[bfs2021]

Problem

sgm-stats counts a wikilink written inside inline code as a real graph edge, so a note that only mentions [[x]] as an example still links to x. Fixes #72.

Root cause

_links_in in brain_ask.py removes fenced code blocks before extracting wikilinks, but never removes inline code spans. Any `[[x]]` outside a fence survives the first pass and is picked up by WIKILINK_RX, which adds an edge that does not exist in the vault graph.

Fix

Added INLINE_CODE_RX = re.compile(r'+[^\n]*+')and strip inline spans right after fenced blocks in_links_in`. Fenced blocks are removed first, so the remaining backticks are inline spans only; multi-backtick spans are consumed as a unit. The docstring now says "code ignored" instead of "code fences ignored".

Testing

• New case in tests/test_brain_ask.py: Write \[[example]]` to link.` yields no links; a real link next to inline code still resolves.
• New test_link_in_inline_code_is_not_an_edge in tests/test_vault_stats.py builds the vault from the issue and asserts build_graph returns no edges.
• Both fail on the previous code (3 failed) and pass with the fix.
• python -m pytest -q: 210 passed, 1 skipped.

[tonydzi]

Hi, Mycroft here, Anton's synthetic AI co-founder. I read this before the human did, which is the one race I always win.

Logged your pull request ("fix: ignore wikilinks inside inline code") after 2 h of silence, which is on us. What a machine can already tell: 3 file(s) touched, +15/-1, CI no checks. A human from the lab reads it within 48 hours, as CONTRIBUTING.md promises; if that clock runs out, ping the thread, it is not rude.

— TonyDzi, Palo Alto AI Research Lab · more of the machine this came from (second brain + fleet coordination) lives at github.com/tonydzi, DMs open.