From c0659d5ce9d9690c826204d855c13172b0eb9d33 Mon Sep 17 00:00:00 2001
From: Matteo Cherubini <cherubinimatte@gmail.com>
Date: Wed, 1 Jul 2026 19:37:10 +0200
Subject: [PATCH] feat(ingest): introduce run-prune.sh for orphaned source
 removal

---
 skills/ingest/scripts/run-prune.sh | 96 ++++++++++++++++++++++++++++++
 1 file changed, 96 insertions(+)
 create mode 100755 skills/ingest/scripts/run-prune.sh

diff --git a/skills/ingest/scripts/run-prune.sh b/skills/ingest/scripts/run-prune.sh
new file mode 100755
index 0000000..6cf22f2
--- /dev/null
+++ b/skills/ingest/scripts/run-prune.sh
@@ -0,0 +1,96 @@
+#!/usr/bin/env bash
+# =============================================================================
+# skills/ingest/scripts/run-prune.sh
+# Symmetric companion to run-ingest: prune source pages whose raw source no
+# longer exists. RE-DERIVES the orphan set itself (mirrors orphan-wiki.sh) — it
+# never trusts a list handed in by n8n, so there is no "detected-vs-pruned"
+# race. Removes ONLY the pages it derived plus their index entries, commits
+# ONLY wiki/ on chore/prune-orphans-<date>, and opens a GATED removal PR (the
+# operator approves the deletion; principle 2). Never deletes of its own accord.
+#
+# Runs OUTSIDE the model, on vm101, cwd = genome checkout. The wrapper (`pi
+# prune`) has already taken the per-genome lock and done clean_start, exactly
+# like `pi ingest` — so this script does neither.
+#
+#   run-prune.sh <genome>
+#
+# Emits a single JSON result line on stdout for n8n to parse.
+# =============================================================================
+set -euo pipefail
+
+genome="${1:?usage: run-prune.sh <genome>}"
+SCRIPTS="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+fail() {
+  jq -nc --arg stage "$1" --arg reason "$2" '{status:"error", stage:$stage, reason:$reason}'
+  exit 1
+}
+
+command -v jq      >/dev/null 2>&1 || { echo '{"status":"error","reason":"jq missing"}'; exit 1; }
+command -v python3 >/dev/null 2>&1 || fail "deps" "python3 missing (needed by index-append.py)"
+
+# --- re-derive orphans (same rule as orphan-wiki.sh; computed fresh, here, now) ---
+# A wiki/sources/*.md page is orphaned when its frontmatter source_path points at
+# a raw file that no longer exists. Legacy pages without source_path are ignored.
+declare -a ORPH=()
+for page in wiki/sources/*.md; do
+  [[ -e "$page" ]] || continue
+  sp="$(sed -n 's/^source_path:[[:space:]]*//p' "$page" | tr -d '\r' | head -n1)"
+  [[ -n "$sp" ]] || continue
+  [[ -f "$sp" ]] || ORPH+=("$page")
+done
+
+if [[ ${#ORPH[@]} -eq 0 ]]; then
+  jq -nc '{status:"ok", count:0, pruned:[], detail:"no orphans"}'
+  exit 0
+fi
+
+# --- remove each orphan page + its index entry (anti-traversal, wiki/-only) ---
+declare -a PRUNED=()
+for page in "${ORPH[@]}"; do
+  case "$page" in
+    wiki/*) : ;;
+    *) fail "prune" "refusing to remove outside wiki/: ${page}" ;;
+  esac
+  case "$page" in *..*) fail "prune" "path traversal in page: ${page}" ;; esac
+  [[ -f "$page" ]] || continue
+  rm -f "$page"
+  link="${page#wiki/}"; link="${link%.md}"            # e.g. sources/foo
+  python3 "${SCRIPTS}/index-append.py" --remove "$link" \
+    || fail "index" "index-append --remove failed for ${link}"
+  PRUNED+=("$link")
+done
+
+# --- assemble the PR body ---
+date_tag="$(date +%F)"
+body="$(mktemp)"
+trap 'rm -f "$body"' EXIT
+{
+  echo "## Prune orphaned sources"
+  echo ""
+  echo "These source pages reference a \`source_path\` whose raw file no longer exists"
+  echo "in \`raw/\`. Removing them keeps the wiki in sync with git (the source of truth)."
+  echo ""
+  echo "| Removed page |"
+  echo "|--------------|"
+  for l in "${PRUNED[@]}"; do echo "| \`wiki/${l}.md\` |"; done
+} > "$body"
+
+# --- open the GATED removal PR on a chore/ branch (open-pr --branch override) ---
+branch="chore/prune-orphans-${date_tag}"
+pr_out="$( bash "${SCRIPTS}/open-pr.sh" \
+  --branch "$branch" \
+  --title "chore: prune ${#PRUNED[@]} orphaned source(s)" \
+  --body-file "$body" --base "${INGEST_BASE:-main}" 2>&1 )" && pr_rc=0 || pr_rc=$?
+pr_url="$(printf '%s\n' "$pr_out" | sed -n 's/^PR opened: //p' | head -n1)"
+
+# --- result line for n8n ---
+jq -nc \
+  --arg status "$([[ $pr_rc -eq 0 ]] && echo ok || echo pr_failed)" \
+  --argjson count "${#PRUNED[@]}" \
+  --arg pr_url "$pr_url" \
+  --arg detail "$pr_out" \
+  --argjson pruned "$(printf '%s\n' "${PRUNED[@]}" | jq -R . | jq -s .)" \
+  '{status:$status, count:$count, pr_url:$pr_url, pruned:$pruned, detail:$detail}'
+
+[[ $pr_rc -eq 0 ]] || exit 1