docs(miner-guide): note that find_product q matches values, not keys (ORO-1927)

[that-guy-wade]

Description

One-line clarification in docs/miner-guide.md on what find_product's q matches. Miner (@boodaoutd on Discord) asked; docs never said.

Issue Link

• Closes ORO-1927

Changes Made

• docs/miner-guide.md: one-sentence note under the Local Testing options table.

[shardi-b]

Review: PR #254 — docs(miner-guide): note that find_product q matches values, not keys

Scope: Documentation-only. The diff adds two lines to docs/miner-guide.md (one sentence + blank line), inserted after the CLI options table:

Note on find_product: the q parameter matches against product title and the values within attributes / sku_options. Field names (keys) themselves are not searchable.

gh pr view confirms: 1 file, +2/-0, no deletions.

Verification performed:

• Traced the search path in src/search_engine/server.py:135 (search) → BM25 via searcher.search(q=...) over the Lucene index. The searchable text is the JsonCollection contents field.
• The index is built externally by docker/index-builder/Dockerfile (pyserini.index.lucene --collection JsonCollection --input /data/resources ... DefaultLuceneDocumentGenerator). The construction of the contents field is not present anywhere in this repo (grep for contents/documents.jsonl builders returns only consumers/tests), so the precise set of BM25-searchable fields is defined by external data prep and cannot be confirmed or refuted from the code under review.
• server.py:78-85: SEARCH_FIELDS includes title; INFORMATION_FIELDS includes short_description, description, sku_options, attributes. These control filtering/returned fields, not the BM25 contents text, so they neither prove nor disprove the sentence.
• No code, callers, or behavior are changed by this diff.

Result: No correctness, reuse, simplification, efficiency, altitude, or CLAUDE.md-convention finding reaches blocking severity. The only observations are sub-blocking editorial nits — the note sits under the test-runner CLI-flags table rather than beside the find_product discussion, and it omits description/short_description which may also be indexed (unverifiable in this repo). Neither is a defect worth blocking on. Approving.