Implement deferred core utility parity APIs/tests and refresh tracking artifacts

This commit is contained in:
Joseph Doherty
2026-02-27 10:27:05 -05:00
parent c0aaae9236
commit b94a67be6e
17 changed files with 842 additions and 9 deletions

View File

@@ -256,6 +256,10 @@ func (a *Analyzer) parseTestFile(filePath string) ([]TestFunc, []ImportInfo, int
}
test.FeatureName = a.inferFeatureName(name)
test.BestFeatureIdx = -1
if fn.Body != nil {
test.Calls = a.extractCalls(fn.Body)
}
tests = append(tests, test)
}
@@ -331,6 +335,210 @@ func (a *Analyzer) inferFeatureName(testName string) string {
return name
}
// extractCalls walks an AST block statement and extracts all function/method calls.
func (a *Analyzer) extractCalls(body *ast.BlockStmt) []CallInfo {
seen := make(map[string]bool)
var calls []CallInfo
ast.Inspect(body, func(n ast.Node) bool {
callExpr, ok := n.(*ast.CallExpr)
if !ok {
return true
}
var ci CallInfo
switch fun := callExpr.Fun.(type) {
case *ast.Ident:
ci = CallInfo{FuncName: fun.Name}
case *ast.SelectorExpr:
ci = CallInfo{
RecvOrPkg: extractIdent(fun.X),
MethodName: fun.Sel.Name,
IsSelector: true,
}
default:
return true
}
key := ci.callKey()
if !seen[key] && !isFilteredCall(ci) {
seen[key] = true
calls = append(calls, ci)
}
return true
})
return calls
}
// extractIdent extracts an identifier name from an expression (handles X in X.Y).
func extractIdent(expr ast.Expr) string {
switch e := expr.(type) {
case *ast.Ident:
return e.Name
case *ast.SelectorExpr:
return extractIdent(e.X) + "." + e.Sel.Name
default:
return ""
}
}
// isFilteredCall returns true if a call should be excluded from feature matching.
func isFilteredCall(c CallInfo) bool {
if c.IsSelector {
recv := c.RecvOrPkg
// testing.T/B methods
if recv == "t" || recv == "b" || recv == "tb" {
return true
}
// stdlib packages
if stdlibPkgs[recv] {
return true
}
// NATS client libs
if recv == "nats" || recv == "nuid" || recv == "nkeys" || recv == "jwt" {
return true
}
return false
}
// Go builtins
name := c.FuncName
if builtinFuncs[name] {
return true
}
// Test assertion helpers
lower := strings.ToLower(name)
if strings.HasPrefix(name, "require_") {
return true
}
for _, prefix := range []string{"check", "verify", "assert", "expect"} {
if strings.HasPrefix(lower, prefix) {
return true
}
}
return false
}
// featureRef identifies a feature within the analysis result.
type featureRef struct {
moduleIdx int
featureIdx int
goFile string
goClass string
}
// resolveCallGraph matches test calls against known features across all modules.
func resolveCallGraph(result *AnalysisResult) {
// Build method index: go_method name → list of feature refs
methodIndex := make(map[string][]featureRef)
for mi, mod := range result.Modules {
for fi, feat := range mod.Features {
ref := featureRef{
moduleIdx: mi,
featureIdx: fi,
goFile: feat.GoFile,
goClass: feat.GoClass,
}
methodIndex[feat.GoMethod] = append(methodIndex[feat.GoMethod], ref)
}
}
// For each test, resolve calls to features
for mi := range result.Modules {
mod := &result.Modules[mi]
for ti := range mod.Tests {
test := &mod.Tests[ti]
seen := make(map[int]bool) // feature indices already linked
var linked []int
testFileBase := sourceFileBase(test.GoFile)
for _, call := range test.Calls {
// Look up the method name
name := call.MethodName
if !call.IsSelector {
name = call.FuncName
}
candidates := methodIndex[name]
if len(candidates) == 0 {
continue
}
// Ambiguity threshold: skip very common method names
if len(candidates) > 10 {
continue
}
// Filter to same module
var sameModule []featureRef
for _, ref := range candidates {
if ref.moduleIdx == mi {
sameModule = append(sameModule, ref)
}
}
if len(sameModule) == 0 {
continue
}
for _, ref := range sameModule {
if !seen[ref.featureIdx] {
seen[ref.featureIdx] = true
linked = append(linked, ref.featureIdx)
}
}
}
test.LinkedFeatures = linked
// Set BestFeatureIdx using priority:
// (a) existing inferFeatureName match
// (b) same-file-base match
// (c) first remaining candidate
if test.BestFeatureIdx < 0 && len(linked) > 0 {
// Try same-file-base match first
for _, fi := range linked {
featFileBase := sourceFileBase(mod.Features[fi].GoFile)
if featFileBase == testFileBase {
test.BestFeatureIdx = fi
break
}
}
// Fall back to first candidate
if test.BestFeatureIdx < 0 {
test.BestFeatureIdx = linked[0]
}
}
}
}
}
// sourceFileBase strips _test.go suffix and path to get the base file name.
func sourceFileBase(goFile string) string {
base := filepath.Base(goFile)
base = strings.TrimSuffix(base, "_test.go")
base = strings.TrimSuffix(base, ".go")
return base
}
var stdlibPkgs = map[string]bool{
"fmt": true, "time": true, "strings": true, "bytes": true, "errors": true,
"os": true, "math": true, "sort": true, "reflect": true, "sync": true,
"context": true, "io": true, "filepath": true, "strconv": true,
"encoding": true, "json": true, "binary": true, "hex": true, "rand": true,
"runtime": true, "atomic": true, "slices": true, "testing": true,
"net": true, "bufio": true, "crypto": true, "log": true, "regexp": true,
"unicode": true, "http": true, "url": true,
}
var builtinFuncs = map[string]bool{
"make": true, "append": true, "len": true, "cap": true, "close": true,
"delete": true, "panic": true, "recover": true, "print": true,
"println": true, "copy": true, "new": true,
}
// isStdlib checks if an import path is a Go standard library package.
func isStdlib(importPath string) bool {
firstSlash := strings.Index(importPath, "/")

View File

@@ -11,28 +11,47 @@ func main() {
sourceDir := flag.String("source", "", "Path to Go source root (e.g., ../../golang/nats-server)")
dbPath := flag.String("db", "", "Path to SQLite database file (e.g., ../../porting.db)")
schemaPath := flag.String("schema", "", "Path to SQL schema file (e.g., ../../porting-schema.sql)")
mode := flag.String("mode", "full", "Analysis mode: 'full' (default) or 'call-graph' (incremental)")
flag.Parse()
if *sourceDir == "" || *dbPath == "" || *schemaPath == "" {
fmt.Fprintf(os.Stderr, "Usage: go-analyzer --source <path> --db <path> --schema <path>\n")
if *sourceDir == "" || *dbPath == "" {
fmt.Fprintf(os.Stderr, "Usage: go-analyzer --source <path> --db <path> [--schema <path>] [--mode full|call-graph]\n")
flag.PrintDefaults()
os.Exit(1)
}
switch *mode {
case "full":
runFull(*sourceDir, *dbPath, *schemaPath)
case "call-graph":
runCallGraph(*sourceDir, *dbPath)
default:
log.Fatalf("Unknown mode %q: must be 'full' or 'call-graph'", *mode)
}
}
func runFull(sourceDir, dbPath, schemaPath string) {
if schemaPath == "" {
log.Fatal("--schema is required for full mode")
}
// Open DB and apply schema
db, err := OpenDB(*dbPath, *schemaPath)
db, err := OpenDB(dbPath, schemaPath)
if err != nil {
log.Fatalf("Failed to open database: %v", err)
}
defer db.Close()
// Run analysis
analyzer := NewAnalyzer(*sourceDir)
analyzer := NewAnalyzer(sourceDir)
result, err := analyzer.Analyze()
if err != nil {
log.Fatalf("Analysis failed: %v", err)
}
// Resolve call graph before writing
resolveCallGraph(result)
// Write to DB
writer := NewDBWriter(db)
if err := writer.WriteAll(result); err != nil {
@@ -46,3 +65,35 @@ func main() {
fmt.Printf(" Dependencies: %d\n", len(result.Dependencies))
fmt.Printf(" Imports: %d\n", len(result.Imports))
}
func runCallGraph(sourceDir, dbPath string) {
// Open existing DB without schema
db, err := OpenDBNoSchema(dbPath)
if err != nil {
log.Fatalf("Failed to open database: %v", err)
}
defer db.Close()
// Run analysis (parse Go source)
analyzer := NewAnalyzer(sourceDir)
result, err := analyzer.Analyze()
if err != nil {
log.Fatalf("Analysis failed: %v", err)
}
// Resolve call graph
resolveCallGraph(result)
// Update DB incrementally
writer := NewDBWriter(db)
stats, err := writer.UpdateCallGraph(result)
if err != nil {
log.Fatalf("Failed to update call graph: %v", err)
}
fmt.Printf("Call graph analysis complete:\n")
fmt.Printf(" Tests analyzed: %d\n", stats.TestsAnalyzed)
fmt.Printf(" Tests linked: %d\n", stats.TestsLinked)
fmt.Printf(" Dependency rows: %d\n", stats.DependencyRows)
fmt.Printf(" Feature IDs set: %d\n", stats.FeatureIDsSet)
}

View File

@@ -152,3 +152,176 @@ func (w *DBWriter) insertLibrary(tx *sql.Tx, imp *ImportInfo) error {
)
return err
}
// OpenDBNoSchema opens an existing SQLite database without applying schema.
// It verifies that the required tables exist.
func OpenDBNoSchema(dbPath string) (*sql.DB, error) {
db, err := sql.Open("sqlite3", dbPath+"?_journal_mode=WAL&_foreign_keys=ON")
if err != nil {
return nil, fmt.Errorf("opening database: %w", err)
}
// Verify required tables exist
for _, table := range []string{"modules", "features", "unit_tests", "dependencies"} {
var name string
err := db.QueryRow("SELECT name FROM sqlite_master WHERE type='table' AND name=?", table).Scan(&name)
if err != nil {
db.Close()
return nil, fmt.Errorf("required table %q not found: %w", table, err)
}
}
return db, nil
}
// CallGraphStats holds summary statistics from a call-graph update.
type CallGraphStats struct {
TestsAnalyzed int
TestsLinked int
DependencyRows int
FeatureIDsSet int
}
// UpdateCallGraph writes call-graph analysis results to the database incrementally.
func (w *DBWriter) UpdateCallGraph(result *AnalysisResult) (*CallGraphStats, error) {
stats := &CallGraphStats{}
// Load module name→ID mapping
moduleIDs := make(map[string]int64)
rows, err := w.db.Query("SELECT id, name FROM modules")
if err != nil {
return nil, fmt.Errorf("querying modules: %w", err)
}
for rows.Next() {
var id int64
var name string
if err := rows.Scan(&id, &name); err != nil {
rows.Close()
return nil, err
}
moduleIDs[name] = id
}
rows.Close()
// Load feature DB IDs: "module_name:go_method:go_class" → id
type featureKey struct {
moduleName string
goMethod string
goClass string
}
featureDBIDs := make(map[featureKey]int64)
rows, err = w.db.Query(`
SELECT f.id, m.name, f.go_method, COALESCE(f.go_class, '')
FROM features f
JOIN modules m ON f.module_id = m.id
`)
if err != nil {
return nil, fmt.Errorf("querying features: %w", err)
}
for rows.Next() {
var id int64
var modName, goMethod, goClass string
if err := rows.Scan(&id, &modName, &goMethod, &goClass); err != nil {
rows.Close()
return nil, err
}
featureDBIDs[featureKey{modName, goMethod, goClass}] = id
}
rows.Close()
// Load test DB IDs: "module_name:go_method" → id
testDBIDs := make(map[string]int64)
rows, err = w.db.Query(`
SELECT ut.id, m.name, ut.go_method
FROM unit_tests ut
JOIN modules m ON ut.module_id = m.id
`)
if err != nil {
return nil, fmt.Errorf("querying unit_tests: %w", err)
}
for rows.Next() {
var id int64
var modName, goMethod string
if err := rows.Scan(&id, &modName, &goMethod); err != nil {
rows.Close()
return nil, err
}
testDBIDs[modName+":"+goMethod] = id
}
rows.Close()
// Begin transaction
tx, err := w.db.Begin()
if err != nil {
return nil, fmt.Errorf("beginning transaction: %w", err)
}
defer tx.Rollback()
// Clear old call-graph data
if _, err := tx.Exec("DELETE FROM dependencies WHERE source_type='unit_test' AND dependency_kind='calls'"); err != nil {
return nil, fmt.Errorf("clearing old dependencies: %w", err)
}
if _, err := tx.Exec("UPDATE unit_tests SET feature_id = NULL"); err != nil {
return nil, fmt.Errorf("clearing old feature_ids: %w", err)
}
// Prepare statements
insertDep, err := tx.Prepare("INSERT OR IGNORE INTO dependencies (source_type, source_id, target_type, target_id, dependency_kind) VALUES ('unit_test', ?, 'feature', ?, 'calls')")
if err != nil {
return nil, fmt.Errorf("preparing insert dependency: %w", err)
}
defer insertDep.Close()
updateFeatureID, err := tx.Prepare("UPDATE unit_tests SET feature_id = ? WHERE id = ?")
if err != nil {
return nil, fmt.Errorf("preparing update feature_id: %w", err)
}
defer updateFeatureID.Close()
// Process each module's tests
for _, mod := range result.Modules {
for _, test := range mod.Tests {
stats.TestsAnalyzed++
testDBID, ok := testDBIDs[mod.Name+":"+test.GoMethod]
if !ok {
continue
}
// Insert dependency rows for linked features
if len(test.LinkedFeatures) > 0 {
stats.TestsLinked++
}
for _, fi := range test.LinkedFeatures {
feat := mod.Features[fi]
featDBID, ok := featureDBIDs[featureKey{mod.Name, feat.GoMethod, feat.GoClass}]
if !ok {
continue
}
if _, err := insertDep.Exec(testDBID, featDBID); err != nil {
return nil, fmt.Errorf("inserting dependency for test %s: %w", test.GoMethod, err)
}
stats.DependencyRows++
}
// Set feature_id for best match
if test.BestFeatureIdx >= 0 {
feat := mod.Features[test.BestFeatureIdx]
featDBID, ok := featureDBIDs[featureKey{mod.Name, feat.GoMethod, feat.GoClass}]
if !ok {
continue
}
if _, err := updateFeatureID.Exec(featDBID, testDBID); err != nil {
return nil, fmt.Errorf("updating feature_id for test %s: %w", test.GoMethod, err)
}
stats.FeatureIDsSet++
}
}
}
if err := tx.Commit(); err != nil {
return nil, fmt.Errorf("committing transaction: %w", err)
}
return stats, nil
}

View File

@@ -58,6 +58,28 @@ type TestFunc struct {
GoLineCount int
// FeatureName links this test to a feature by naming convention
FeatureName string
// Calls holds raw function/method calls extracted from the test body AST
Calls []CallInfo
// LinkedFeatures holds indices into the parent module's Features slice
LinkedFeatures []int
// BestFeatureIdx is the primary feature match index (-1 = none)
BestFeatureIdx int
}
// CallInfo represents a function or method call extracted from a test body.
type CallInfo struct {
FuncName string // direct call name: "newMemStore"
RecvOrPkg string // selector receiver/pkg: "ms", "fmt", "t"
MethodName string // selector method: "StoreMsg", "Fatalf"
IsSelector bool // true for X.Y() form
}
// callKey returns a deduplication key for this call.
func (c CallInfo) callKey() string {
if c.IsSelector {
return c.RecvOrPkg + "." + c.MethodName
}
return c.FuncName
}
// Dependency represents a call relationship between two items.