PDL-literal agent counts (n=1) and wholesaler passthrough

[Hackatro]

Context

Already discussed: the simulation should follow the PDL completely instead of inventing population sizes and wholesaler behaviour the PDL does not contain.

Do not implement on #33. #33 is a byte-for-byte identity refactor. This issue changes volumes and wholesaler economics. Start after #33 is merged (or after Minu explicitly unfreezes behaviour).

Vault write-up (for other agents): Projects/Work/OFFIS-PROVIDER/Tasks/pdl-literal-counts-passthrough-wholesaler.md

This reverses #33 spec §4.3 ("wholesalers are real actors") for this follow-up only. #33 itself still lands the current wholesaler until that gate is done.

Decision 1 — count = PDL entities, default 1

Today one PDL entity (brazil_farms) is instantiated as 10 Farmer agents (n_brazil_farms=10; ARG=5, US=8). That count is not in the PDL.

• One PDL entity → one agent, unless the PDL lists more of the same agent.
• No count field → 1.
• Total volume will drop (Brazil ~10×). Accepted.
• size_sigma becomes inert at n=1. Do not add a sidecar count: as a permanent substitute.

Decision 2 — wholesaler is a passthrough (Brazil split stays)

• Wholesaler = middleman that passes volume through to its sea lane(s).
• Argentina / US: one wholesaler → one outbound lane.
• Brazil exception: one wholesaler, two ports (Santos + Paranaguá). Keep a share split (share_{port} / equal-split fallback).

Likely files

• scenario.py + both CSVs — stop using invented n_{eid} population columns
• topology.py count_attr / build_roster
• agents/trader.py — passthrough; Brazil split only
• agents/farmer.py — size_sigma unused at n=1
• Re-baseline cascades; report volume/price deltas (not byte-for-byte)

Out of scope

• Collapse BRA/ARG/USA producers into one PDL-derived producer archetype #33 producer-identity collapse, sidecar, fertilizer-from-deps, output-schema rename
• Adding a new PDL count: field without partner agreement