Repository navigation
feat(B.8): paridad GUI/servicio/CLI verificada por ejecución + alias canónicos en resultado GUI - #8
Merged
Merged
Conversation
… the GUI result B.8: same input + target + profile must yield the same semantic IR, the same compiled artifacts and the same quality_report from every entry point — verified by execution, not declared (tests/test_parity_b8.py, 7 tests). Hallazgos de la comparación, corregidos: 1. compile_for_gui's result schema had drifted from compile_prompt's canonical one: chosen NSL was only exposed as 'nsl' (not 'chosen_nsl'), the execution prompt only as 'optimized' (not 'optimized_prompt'), and the profile/level metadata (applied_profile, requested_profile, target, requested_level, chosen_level, context_loss_report) was missing from the top level — unreachable for parity checks and exports. Fixed ADDITIVELY: the canonical names are now aliases to the same objects; every legacy key the GUI consumes stays untouched, so no GUI behaviour changes. 2. The CLI already routes through compile_prompt (the canonical entry point); the parity test now proves its exported artifacts (canonical_nsl.nsl, optimized_prompt.txt) are ID-line-identical to the service's, and that the exported hybrid JSON carries the B.2 extended IR (ambiguities/confidence) end to end. Also re-executed the six A.7 critical-path routes against the enriched results (quality_report/four_layers/policy_check/clarification now in every result): all six VERIFIED_BY_EXECUTION. Suite: 304 passed + 10 subtests, ruff clean.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Eje B — B.8
Misma entrada + target + perfil ⇒ mismo IR semántico, mismos artefactos compilados y mismo
quality_reportdesde los tres puntos de entrada — verificado por ejecución (7 tests entests/test_parity_b8.py), no declarado.Hallazgos corregidos
compile_for_gui: el NSL elegido solo se exponía comonsl(nochosen_nsl), el prompt solo comooptimized(nooptimized_prompt), y los metadatos de perfil/nivel (applied_profile,target,chosen_level,context_loss_report…) faltaban en el nivel superior. Fix aditivo: alias canónicos a los mismos objetos; las claves legacy que la GUI consume quedan intactas.compile_prompt(entrada canónica); el test prueba que sus artefactos exportados son idénticos a los del servicio (salvo la líneaID=de traza) y que el JSON híbrido lleva el IR extendido B.2 de extremo a extremo.Verificación final del Eje B
Las 6 rutas críticas de A.7 re-ejecutadas contra los resultados enriquecidos (
quality_report/four_layers/policy_check/clarification): 6/6 VERIFIED_BY_EXECUTION. Suite: 304 passed + 10 subtests, ruff limpio.