Ingest a harness's semantic assessments

The second half of the review-packet contract: reads the CSV a harness wrote against a packet from write_semantic_review_packet(), in the frozen 30-column assessment row, and turns it into the package's record. It refuses a selection whose IRI the packet did not offer (recorded as an error row, never applied), runs the retry bookkeeping, escalates a final reject_shortlist to request_new_term so the ontology gap is surfaced, runs the existing bundle validators, merges the result into the suggestions and persists it under review/, so that semantic_llm_assessments() returns it for a package path.

Usage

ingest_semantic_assessments(
  x,
  assessments = NULL,
  packet = NULL,
  packet_id = NULL,
  provider = NULL,
  model = NULL,
  search_fn = find_terms,
  review_dir = NULL,
  quiet = FALSE
)

Arguments

x

The package directory or the in-memory dictionary the packet was built from.

assessments

The harness's file: a CSV path or a data frame. Defaults to review/semantic-assessments-pass-<n>.csv for the packet's pass.

packet

The packet the file answers: a path or a parsed packet. Defaults to the latest pass in review/.

packet_id

Optional. The packet_id this ingest expects; a different packet is refused with code packet_mismatch.

provider, model

Optional overrides for the llm_provider and llm_model columns of every row.

search_fn

The search function used for a retry. Defaults to find_terms().

review_dir

Where the session lives. Defaults to review/ under the package directory; required for in-memory input.

quiet

Logical. Suppress the summary message.

Value

Invisibly, a list: status ("complete" or "awaiting_pass_2"), pass, packet_id, next_packet (the continuation packet's path, or NULL), assessments (the record, with the validator findings as its semantic_validator_findings attribute), findings, suggestions, targets, dictionary (with semantic_suggestions, semantic_targets and semantic_llm_assessments attached, so detect_semantic_term_gaps() and the accessors work unchanged), and summary (counts of decisions, errors, downgrades, escalations and retries).

Details

A retry is a second harness pass. A retry_search with a usable query is retrieved here through search_fn, and when the search widens the shortlist a continuation packet (review/semantic-review-packet-pass-2.json) is written holding the widened shortlist; nothing from that target is merged or escalated until the harness has answered it. Calling this again after the harness answers the second packet completes the session. There is no third pass.

This never calls a model, and it reaches the network only through search_fn, for a retry. It never touches the metadata CSVs: applying a choice stays review_semantics() -> accept_suggestion() -> apply_sdp_semantics().

A file-level problem aborts the ingest and writes nothing, with a stable code carried as the condition's code field and class metasalmon_semantic_review_<code>: packet_version, packet_integrity (the packet no longer matches its packet_id), packet_mismatch, header, unknown_target, duplicate_target, provenance and no_pass_2. A row-level problem makes that row an error row or downgrades it with a note, and processing continues.

See also

write_semantic_review_packet(), semantic_llm_assessments()

Examples

if (FALSE) { # \dontrun{
# Run in checks once a harness response fixture is bundled.
# Start with an SDP whose semantic targets are ready for review.
sdp_path <- "/path/to/reviewed-sdp"
packet <- write_semantic_review_packet(sdp_path)
# Have the external harness judge packet$path and write its 30-column
# response and a matching .packet-id sidecar in the SDP's review directory.
response <- file.path(sdp_path, "review", "semantic-assessments-pass-1.csv")
result <- ingest_semantic_assessments(
  sdp_path, assessments = response, packet_id = packet$packet_id
)
if (identical(result$status, "awaiting_pass_2")) {
  # The harness judges result$next_packet, then ingest that second response.
  second_response <- file.path(
    sdp_path, "review", "semantic-assessments-pass-2.csv"
  )
  result <- ingest_semantic_assessments(
    sdp_path, assessments = second_response
  )
}
} # }