zzet--gortex/internal/codegen/annotations.go

package codegen

import (
	"sort"
	"strings"

	"github.com/zzet/gortex/internal/graph"
)

// annotationGen describes the codegen tool an annotation implies and
// the member kinds it produces — members that exist after compilation
// but never appear in source, so the indexer cannot see them.
type annotationGen struct {
	tool    string
	members []string
}

// annotationGenerators maps a generation-implying annotation (Java
// Lombok, MapStruct, Kotlin compiler plugins) to the tool plus the
// members it synthesises. Keyed by the bare annotation name — no `@`,
// no package qualifier.
var annotationGenerators = map[string]annotationGen{
	// Lombok.
	"Data":                    {"lombok", []string{"getters", "setters", "equals_hashcode", "tostring", "required_args_constructor"}},
	"Value":                   {"lombok", []string{"getters", "equals_hashcode", "tostring", "all_args_constructor"}},
	"Getter":                  {"lombok", []string{"getters"}},
	"Setter":                  {"lombok", []string{"setters"}},
	"Builder":                 {"lombok", []string{"builder"}},
	"SuperBuilder":            {"lombok", []string{"builder"}},
	"AllArgsConstructor":      {"lombok", []string{"all_args_constructor"}},
	"NoArgsConstructor":       {"lombok", []string{"no_args_constructor"}},
	"RequiredArgsConstructor": {"lombok", []string{"required_args_constructor"}},
	"EqualsAndHashCode":       {"lombok", []string{"equals_hashcode"}},
	"ToString":                {"lombok", []string{"tostring"}},
	"Slf4j":                   {"lombok", []string{"logger"}},
	// MapStruct — generates the <Mapper>Impl implementation class.
	"Mapper": {"mapstruct", []string{"mapper_impl"}},
	// Kotlin compiler plugins (KAPT / kotlin-parcelize).
	"Parcelize": {"kapt", []string{"parcelable"}},
	// CommunityToolkit.Mvvm .NET source generators: [ObservableProperty]
	// on a field generates the public property + change notification;
	// [RelayCommand] on a method generates an ICommand property.
	"ObservableProperty": {"mvvm_toolkit", []string{"observable_property"}},
	"RelayCommand":       {"mvvm_toolkit", []string{"relay_command"}},
}

// AnnotationGeneratedStats reports what MarkAnnotatedGenerated did.
type AnnotationGeneratedStats struct {
	NodesMarked int
	EdgesAdded  int
}

// normalizeAnnotationName strips a leading `@` and any package
// qualifier, leaving the bare annotation name.
func normalizeAnnotationName(name string) string {
	name = strings.TrimPrefix(strings.TrimSpace(name), "@")
	if i := strings.LastIndexByte(name, '.'); i >= 0 {
		name = name[i+1:]
	}
	return name
}

// MarkAnnotatedGenerated scans one file's extraction output for
// EdgeAnnotated edges that point at a generation-implying annotation
// (Lombok / MapStruct / Kotlin codegen) and stamps each annotated
// symbol with codegen-visibility metadata: has_generated_members,
// codegen_tool, and the generated_members list. It returns the
// EdgeGeneratedBy edges the caller must append to result.Edges.
//
// This makes generated members visible without materialising them: a
// `@Data` class is flagged as carrying Lombok-generated accessors, so
// dead-code analysis and agents know the fields are not unused and
// that a missing getName() is compiler output, not a bug.
func MarkAnnotatedGenerated(nodes []*graph.Node, edges []*graph.Edge) (extra []*graph.Edge, stats AnnotationGeneratedStats) {
	if len(edges) == 0 {
		return nil, stats
	}
	annoName := map[string]string{} // annotation node ID → bare name
	byID := make(map[string]*graph.Node, len(nodes))
	for _, n := range nodes {
		byID[n.ID] = n
		if n.Meta != nil {
			if k, _ := n.Meta["kind"].(string); k == "annotation" {
				annoName[n.ID] = normalizeAnnotationName(n.Name)
			}
		}
	}
	seenEdge := map[string]bool{}
	for _, e := range edges {
		if e.Kind != graph.EdgeAnnotated {
			continue
		}
		gen, ok := annotationGenerators[annoName[e.To]]
		if !ok {
			continue
		}
		host := byID[e.From]
		if host == nil {
			continue
		}
		if host.Meta == nil {
			host.Meta = map[string]any{}
		}
		host.Meta["has_generated_members"] = true
		host.Meta["codegen_tool"] = gen.tool
		host.Meta["generated_members"] = mergeStringSet(host.Meta["generated_members"], gen.members)
		stats.NodesMarked++

		key := e.From + "\x00" + gen.tool
		if !seenEdge[key] {
			seenEdge[key] = true
			extra = append(extra, &graph.Edge{
				From:     e.From,
				To:       "external::generator-tool:" + gen.tool,
				Kind:     graph.EdgeGeneratedBy,
				FilePath: host.FilePath,
				Origin:   graph.OriginASTResolved,
				Meta:     map[string]any{"tool": gen.tool},
			})
			stats.EdgesAdded++
		}
	}
	return extra, stats
}

// lombokGeneratorTool is the external node every materialized Lombok member
// is generated by.
const lombokGeneratorTool = "external::generator-tool:lombok"

// MaterializeLombokAccessors mints the actual member nodes a Lombok class
// generates — getters / setters / a builder() factory + Builder type / an
// Slf4j log field — for every class MarkAnnotatedGenerated already flagged
// with codegen_tool="lombok". Without these nodes, `obj.getName()` cannot
// resolve to a graph node and a call chain through a Lombok accessor breaks.
// The synthetic methods carry the same Meta["receiver"] + bare Name shape the
// Java method-call resolver expects, so a `user.getName()` lands on the
// minted getName. IDs are deterministic and existence-checked, so a
// hand-written accessor wins and a reindex never duplicates.
func MaterializeLombokAccessors(nodes []*graph.Node) (newNodes []*graph.Node, newEdges []*graph.Edge) {
	byID := make(map[string]*graph.Node, len(nodes))
	for _, n := range nodes {
		byID[n.ID] = n
	}
	// Group instance fields by their owning class (file + receiver). Static
	// constants (KindConstant) are excluded — Lombok does not accessor them.
	fieldsByClass := map[string][]*graph.Node{}
	for _, n := range nodes {
		if n.Kind != graph.KindField || n.Meta == nil {
			continue
		}
		if recv, _ := n.Meta["receiver"].(string); recv != "" {
			key := n.FilePath + "\x00" + recv
			fieldsByClass[key] = append(fieldsByClass[key], n)
		}
	}

	for _, host := range nodes {
		if host.Kind != graph.KindType && host.Kind != graph.KindInterface {
			continue
		}
		if host.Meta == nil {
			continue
		}
		if tool, _ := host.Meta["codegen_tool"].(string); tool != "lombok" {
			continue
		}
		members := metaStringSlice(host.Meta["generated_members"])
		if len(members) == 0 {
			continue
		}
		mset := map[string]bool{}
		for _, m := range members {
			mset[m] = true
		}
		class := host.Name

		mint := func(id, name string, kind graph.NodeKind, member string, meta map[string]any) bool {
			if byID[id] != nil {
				return false // a hand-written member or an already-minted one wins
			}
			meta["synthesized"] = true
			meta["generated"] = true
			meta["codegen_tool"] = "lombok"
			meta["generated_member"] = member
			node := &graph.Node{
				ID: id, Kind: kind, Name: name,
				FilePath: host.FilePath, StartLine: host.StartLine, EndLine: host.StartLine,
				Language: host.Language, Meta: meta,
			}
			byID[id] = node
			newNodes = append(newNodes, node)
			newEdges = append(newEdges,
				&graph.Edge{From: host.ID, To: id, Kind: graph.EdgeDefines, FilePath: host.FilePath, Line: host.StartLine},
				&graph.Edge{From: id, To: lombokGeneratorTool, Kind: graph.EdgeGeneratedBy, FilePath: host.FilePath, Origin: graph.OriginASTInferred, Meta: map[string]any{"tool": "lombok"}},
			)
			return true
		}

		for _, f := range fieldsByClass[host.FilePath+"\x00"+class] {
			ftype, _ := f.Meta["field_type"].(string)
			if mset["getters"] {
				g := lombokGetterName(f.Name, ftype)
				id := host.FilePath + "::" + class + "." + g
				if mint(id, g, graph.KindMethod, "getter", map[string]any{"receiver": class, "scope_class": class, "return_type": ftype, "signature": g + "()"}) && ftype != "" {
					newEdges = append(newEdges, &graph.Edge{From: id, To: "unresolved::" + simpleTypeName(ftype), Kind: graph.EdgeReferences, FilePath: host.FilePath, Origin: graph.OriginASTInferred, Meta: map[string]any{"ref_context": "return"}})
				}
			}
			if mset["setters"] {
				s := "set" + pascalCase(f.Name)
				mint(host.FilePath+"::"+class+"."+s, s, graph.KindMethod, "setter", map[string]any{"receiver": class, "scope_class": class, "return_type": "void", "signature": s + "(" + ftype + ")"})
			}
		}
		if mset["builder"] {
			mint(host.FilePath+"::"+class+".builder", "builder", graph.KindMethod, "builder", map[string]any{"receiver": class, "scope_class": class, "is_static": true, "return_type": class + "Builder", "signature": "builder()"})
			mint(host.FilePath+"::"+class+"Builder", class+"Builder", graph.KindType, "builder", map[string]any{})
		}
		if mset["logger"] {
			mint(host.FilePath+"::"+class+".log", "log", graph.KindField, "logger", map[string]any{"receiver": class, "field_type": "org.slf4j.Logger"})
		}
	}
	return newNodes, newEdges
}

// lombokGetterName returns the Lombok getter name for a field — `is<Name>` for
// a boolean (kept as-is when already `is`-prefixed), else `get<Name>`.
func lombokGetterName(field, ftype string) string {
	if ftype == "boolean" {
		if len(field) > 2 && field[0:2] == "is" && field[2] >= 'A' && field[2] <= 'Z' {
			return field
		}
		return "is" + pascalCase(field)
	}
	return "get" + pascalCase(field)
}

// pascalCase upper-cases the first letter of s.
func pascalCase(s string) string {
	if s == "" {
		return s
	}
	return strings.ToUpper(s[:1]) + s[1:]
}

// simpleTypeName strips generic arguments and a package qualifier from a Java
// type, leaving the bare type name for a reference edge.
func simpleTypeName(t string) string {
	if i := strings.IndexByte(t, '<'); i >= 0 {
		t = t[:i]
	}
	t = strings.TrimSpace(t)
	if i := strings.LastIndexByte(t, '.'); i >= 0 {
		t = t[i+1:]
	}
	return strings.TrimSuffix(t, "[]")
}

// metaStringSlice coerces a Meta value to []string, tolerating the []any form
// a persistence round-trip can produce.
func metaStringSlice(v any) []string {
	switch s := v.(type) {
	case []string:
		return s
	case []any:
		out := make([]string, 0, len(s))
		for _, e := range s {
			if str, ok := e.(string); ok {
				out = append(out, str)
			}
		}
		return out
	}
	return nil
}

// mergeStringSet merges add into an existing []string-typed meta value,
// returning a sorted, de-duplicated slice.
func mergeStringSet(existing any, add []string) []string {
	set := map[string]bool{}
	if cur, ok := existing.([]string); ok {
		for _, s := range cur {
			set[s] = true
		}
	}
	for _, s := range add {
		set[s] = true
	}
	out := make([]string, 0, len(set))
	for s := range set {
		out = append(out, s)
	}
	sort.Strings(out)
	return out
}