gitnexus 1.6.3-rc.25 → 1.6.3-rc.27

package/dist/core/group/extractors/http-patterns/php.js

@@ -1,27 +1,71 @@
 import PHP from 'tree-sitter-php';
 import { compilePatterns, runCompiledPatterns, unquoteLiteral, } from '../tree-sitter-scanner.js';
 /**
- * PHP HTTP plugin — Laravel `Route::get/post/...` declarations.
+ * PHP HTTP plugin.
+ *
+ * Providers:
+ *   - Laravel `Route::get/post/...`
+ *
+ * Consumers (string-literal URLs only):
+ *   - Laravel HTTP client: `Http::get/post/put/delete/patch($url)`
+ *   - Guzzle / generic object method: `$client->get/post/...($url)`
+ *   - `file_get_contents($url)`
  *
  * The pipeline already uses `PHP.php_only` for ingesting plain `.php`
  * files (see `core/tree-sitter/parser-loader.ts`), and we do the same
  * here so Laravel route files are parsed with the right grammar dialect.
+ *
+ * Scope notes: consumer patterns match string literals only. URLs built
+ * via binary concatenation (`$base . '/path'`), `sprintf`, or config
+ * lookup (`config('services.foo.base').'/path'`) are intentionally left
+ * for a follow-up — they require constant-folding the surrounding
+ * scope to be meaningful.
  */
-const LARAVEL_PATTERNS = compilePatterns({
-    name: 'php-laravel',
+const LARAVEL_ROUTE_SPEC = {
+    meta: {},
+    query: `
+    (scoped_call_expression
+      scope: (name) @scope (#eq? @scope "Route")
+      name: (name) @method (#match? @method "^(get|post|put|delete|patch)$")
+      arguments: (arguments . (argument (string) @path)))
+  `,
+};
+const HTTP_FACADE_SPEC = {
+    meta: {},
+    query: `
+    (scoped_call_expression
+      scope: (name) @scope (#eq? @scope "Http")
+      name: (name) @method (#match? @method "^(get|post|put|delete|patch)$")
+      arguments: (arguments . (argument (string) @path)))
+  `,
+};
+const GUZZLE_MEMBER_SPEC = {
+    meta: {},
+    query: `
+    (member_call_expression
+      name: (name) @method (#match? @method "^(get|post|put|delete|patch)$")
+      arguments: (arguments . (argument (string) @path)))
+  `,
+};
+const FILE_GET_CONTENTS_SPEC = {
+    meta: {},
+    query: `
+    (function_call_expression
+      function: (name) @fn (#eq? @fn "file_get_contents")
+      arguments: (arguments . (argument (string) @path)))
+  `,
+};
+const mk = (spec, suffix) => compilePatterns({
+    name: `php-${suffix}`,
     language: PHP.php_only,
-    patterns: [
-        {
-            meta: {},
-            query: `
-        (scoped_call_expression
-          scope: (name) @scope (#eq? @scope "Route")
-          name: (name) @method (#match? @method "^(get|post|put|delete|patch)$")
-          arguments: (arguments . (argument (string) @path)))
-      `,
-        },
-    ],
+    patterns: [spec],
 });
+const PHP_PATTERNS = {
+    laravelRoute: mk(LARAVEL_ROUTE_SPEC, 'laravel-route'),
+    httpFacade: mk(HTTP_FACADE_SPEC, 'http-facade'),
+    guzzleMember: mk(GUZZLE_MEMBER_SPEC, 'guzzle-member'),
+    fileGetContents: mk(FILE_GET_CONTENTS_SPEC, 'file-get-contents'),
+};
 /**
  * Extract the inner text of a PHP `string` node. The tree-sitter-php
  * grammar wraps single / double-quoted literals differently depending
@@ -30,12 +74,9 @@ const LARAVEL_PATTERNS = compilePatterns({
  * child nodes.
  */
 function phpStringText(node) {
-    // Most single-quoted strings expose their inner content through the
-    // full node text (including quotes), which unquoteLiteral strips.
     const direct = unquoteLiteral(node.text);
     if (direct !== null && direct !== node.text)
         return direct;
-    // Fall back to child string_content / string_value node if present.
     for (const child of node.children) {
         if (child.type === 'string_content' || child.type === 'string_value') {
             return child.text;
@@ -43,12 +84,29 @@ function phpStringText(node) {
     }
     return direct;
 }
+/**
+ * HTTP client helpers (`Http::`, Guzzle) are almost always called with
+ * a path relative to a configured base URL, or a full URL. File paths
+ * are rare. Accept both relative (`/api/...`) and absolute (`http(s)://`).
+ */
+function isHttpClientPath(path) {
+    return path.startsWith('/') || path.startsWith('http://') || path.startsWith('https://');
+}
+/**
+ * `file_get_contents` is used for both HTTP and filesystem reads. Only
+ * emit a consumer contract when the URL is an absolute HTTP(S) URL to
+ * avoid false positives for local file paths and stream wrappers
+ * (`php://input`, `file://`, `data:`, ...).
+ */
+function isHttpUrlLiteral(path) {
+    return path.startsWith('http://') || path.startsWith('https://');
+}
 export const PHP_HTTP_PLUGIN = {
     name: 'php-http',
     language: PHP.php_only,
     scan(tree) {
         const out = [];
-        for (const match of runCompiledPatterns(LARAVEL_PATTERNS, tree)) {
+        for (const match of runCompiledPatterns(PHP_PATTERNS.laravelRoute, tree)) {
             const methodNode = match.captures.method;
             const pathNode = match.captures.path;
             if (!methodNode || !pathNode)
@@ -65,6 +123,56 @@ export const PHP_HTTP_PLUGIN = {
                 confidence: 0.8,
             });
         }
+        for (const match of runCompiledPatterns(PHP_PATTERNS.httpFacade, tree)) {
+            const methodNode = match.captures.method;
+            const pathNode = match.captures.path;
+            if (!methodNode || !pathNode)
+                continue;
+            const path = phpStringText(pathNode);
+            if (path === null || !isHttpClientPath(path))
+                continue;
+            out.push({
+                role: 'consumer',
+                framework: 'laravel-http',
+                method: methodNode.text.toUpperCase(),
+                path,
+                name: null,
+                confidence: 0.7,
+            });
+        }
+        for (const match of runCompiledPatterns(PHP_PATTERNS.guzzleMember, tree)) {
+            const methodNode = match.captures.method;
+            const pathNode = match.captures.path;
+            if (!methodNode || !pathNode)
+                continue;
+            const path = phpStringText(pathNode);
+            if (path === null || !isHttpClientPath(path))
+                continue;
+            out.push({
+                role: 'consumer',
+                framework: 'guzzle',
+                method: methodNode.text.toUpperCase(),
+                path,
+                name: null,
+                confidence: 0.7,
+            });
+        }
+        for (const match of runCompiledPatterns(PHP_PATTERNS.fileGetContents, tree)) {
+            const pathNode = match.captures.path;
+            if (!pathNode)
+                continue;
+            const path = phpStringText(pathNode);
+            if (path === null || !isHttpUrlLiteral(path))
+                continue;
+            out.push({
+                role: 'consumer',
+                framework: 'file-get-contents',
+                method: 'GET',
+                path,
+                name: null,
+                confidence: 0.7,
+            });
+        }
         return out;
     },
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gitnexus",
-  "version": "1.6.3-rc.25",
+  "version": "1.6.3-rc.27",
   "description": "Graph-powered code intelligence for AI agents. Index any codebase, query via MCP or CLI.",
   "author": "Abhigyan Patwari",
   "license": "PolyForm-Noncommercial-1.0.0",