Every relevant operation and original assigned row has test evidence or a truthful ownership disposition. | A method/option is omitted, unsupported essential work is excluded, or a generic CI result is used as parity proof. |
Closes: installed-migration-artifacts.record, installed-migration-artifacts.absent-record, installed-migration-artifacts.ordering-spacing, installed-migration-artifacts.entity-identity, installed-migration-artifacts.known-at, installed-migration-artifacts.institution, installed-migration, coverage-accountable
Intake and completion
- Reproduce public producer defects before repairs and retain exact failing/passing revisions.
- Complete the methods, negative/error cases, runnable tests and migration route above.
- Verify actual installed outputs and the applicable final-revision quality/platform gates.
- Record producer/consumer dependencies separately from standalone qualification.
Plan reference: accepted Engineer preparation and Reloaded migration plan, unit 6 in milestone unreleased - Engineer preparation and Reloaded migration.
Completion evidence: #116 at reviewed head 254905b723c5d96a385f150bc0bcd4ce3d25d1e0 passed 5,670 native and 5,670 no-JIT tests, critical coverage, strict docs, types, lint and hooks. All 49 hosted contexts completed with 48 successes and one configured exception-validator skip, including all 12 installed/native/shared matrix cells and final Build Package. Actual built-wheel/source-archive and identified producer/consumer journeys passed. These qualify Engineer preparation; native Diagnostic capacity/supplied-loading analytics and whole-portfolio replacement are not claimed.
AI assistance: Codex assisted with implementation, regressions, documentation, installed qualification and these updates. The source, independent expectations and actual workflow results were reviewed in the owning maintainer session.
Activity
fix(packaging): include runnable migration recipes and documentation …
feat(migration): prepare optional inputs and qualify installed workfl…