Add TapasXq::DeriveReaderService to generate reading interface XHTML

[candyhazlett]

Background

The tapas-xq endpoint POST /tapas-xq/derive-reader/:type transforms a TEI file into an XHTML snippet (a <div>) for display in the TAPAS reading interface. This is the core reader functionality and is not yet implemented in Rails.

Work required

app/services/tapas_xq/derive_reader_service.rb

Create with:

• Constructor: view_package_type:, tei_content:, extra_params: {}, client: nil
• #derive — POSTs to /derive-reader/#{view_package_type} with { file: tei_content }.merge(extra_params), returns the XHTML string

Error handling:

• TapasXq::InvalidTeiError (422) — raised by client.rb when tapas-xq rejects the TEI file
• TapasXq::InvalidResponseError — raised on 400 (unknown view package type); callers should validate the type against the ViewPackage table before calling
• Other TapasXq::Error subclasses propagate normally

The extra_params argument supports view-package-specific parameters (each package may define additional form parameters in its configuration file).

Controller / job integration (to be designed)

The derive-reader call is stateless — tapas-xq does not store anything. It can be called:

• On demand per request (acceptable for small files, but may be slow)
• As a background job with result cached on CoreFile (e.g., a reader_xhtml column) and invalidated on TEI update

The integration point (controller action, background job, or caching strategy) is left to the implementer to decide based on performance requirements.

Dependencies

• TapasXq::ViewPackagesService must be implemented first so that valid view_package_type values are known
• Issue Handle 422 from tapas-xq as InvalidTeiError with descriptive message #51 (422 → InvalidTeiError) must be resolved

Tests

See spec/services/tapas_xq/derive_reader_service_spec.rb in the testing plan doc.