hyperguild/ingestion/internal/watcher/watcher_wing_test.go

package watcher

import (
	"context"
	"encoding/json"
	"os"
	"path/filepath"
	"regexp"
	"strings"
	"sync/atomic"
	"testing"

	"github.com/stretchr/testify/assert"
	"github.com/stretchr/testify/require"

	"github.com/mathiasbq/hyperguild/ingestion/internal/pipeline"
	"github.com/mathiasbq/hyperguild/ingestion/internal/wing"
)

func wingTestCfg(t *testing.T, mk func(brainDir string) wing.Placer, llmCalls *atomic.Int32) (Config, string) {
	t.Helper()
	brainDir := setupBrainDir(t)
	cfg := Config{
		BrainDir: brainDir,
		Pipeline: pipeline.Config{
			Schema: "# Schema",
			Complete: func(ctx context.Context, s, u string) (string, error) {
				llmCalls.Add(1)
				return successComplete(pipeline.RawPage{Title: "Talk", Type: "source", Subtype: "video", Content: "## Summary\n\nx\n"})(ctx, s, u)
			},
		},
	}
	if mk != nil {
		cfg.Resolver = mk(brainDir)
	}
	return cfg, brainDir
}

// router is the production Router with a client-content marker.
func router(brainDir string) wing.Placer {
	return wing.Router{BrainDir: brainDir, Default: "web-clip", ClientBlock: regexp.MustCompile(`(?i)acme-client`)}
}

func makeWing(t *testing.T, brainDir, name string) {
	t.Helper()
	require.NoError(t, os.MkdirAll(filepath.Join(brainDir, "wiki", name, "facts"), 0o755))
}

func put(t *testing.T, brainDir, rel, body string) {
	t.Helper()
	full := filepath.Join(brainDir, "raw", rel)
	require.NoError(t, os.MkdirAll(filepath.Dir(full), 0o755))
	require.NoError(t, os.WriteFile(full, []byte(body), 0o644))
}

func has(brainDir, rel string) bool {
	_, err := os.Stat(filepath.Join(brainDir, rel))
	return err == nil
}

func mustJSONPages(pages []pipeline.RawPage) string {
	b, err := json.Marshal(pages)
	if err != nil {
		panic(err)
	}
	return string(b)
}

func parkedBody(t *testing.T, brainDir, name string) string {
	t.Helper()
	b, err := os.ReadFile(filepath.Join(brainDir, "raw", "needs-review", name))
	require.NoError(t, err)
	return string(b)
}

func linesStartingWith(s, prefix string) int {
	n := 0
	for _, l := range strings.Split(s, "\n") {
		if strings.HasPrefix(l, prefix) {
			n++
		}
	}
	return n
}

func TestProcessDir_ASubfolderNamesTheWing(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "radar")
	put(t, brainDir, "radar/dark-factory.md", "A talk about factories.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.True(t, has(brainDir, "wiki/radar/sources/dark-factory.md"), "lands in the named wing")
	assert.False(t, has(brainDir, "raw/radar/dark-factory.md"), "inbox zero holds for subfolders too")
}

func TestProcessDir_ADeliberateSubfolderIntoAnyExistingWingPlaces(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "rikatillsammans")
	put(t, brainDir, "rikatillsammans/meeting.md", "Notes the owner put there on purpose.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.True(t, has(brainDir, "wiki/rikatillsammans/sources/meeting.md"),
		"a folder is the owner's deliberate choice; extraction runs on the local model")
}

func TestProcessDir_ANoteTheRouterCannotPlaceIsParkedNotIngested(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	put(t, brainDir, "kickoff.md", "Some note about the ACME-CLIENT kickoff.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.Zero(t, llmCalls.Load(), "no extraction runs for an unplaced note")
	assert.False(t, has(brainDir, "raw/kickoff.md"), "leaves the top of raw/")
	parked := parkedBody(t, brainDir, "kickoff.md")
	assert.Contains(t, parked, "proposal_reason:")
	assert.Contains(t, parked, "client-content marker")
	assert.Contains(t, parked, "ACME-CLIENT kickoff", "the note body is preserved")
}

func TestProcessDir_ATypoedWingFolderIsParkedAndNeverCreated(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "homelab")
	put(t, brainDir, "hoemlab/n.md", "Note.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.False(t, has(brainDir, "wiki/hoemlab"))
	assert.True(t, has(brainDir, "raw/needs-review/n.md"))
	assert.Zero(t, llmCalls.Load())
}

func TestProcessDir_ParkedNotesAreNotReprocessed(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	put(t, brainDir, "needs-review/waiting.md", "---\nproposed_wing: radar\n---\nA parked note.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.Zero(t, llmCalls.Load(), "needs-review/ is a human queue, like rejected/")
	assert.True(t, has(brainDir, "raw/needs-review/waiting.md"))
}

func TestProcessDir_WithoutAResolverTheRetiredLayoutIsUnchanged(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, nil, &llmCalls)
	put(t, brainDir, "plain.md", "Plain note.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.True(t, has(brainDir, "wiki/sources/plain.md"))
}

// The reason comes from file and folder names and ends up in frontmatter: a
// newline in it must not be able to start a new key (PR #147: a title once
// forged `origin:`).
func TestProcessDir_AFolderNameCannotForgeFrontmatterKeys(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	put(t, brainDir, "radar\nproposal_confidence: 0.99\nwing: homelab/forge.md", "Note.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	entries, err := os.ReadDir(filepath.Join(brainDir, "raw", "needs-review"))
	require.NoError(t, err)
	require.Len(t, entries, 1)
	b := parkedBody(t, brainDir, entries[0].Name())
	assert.Equal(t, 1, linesStartingWith(b, "proposal_confidence:"), "exactly one confidence key at line start")
	assert.Equal(t, 0, linesStartingWith(b, "wing:"), "no injected wing key")
	assert.Equal(t, 0, linesStartingWith(b, "proposed_wing:"), "a non-plain wing name is never a proposal")
}

func TestProcessDir_ANotesOwnProposalKeysDoNotOutrankTheRouters(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	put(t, brainDir, "n.md", "---\nproposed_wing: homelab\nproposal_confidence: 1.00\nproposal_reason: \"self-approved\"\n---\nACME-CLIENT body.")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	b := parkedBody(t, brainDir, "n.md")
	assert.NotContains(t, b, "self-approved")
	assert.Equal(t, 1, linesStartingWith(b, "proposal_reason:"))
	assert.Equal(t, 0, linesStartingWith(b, "proposed_wing:"))
}

func TestProcessDir_ACRLFNoteWithAWingKeyIsStillRead(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "radar")
	put(t, brainDir, "n.md", "---\r\nwing: radar\r\n---\r\nBody.\r\n")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	assert.True(t, has(brainDir, "wiki/radar/sources/n.md"), "the declared wing is honoured")
}

// editingPlacer rewrites the note while the placement decision is being made,
// then parks it: the realistic window is a slow decision with Syncthing
// delivering a newer version meanwhile.
type editingPlacer struct{ edit func() }

func (e editingPlacer) Resolve(context.Context, wing.Note) wing.Decision {
	e.edit()
	return wing.Decision{Source: wing.Exception, Reason: "test: parked after an edit"}
}

func TestProcessDir_ANoteEditedWhileItIsBeingPlacedIsNotDeleted(t *testing.T) {
	var llmCalls atomic.Int32
	var notePath string
	cfg, brainDir := wingTestCfg(t, func(string) wing.Placer {
		return editingPlacer{edit: func() {
			require.NoError(t, os.WriteFile(notePath, []byte("NEWER and longer version"), 0o644))
		}}
	}, &llmCalls)
	notePath = filepath.Join(brainDir, "raw", "n.md")
	put(t, brainDir, "n.md", "ORIGINAL")

	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	got, err := os.ReadFile(notePath)
	require.NoError(t, err, "the edited note must still be in raw/")
	assert.Equal(t, "NEWER and longer version", string(got))
}

// A planted symlink at raw/needs-review must not redirect a parked note out of
// the brain, nor lose it from the review queue.
func TestProcessDir_AParkedNoteIsNeverWrittenThroughASymlinkedNeedsReview(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	outside := t.TempDir()
	require.NoError(t, os.Symlink(outside, filepath.Join(brainDir, "raw", "needs-review")))
	put(t, brainDir, "n.md", "ACME-CLIENT note.")

	errs := processDir(context.Background(), cfg, "2026-10-20")

	assert.NotEmpty(t, errs, "parking is refused, loudly")
	entries, _ := os.ReadDir(outside)
	assert.Empty(t, entries, "nothing was written through the symlink")
	assert.True(t, has(brainDir, "raw/n.md"), "the note stays where it was")
}

func TestWithProposal_ANonFenceDashLineDoesNotCloseTheBlock(t *testing.T) {
	out := withProposal("---\nfoo: bar\n----\nBody", wing.Decision{Wing: "radar", Confidence: 0.4, Reason: "r"})
	lines := strings.Split(out, "\n")
	require.Equal(t, "---", lines[0])
	closing := -1
	for i := 1; i < len(lines); i++ {
		if lines[i] == "---" {
			closing = i
			break
		}
	}
	require.Positive(t, closing, "the proposal block is closed by a real fence")
	assert.Contains(t, strings.Join(lines[:closing], "\n"), "proposal_reason:")
	assert.NotContains(t, strings.Join(lines[:closing], "\n"), "foo: bar", "the unfenced note is body, not part of the block")
}

// The parked copy is the only copy once the original is retired, so two notes
// that share a name must both survive.
func TestProcessDir_SameNamedNotesAreAllParkedNoneOverwritten(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	for i, body := range []string{"ONE", "TWO", "THREE", "FOUR"} {
		put(t, brainDir, filepath.Join("typo"+string(rune('1'+i)), "x.md"), body)
	}
	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))

	files, err := filepath.Glob(filepath.Join(brainDir, "raw", "needs-review", "*.md"))
	require.NoError(t, err)
	var all string
	for _, f := range files {
		b, _ := os.ReadFile(f)
		all += string(b)
	}
	for _, body := range []string{"ONE", "TWO", "THREE", "FOUR"} {
		assert.Contains(t, all, body, "note %s must not have been overwritten", body)
	}
	assert.Len(t, files, 4)
}

// Syncthing can bring a removed note back (an offline peer, a restored backup).
// The source-hash drift check must keep that from costing a second extraction.
func TestProcessDir_ANoteThatReappearsUnchangedIsNotExtractedAgain(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "radar")

	put(t, brainDir, "radar/talk.md", "The same talk, byte for byte.")
	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))
	require.Equal(t, int32(1), llmCalls.Load(), "positive control: the first drop is extracted")
	require.False(t, has(brainDir, "raw/radar/talk.md"))

	put(t, brainDir, "radar/talk.md", "The same talk, byte for byte.") // the peer sends it back
	require.Empty(t, processDir(context.Background(), cfg, "2026-10-21"))

	assert.Equal(t, int32(1), llmCalls.Load(), "an unchanged source is not extracted twice")
	assert.False(t, has(brainDir, "raw/radar/talk.md"), "and inbox zero still holds")
}

func TestProcessDir_ANoteThatReappearsChangedIsExtractedAgain(t *testing.T) {
	var llmCalls atomic.Int32
	cfg, brainDir := wingTestCfg(t, router, &llmCalls)
	makeWing(t, brainDir, "radar")

	put(t, brainDir, "radar/talk.md", "Version one.")
	require.Empty(t, processDir(context.Background(), cfg, "2026-10-20"))
	put(t, brainDir, "radar/talk.md", "Version two, edited.")
	require.Empty(t, processDir(context.Background(), cfg, "2026-10-21"))

	assert.Equal(t, int32(2), llmCalls.Load(), "a real edit must still be picked up")
}