Enhancement: Re-document & test `output_get()` (Eurostat P1 Output)

[antaldaniel]

Problem

• output_get() currently wraps primary_input_get() (semantic mismatch).
• Roxygen mentions “multipliers” and doesn’t align with Eurostat P1 definition.
• Direct positional indexing ([, 1]) discards column names, may coerce factors oddly, and can trigger unintended recycling.

Proposal

1. Make output_get() a dedicated helper that selects the Output (P1) row
   (accept "output", "output_bp", "P1", "p1") without calling
   primary_input_get().
2. Update roxygen to Eurostat terms:
   • “Output (P1) at basic prices”
   • Clarify it’s not a primary input
   • Return: one-row data frame (key + numeric columns).
3. Add concise tests (testthat 3e):
   • One row returned; first column is key; all non-key cols numeric
   • Label ∈ {output, output_bp, P1, p1}
   • No NAs in returned row.
4. (Optional follow-up) Consider renaming primary_input_get() to
   primary_inputs_get() (or value_added_get()) to avoid confusion.

Acceptance

• Documentation follows rOpenSci markdown style; ≤80-char code lines.
• New tests pass on package demo data.
• No tidyselect warnings.