This reference defines how to build a frontend-to-backend or cross-layer interface mapping page.
Use this page when the goal is to make request chains and interface ownership easier to understand.
Typical targets:
- H5 / frontend request layer to backend route layer
- app request declaration to backend API layer
- page action to API wrapper to backend handler
- bridge call to native handler to downstream request
Use this page when any of these are true:
- the user asks whether interface paths are clear
- the repo contains both request declarations and backend handlers
- cross-team collaboration depends on knowing who owns which interface
- communication mapping alone is not enough to show caller-handler correspondence
Recommended sections:
- current mapping boundary
- interface mapping table
- unresolved interfaces
- confidence and evidence notes
Recommended table:
| caller side | request path | method | wrapper / sdk / api fn | handler side | mapping status | evidence level | code locations |
| --- | --- | --- | --- | --- | --- | --- | --- |Examples of caller side:
- page component
- frontend api module
- app service interface
- bridge wrapper
- python client module
Examples of handler side:
- backend controller
- FastAPI router
- gateway route
- callback handler
- native handler
Use clear labels such as:
direct matchpath match, handler not closedwrapper visible, route unresolvedcontract visible, provider not foundpartial match
For each important row, provide:
- code-visible request declaration or invocation
- code-visible handler declaration if found
- evidence level
- exact code locations when practical
Do not claim a direct mapping unless both sides are visible or a clearly defensible closure path exists.
Useful matching signals include:
- same HTTP method and path
- same route prefix plus handler suffix
- wrapper function naming plus imported backend contract
- gateway path forwarding config
- shared constant or route registration source
These are hints, not automatic proof.
If only the caller side is visible:
- record the contract
- mark the provider as unresolved
- keep the evidence level aligned with what is actually visible
If only the handler side is visible:
- record the handler
- mark the caller as unresolved
- do not invent a frontend or app owner
Use phrases like:
Caller-side contract visibleHandler-side route visibleDirect match confirmed by method + pathOnly wrapper layer visible in current repositoryProvider not closed in current scope