Fix Magma import/call extraction, resolve file-path imports in linker

- Add field('path', ...) to Magma load_statement grammar rule so
  parse_generic_imports() finds the import path via field lookup
  instead of the broken text fallback (which only extracted 1 per file)
- Fix passImports() to resolve file-path imports (e.g. "utils.mag",
  "lib/helpers.h") via fqn.ModuleQN() when raw path doesn't match
  any node QN — general fix benefiting any file-path-based import
- Add TestMagmaImport_Regression and TestMagmaCall_Regression
- Update language count 59 → 63 in README, docs/index.html, marketing

Author: DeusData

BENCHMARK.md

@@ -2,7 +2,7 @@
 
 ## Methodology
 
-- **35 languages** (27 programming + 8 config/markup), 12 questions each (4 for config languages)
+- **63 languages** (27 programming + 8 config/markup), 12 questions each (4 for config languages)
 - **Up to 5 attempts** per question with escalating retry strategies
 - **Real open-source repos** (medium to large: 78--49K nodes)
 - **Grading**: PASS (1.0) / PARTIAL (0.5) / FAIL (0.0), N/A excluded from denominator
@@ -847,7 +847,7 @@ The tool handles 20K-node kernel subsystems without timeouts. Deep traces on wel
 ## Cross-Cutting Findings
 
 ### Strengths
-1. **Zero indexing failures** across 35 repos of all sizes (78 to 49K nodes)
+1. **Zero indexing failures** across 63 repos of all sizes (78 to 49K nodes)
 2. **100% on 17 languages** -- half the languages achieved perfect scores
 3. **No language below 62%** -- even the weakest performs core operations
 4. **Massive codebases handled**: Django (49K nodes), Laravel (38K nodes), neovim (24K nodes) -- no performance issues

CONTRIBUTING.md

@@ -37,7 +37,7 @@ golangci-lint run ./...
 ```
 cmd/codebase-memory-mcp/  Entry point (MCP server + CLI + install/update)
 internal/
-  lang/                   Language specs (35 languages, tree-sitter node types)
+  lang/                   Language specs (63 languages, tree-sitter node types)
   parser/                 Tree-sitter grammar loading
   pipeline/               Multi-pass indexing pipeline
   httplink/               Cross-service HTTP route matching

README.md

@@ -8,7 +8,7 @@ Parses source code with [tree-sitter](https://tree-sitter.github.io/tree-sitter/
 
 ## Features
 
-- **59 languages**: Python, Go, JavaScript, TypeScript, TSX, Rust, Java, C++, C#, C, PHP, Lua, Scala, Kotlin, Ruby, Bash, Zig, Elixir, Haskell, OCaml, Objective-C, Swift, Dart, Perl, Groovy, Erlang, R, Clojure, F#, Julia, Vim Script, Nix, Common Lisp, Elm, Fortran, CUDA, COBOL, Verilog, Emacs Lisp, HTML, CSS, SCSS, YAML, TOML, HCL, SQL, Dockerfile, JSON, XML, Markdown, Makefile, CMake, Protobuf, GraphQL, Vue, Svelte, Meson, GLSL, INI
+- **63 languages**: Python, Go, JavaScript, TypeScript, TSX, Rust, Java, C++, C#, C, PHP, Lua, Scala, Kotlin, Ruby, Bash, Zig, Elixir, Haskell, OCaml, Objective-C, Swift, Dart, Perl, Groovy, Erlang, R, Clojure, F#, Julia, Vim Script, Nix, Common Lisp, Elm, Fortran, CUDA, COBOL, Verilog, Emacs Lisp, MATLAB, Lean 4, FORM, Magma, HTML, CSS, SCSS, YAML, TOML, HCL, SQL, Dockerfile, JSON, XML, Markdown, Makefile, CMake, Protobuf, GraphQL, Vue, Svelte, Meson, GLSL, INI
 - **Architecture overview**: `get_architecture` returns languages, packages, entry points, routes, hotspots, boundaries, layers, and clusters in a single call — instant codebase orientation
 - **Architecture Decision Records**: `manage_adr` persists architectural decisions (PURPOSE, STACK, ARCHITECTURE, PATTERNS, TRADEOFFS, PHILOSOPHY) across sessions with section filtering and validation
 - **Louvain community detection**: Discovers hidden functional modules across packages by clustering CALLS, HTTP_CALLS, and ASYNC_CALLS edges
@@ -57,7 +57,7 @@ Claude Code formats and explains the results.
 
 **Why no built-in LLM?** Other code graph tools embed an LLM to translate natural language into graph queries. This means extra API keys, extra cost per query, and another model to configure. With MCP, the AI assistant you're already talking to *is* the query translator — no duplication needed.
 
-**Token efficiency**: Compared to having an AI agent grep through your codebase file by file, graph queries return precise results in a single tool call. In benchmarks across 59 real-world repos (78 to 49K nodes), five structural queries consumed ~3,400 tokens via codebase-memory-mcp versus ~412,000 tokens via file-by-file exploration — a **99.2% reduction**. All 59 supported languages use the same efficient graph backend.
+**Token efficiency**: Compared to having an AI agent grep through your codebase file by file, graph queries return precise results in a single tool call. In benchmarks across 63 real-world repos (78 to 49K nodes), five structural queries consumed ~3,400 tokens via codebase-memory-mcp versus ~412,000 tokens via file-by-file exploration — a **99.2% reduction**. All 63 supported languages use the same efficient graph backend.
 
 ## Performance
 
@@ -663,7 +663,7 @@ make install  # go install
 
 ## Language Benchmark
 
-59 languages supported. Benchmarked against 59 real open-source repositories (78 to 49K nodes). 12 standardized questions per language. Grading: HIGH (1.0) / MEDIUM (0.5) / LOW (0.1). Overall: **76%** average MCP score across all languages (97% for explorer-based agents).
+63 languages supported. Benchmarked against 63 real open-source repositories (78 to 49K nodes). 12 standardized questions per language. Grading: HIGH (1.0) / MEDIUM (0.5) / LOW (0.1). Overall: **76%** average MCP score across all languages (97% for explorer-based agents).
 
 | Tier | Score | Languages |
 |------|-------|-----------|
@@ -682,8 +682,8 @@ See [`BENCHMARK.md`](BENCHMARK.md) for the full 35-language benchmark with per-q
 cmd/codebase-memory-mcp/  Entry point (MCP stdio server + CLI mode + install/update commands)
 internal/
   store/                  SQLite graph storage (nodes, edges, traversal, search, architecture, Louvain clustering)
-  lang/                   Language specs (59 languages, tree-sitter node types)
-  cbm/                    Vendored tree-sitter C grammars (59 languages) and AST extraction engine
+  lang/                   Language specs (63 languages, tree-sitter node types)
+  cbm/                    Vendored tree-sitter C grammars (63 languages) and AST extraction engine
   pipeline/               Multi-pass indexing (structure → definitions → calls → HTTP links → config links → communities → tests)
   httplink/               Cross-service HTTP route/call-site matching
   cypher/                 Cypher query lexer, parser, planner, executor

docs/index.html

@@ -4,19 +4,19 @@
     <meta charset="UTF-8">
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
     <title>codebase-memory-mcp — Code Knowledge Graph for AI Assistants</title>
-    <meta name="description" content="MCP server that indexes codebases into a persistent knowledge graph. 59 languages, 120x fewer tokens, single Go binary. Works with Claude Code, Codex CLI, Cursor, Windsurf, Gemini CLI, VS Code, Zed.">
+    <meta name="description" content="MCP server that indexes codebases into a persistent knowledge graph. 63 languages, 120x fewer tokens, single Go binary. Works with Claude Code, Codex CLI, Cursor, Windsurf, Gemini CLI, VS Code, Zed.">
     <meta name="keywords" content="MCP server, code analysis, knowledge graph, tree-sitter, Claude Code, Codex, Cursor, Windsurf, Gemini CLI, VS Code, Zed, developer tools, code exploration, token reduction">
 
     <!-- Open Graph -->
     <meta property="og:title" content="codebase-memory-mcp — Code Knowledge Graph for AI Assistants">
-    <meta property="og:description" content="120x fewer tokens for AI code exploration. 59 languages, sub-ms queries, single Go binary.">
+    <meta property="og:description" content="120x fewer tokens for AI code exploration. 63 languages, sub-ms queries, single Go binary.">
     <meta property="og:type" content="website">
     <meta property="og:url" content="https://deusdata.github.io/codebase-memory-mcp">
 
     <!-- Twitter Card -->
     <meta name="twitter:card" content="summary_large_image">
     <meta name="twitter:title" content="codebase-memory-mcp — Code Knowledge Graph for AI Assistants">
-    <meta name="twitter:description" content="120x fewer tokens for AI code exploration. 59 languages, sub-ms queries, single Go binary.">
+    <meta name="twitter:description" content="120x fewer tokens for AI code exploration. 63 languages, sub-ms queries, single Go binary.">
 
     <link rel="canonical" href="https://deusdata.github.io/codebase-memory-mcp">
 
@@ -210,7 +210,7 @@ <h1>codebase-memory-mcp</h1>
                 <div class="label">fewer tokens</div>
             </div>
             <div class="stat">
-                <div class="number">59</div>
+                <div class="number">63</div>
                 <div class="label">languages</div>
             </div>
             <div class="stat">
@@ -322,8 +322,8 @@ <h2>Benchmark results</h2>
         <h2>Features</h2>
         <div class="features">
             <div class="feature">
-                <h3>59 Languages</h3>
-                <p>Python, Go, JS, TS, TSX, Rust, Java, C++, C#, C, PHP, Ruby, Kotlin, Scala, Zig, Elixir, Haskell, OCaml, Swift, Dart, and 39 more via vendored tree-sitter grammars.</p>
+                <h3>63 Languages</h3>
+                <p>Python, Go, JS, TS, TSX, Rust, Java, C++, C#, C, PHP, Ruby, Kotlin, Scala, Zig, Elixir, Haskell, OCaml, Swift, Dart, MATLAB, Lean 4, and 41 more via vendored tree-sitter grammars.</p>
             </div>
             <div class="feature">
                 <h3>Call Graph Tracing</h3>
@@ -371,7 +371,7 @@ <h2>How it compares</h2>
                 </tr>
             </thead>
             <tbody>
-                <tr><td>Languages</td><td class="win">59</td><td>8-11</td></tr>
+                <tr><td>Languages</td><td class="win">63</td><td>8-11</td></tr>
                 <tr><td>Runtime</td><td class="win">Single Go binary</td><td>Node.js (npx)</td></tr>
                 <tr><td>Runtime dependency</td><td class="win">None</td><td>Node.js</td></tr>
                 <tr><td>Embedded LLM</td><td class="win">No (uses your MCP client)</td><td>Yes (extra API key + cost)</td></tr>

internal/cbm/regression_test.go

@@ -1031,3 +1031,34 @@ func TestMagmaParse_Regression(t *testing.T) {
 		t.Fatal(err)
 	}
 }
+
+func TestMagmaImport_Regression(t *testing.T) {
+	src := []byte("load \"utils.mag\";\nload \"lib/helpers.mag\";\n")
+	r, err := ExtractFile(src, lang.Magma, "t", "main.mag")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(r.Imports) < 2 {
+		t.Fatalf("expected at least 2 imports, got %d", len(r.Imports))
+	}
+}
+
+func TestMagmaCall_Regression(t *testing.T) {
+	src := []byte("function Foo(x)\n  y := Bar(x);\n  return y;\nend function;\n")
+	r, err := ExtractFile(src, lang.Magma, "t", "calls.mag")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(r.Calls) == 0 {
+		t.Fatal("expected calls to be extracted, got 0")
+	}
+	found := false
+	for _, c := range r.Calls {
+		if c.CalleeName == "Bar" {
+			found = true
+		}
+	}
+	if !found {
+		t.Errorf("expected call to 'Bar', got calls: %v", r.Calls)
+	}
+}