Keru/knowledge-genome-orchestrator
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

test(ingest): add tests for run-prune.sh

Browse source
This commit is contained in:
Matteo Cherubini 2026-07-01 19:37:10 +02:00
parent c0659d5ce9
commit 066db00e89
1 changed files with 68 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

68
tests/run-prune.bats Normal file
View file

@ -0,0 +1,68 @@
#!/usr/bin/env bats
# tests/run-prune.bats — prune orphaned sources (no LLM, no network; DRY_RUN).
setup() {
load 'helpers'
export PRUNE="${SKILL_SCRIPTS}/run-prune.sh"
export GENOMES_ROOT="${BATS_TEST_TMPDIR}"
export INGEST_BASE="main"
export KG_LIB_DIR="${LIB_DIR}"
export FORGEJO_URL="http://forgejo.local" FORGEJO_USER="u" FORGEJO_TOKEN="t"
export DRY_RUN=1
g_src="$(make_fixture_genome)"; export g_name="fixture-genome"
mv "$g_src" "${GENOMES_ROOT}/${g_name}"; export g="${GENOMES_ROOT}/${g_name}"
( cd "$g" && rm -f raw/articles/test.md && git add -A && git commit -q -m clear && git push -q )
}
@test "run-prune: removes only the orphaned source + its index entry, opens a dry PR" {
command -v jq >/dev/null 2>&1 || skip "jq not installed"
cd "$g"
# kept: raw exists. orphan: raw missing.
echo content > raw/articles/kept.md
h="$(sha256sum raw/articles/kept.md | cut -d' ' -f1)"
printf -- '---\nsource_path: raw/articles/kept.md\nsource_sha256: %s\n---\nbody\n' "$h" > wiki/sources/kept.md
printf -- '---\nsource_path: raw/articles/gone.md\nsource_sha256: abc\n---\nbody\n' > wiki/sources/orphan.md
python3 "$SKILL_SCRIPTS/index-append.py" --section Sources --entry '- [[sources/kept]] — kept. `maturity: draft`'
python3 "$SKILL_SCRIPTS/index-append.py" --section Sources --entry '- [[sources/orphan]] — orphan. `maturity: draft`'
git add -A && git commit -q -m setup && git push -q
run bash "$PRUNE" "$g_name"
[ "$status" -eq 0 ]
[[ "$output" == *'"status":"ok"'* ]]
[[ "$output" == *'"count":1'* ]]
# only the orphan page is gone
[ ! -f wiki/sources/orphan.md ]
[ -f wiki/sources/kept.md ]
# index reflects the removal
! grep -q 'sources/orphan' wiki/index.md
grep -q 'sources/kept' wiki/index.md
# committed on a chore/ branch (NOT feat/ai-ingest-*)
git rev-parse --verify "chore/prune-orphans-$(date +%F)"
}
@test "run-prune: no orphans -> count 0 and no PR/branch" {
command -v jq >/dev/null 2>&1 || skip "jq not installed"
cd "$g"
echo content > raw/articles/kept.md
h="$(sha256sum raw/articles/kept.md | cut -d' ' -f1)"
printf -- '---\nsource_path: raw/articles/kept.md\nsource_sha256: %s\n---\nbody\n' "$h" > wiki/sources/kept.md
git add -A && git commit -q -m setup && git push -q
run bash "$PRUNE" "$g_name"
[ "$status" -eq 0 ]
[[ "$output" == *'"count":0'* ]]
run git rev-parse --verify "chore/prune-orphans-$(date +%F)"
[ "$status" -ne 0 ]
}
@test "run-prune: refuses when an orphan path would escape wiki/ (defense in depth)" {
command -v jq >/dev/null 2>&1 || skip "jq not installed"
cd "$g"
# legacy page without source_path is ignored; a page with a missing raw is the orphan.
printf -- '---\nsource_path: raw/articles/gone.md\nsource_sha256: abc\n---\nbody\n' > wiki/sources/orphan.md
git add -A && git commit -q -m setup && git push -q
run bash "$PRUNE" "$g_name"
[ "$status" -eq 0 ]
[[ "$output" == *'"count":1'* ]]
[ ! -f wiki/sources/orphan.md ]
}