fix(python): resolve ClassName() calls to __init__

A Python construction call resolved to the Class node, so a user
__init__ never received a caller from ClassName(...) and the dead-code
view reported live constructors as unreferenced.

When the called name is the class itself, resolve it to the __init__
found on the class or, depth-first, its declared bases (the lookup
super().__init__() already uses), with a new lsp_constructor_init
strategy that joins the textual call site through its reason. Classes
with no user __init__ keep the Class target (lsp_constructor), as does
a builtins.* __init__. Calls through an instance (f = Foo(); f()) are
unchanged, since they dispatch to __call__.

This changes the documented Class-target contract for classes that
define __init__; the three tests that pinned it are updated.

Refs #1642

Signed-off-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>

internal/cbm/lsp/py_lsp.c

@@ -3249,7 +3249,26 @@ static void py_emit_call_for(PyLSPContext *ctx, TSNode call_node) {
                     py_emit_resolved_call_reason(ctx, qn, "lsp_import_alias", 0.85f, fname,
                                                  call_node);
                 } else if (qn && cbm_registry_lookup_type(ctx->registry, qn)) {
-                    py_emit_resolved_call(ctx, qn, "lsp_constructor", 0.85f, call_node);
+                    /* ClassName() runs a user __init__ found on the class or,
+                     * depth-first, a declared base (the same walk
+                     * super().__init__ uses, not C3 MRO). Retarget only when
+                     * this binding IS the class being called. An instance
+                     * (`f = Foo(); f()`) keeps the class row: that call
+                     * dispatches to __call__, and a reason-join would invent
+                     * a Foo.__init__ edge. A builtins.* __init__ is not a
+                     * project constructor, so the class target stays. */
+                    const CBMRegisteredFunc *fi = NULL;
+                    if (qn_short && strcmp(qn_short, fname) == 0) {
+                        fi = py_lookup_attribute(ctx, qn, "__init__");
+                    }
+                    if (fi && fi->qualified_name &&
+                        strncmp(fi->qualified_name, "builtins.", sizeof("builtins.") - 1) != 0) {
+                        py_emit_resolved_call_reason(ctx, fi->qualified_name,
+                                                     "lsp_constructor_init", 0.85f, fname,
+                                                     call_node);
+                    } else {
+                        py_emit_resolved_call(ctx, qn, "lsp_constructor", 0.85f, call_node);
+                    }
                 }
             }
             return;

src/pipeline/lsp_resolve.h

@@ -388,6 +388,7 @@ static inline bool cbm_pipeline_invocation_reason_join_strategy(const char *stra
             strcmp(strategy, "lsp_method_ref_ctor_synth") == 0 ||
             strcmp(strategy, "lsp_dict_dispatch") == 0 ||
             strcmp(strategy, "lsp_import_alias") == 0 ||
+            strcmp(strategy, "lsp_constructor_init") == 0 ||
             strcmp(strategy, "php_method_dynamic") == 0);
 }
 

tests/repro/repro_lsp_go_py.c

@@ -64,8 +64,10 @@
  * PYTHON STRATEGY INVENTORY — every literal "lsp_..." emitted by py_lsp.c
  *   (grep '"lsp_' internal/cbm/lsp/py_lsp.c), with its keying site:
  *     lsp_direct                (py_lsp.c:1631)  module-local f()
- *     lsp_constructor           (py_lsp.c:1624)  ClassName() where the name is a
- *                                                NAMED type in scope
+ *     lsp_constructor           ClassName() with no user __init__ → the Class
+ *     lsp_constructor_init      ClassName() with a user __init__ (on the class
+ *                               or a declared base) → that __init__. The textual
+ *                               callee leaf stays the class name; reason joins it.
  *     lsp_method                (py_lsp.c:1731)  obj.method() on a NAMED-typed
  *                                                receiver (covers self.other())
  *     lsp_super                 (py_lsp.c:1693)  super().method() resolved on a
@@ -340,13 +342,24 @@ static const char kPyDirect[] =
     "def caller(v):\n"
     "    return helper(v)\n";
 
-/* lsp_constructor — ClassName() where the name is a NAMED type in scope
- * (py_lsp.c:1620-1624: cbm_scope_lookup yields a NAMED type → emit constructor
- * edge to the class QN). */
+/* lsp_constructor — ClassName() where the class has no user __init__.
+ * The edge stays on the Class node. A class that defines __init__ is
+ * lsp_constructor_init below; strstr would otherwise treat that strategy
+ * as a hit for "lsp_constructor". */
 static const char kPyConstructor[] =
+    "class Widget:\n"
+    "    def label(self):\n"
+    "        return 1\n"
+    "def caller():\n"
+    "    return Widget()\n";
+
+/* lsp_constructor_init — ClassName() resolves to the user __init__
+ * (issue #1642). reason carries the class-name leaf so the pipeline join
+ * still matches the textual call. */
+static const char kPyConstructorInit[] =
     "class Widget:\n"
     "    def __init__(self):\n"
-    "        pass\n"
+    "        self.ready = True\n"
     "def caller():\n"
     "    return Widget()\n";
 
@@ -538,6 +551,10 @@ TEST(repro_lsp_py_constructor) {
     return assert_lsp_strategy("main.py", kPyConstructor, "lsp_constructor");
 }
 
+TEST(repro_lsp_py_constructor_init) {
+    return assert_lsp_strategy("main.py", kPyConstructorInit, "lsp_constructor_init");
+}
+
 TEST(repro_lsp_py_method) {
     return assert_lsp_strategy("main.py", kPyMethod, "lsp_method");
 }
@@ -623,6 +640,7 @@ SUITE(repro_lsp_go_py) {
 
     RUN_TEST(repro_lsp_py_direct);
     RUN_TEST(repro_lsp_py_constructor);
+    RUN_TEST(repro_lsp_py_constructor_init);
     RUN_TEST(repro_lsp_py_method);
     RUN_TEST(repro_lsp_py_super);
     RUN_TEST(repro_lsp_py_super_init);

tests/test_edge_structural.c

@@ -826,14 +826,31 @@ TEST(es_decorates_attribute_csharp) {
  * resolves in the registry.  pass_usages.c handles this.
  * ══════════════════════════════════════════════════════════════════ */
 
-/* Python constructor syntax is a CALLS edge to the materialized Class node. */
+/* Python ClassName() with a user __init__ is a CALLS edge to that method.
+ * The Class node no longer receives the instantiation edge. A class with no
+ * user __init__ still targets the Class (lsp_constructor). */
 TEST(es_usage_crossfile_python) {
     static const ES_LangFile f[] = {
         {"models.py",
          "class User:\n    def __init__(self, name):\n        self.name = name\n"},
         {"main.py",
          "from .models import User\n\n\ndef create_user(name):\n    return User(name)\n"}};
-    ASSERT_EQ(es_exact_edge_by_name(f, 2, "CALLS", "create_user", "User"), 1);
+    ASSERT_EQ(es_exact_edge_by_name(f, 2, "CALLS", "create_user", "__init__"), 1);
+    ASSERT_EQ(es_exact_edge_by_name(f, 2, "CALLS", "create_user", "User"), 0);
+    PASS();
+}
+
+/* Issue #1642 repro: build() -> Embedder.__init__, not the Embedder class. */
+TEST(es_calls_python_constructor_init) {
+    static const ES_LangFile f[] = {
+        {"main.py", "class Embedder:\n"
+                    "    def __init__(self, model=\"x\"):\n"
+                    "        self.model = model\n"
+                    "\n"
+                    "def build():\n"
+                    "    return Embedder(model=\"y\")\n"}};
+    ASSERT_EQ(es_exact_edge_by_name(f, 1, "CALLS", "build", "__init__"), 1);
+    ASSERT_EQ(es_exact_edge_by_name(f, 1, "CALLS", "build", "Embedder"), 0);
     PASS();
 }
 
@@ -982,6 +999,7 @@ SUITE(edge_structural) {
     /* ── FAMILY 7: USAGE cross-file ─────────────────────────── */
     /* Uncertain: depends on USAGE extraction for each language. */
     RUN_TEST(es_usage_crossfile_python);
+    RUN_TEST(es_calls_python_constructor_init);
     RUN_TEST(es_usage_crossfile_typescript);
     RUN_TEST(es_usage_crossfile_go);
 

tests/test_lsp_resolution_probe.c

@@ -1837,16 +1837,19 @@ TEST(lrp_tsx_usage_import_type_with_decoy) {
     PASS();
 }
 
-/* Python: constructor syntax is an exact call to the materialized Class node. */
+/* Python: ClassName() with a user __init__ is an exact CALLS edge to that
+ * method. The Class node no longer receives the instantiation edge. */
 TEST(lrp_python_usage_instantiation) {
     static const LRP_File f[] = {
         {"model.py", "class User:\n    def __init__(self, name):\n        self.name = name\n"},
         {"main.py", "from .model import User\n\n\ndef create(name):\n    return User(name)\n"}};
     LRP_Proj lp;
     cbm_store_t *store = lrp_index(&lp, f, 2);
-    int exact = store ? lrp_exact_calls_by_name(store, lp.project, "create", "User") : -1;
+    int exact = store ? lrp_exact_calls_by_name(store, lp.project, "create", "__init__") : -1;
+    int class_edge = store ? lrp_exact_calls_by_name(store, lp.project, "create", "User") : -1;
     lrp_cleanup(&lp, store);
     ASSERT_EQ(exact, 1);
+    ASSERT_EQ(class_edge, 0);
     PASS();
 }
 

tests/test_py_lsp.c

@@ -356,6 +356,164 @@ TEST(pylsp_constructor_call_returns_instance) {
     PASS();
 }
 
+/* Issue #1642: Embedder(...) must resolve to Embedder.__init__, and the
+ * textual callee leaf "Embedder" must still join that row. */
+TEST(pylsp_constructor_call_targets_init) {
+    const char *source = "class Embedder:\n"
+                         "    def __init__(self, model=\"x\"):\n"
+                         "        self.model = model\n"
+                         "def build():\n"
+                         "    return Embedder(model=\"y\")\n";
+    CBMFileResult *r = extract_py(source);
+    ASSERT_NOT_NULL(r);
+    int idx = require_resolved(r, "build", "Embedder.__init__");
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[idx];
+        ASSERT_STR_EQ(rc->strategy, "lsp_constructor_init");
+        ASSERT_STR_EQ(rc->reason, "Embedder");
+        ASSERT(strncmp(rc->callee_qn, "builtins.", 9) != 0);
+    }
+    for (int i = 0; i < r->resolved_calls.count; i++) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[i];
+        if (rc->kind != CBM_RESOLVED_INVOCATION || !rc->caller_qn || !rc->callee_qn ||
+            !strstr(rc->caller_qn, "build"))
+            continue;
+        /* One site, one target: the class node no longer receives the call. */
+        ASSERT(strstr(rc->callee_qn, ".__init__") != NULL);
+        ASSERT(strncmp(rc->callee_qn, "builtins.", 9) != 0);
+    }
+
+    const CBMCall *carrier = NULL;
+    for (int i = 0; i < r->calls.count; i++) {
+        const CBMCall *call = &r->calls.items[i];
+        if (call->callee_name && strcmp(call->callee_name, "Embedder") == 0 &&
+            call->enclosing_func_qn && strstr(call->enclosing_func_qn, "build")) {
+            carrier = call;
+            break;
+        }
+    }
+    ASSERT_NOT_NULL(carrier);
+    const CBMResolvedCall *joined =
+        cbm_pipeline_find_lsp_resolution(&r->resolved_calls, carrier, false);
+    ASSERT_NOT_NULL(joined);
+    ASSERT(strstr(joined->callee_qn, "Embedder.__init__") != NULL);
+    ASSERT_STR_EQ(joined->strategy, "lsp_constructor_init");
+    cbm_free_result(r);
+    PASS();
+}
+
+/* class B(A) with only A.__init__ resolves construction of B to A.__init__. */
+TEST(pylsp_constructor_inherited_init) {
+    CBMFileResult *r = extract_py("class A:\n"
+                                  "    def __init__(self):\n"
+                                  "        self.ready = True\n"
+                                  "class B(A):\n"
+                                  "    def go(self):\n"
+                                  "        return 1\n"
+                                  "def build():\n"
+                                  "    return B()\n");
+    ASSERT_NOT_NULL(r);
+    int idx = require_resolved(r, "build", "A.__init__");
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[idx];
+        ASSERT_STR_EQ(rc->strategy, "lsp_constructor_init");
+        ASSERT(strstr(rc->callee_qn, "B.__init__") == NULL);
+        ASSERT(strncmp(rc->callee_qn, "builtins.", 9) != 0);
+    }
+    ASSERT_EQ(find_resolved(r, "build", "B.__init__"), -1);
+    cbm_free_result(r);
+    PASS();
+}
+
+/* No user __init__ (and no inherited one): the call stays on the class. */
+TEST(pylsp_constructor_without_init_stays_on_class) {
+    CBMFileResult *r = extract_py("class Bare:\n"
+                                  "    def method(self):\n"
+                                  "        return 1\n"
+                                  "def build():\n"
+                                  "    return Bare()\n");
+    ASSERT_NOT_NULL(r);
+    ASSERT_EQ(find_resolved(r, "build", "__init__"), -1);
+    int idx = -1;
+    for (int i = 0; i < r->resolved_calls.count; i++) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[i];
+        if (rc->kind == CBM_RESOLVED_INVOCATION && rc->caller_qn &&
+            strstr(rc->caller_qn, "build") && rc->callee_qn && strstr(rc->callee_qn, "Bare") &&
+            !strstr(rc->callee_qn, "__init__")) {
+            idx = i;
+            break;
+        }
+    }
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0)
+        ASSERT_STR_EQ(r->resolved_calls.items[idx].strategy, "lsp_constructor");
+    cbm_free_result(r);
+    PASS();
+}
+
+/* f = Foo(); f() is a callable instance, not a constructor. It must not
+ * gain an edge to Foo.__init__. Foo() itself still does. */
+TEST(pylsp_callable_instance_call_skips_init) {
+    CBMFileResult *r = extract_py("class Foo:\n"
+                                  "    def __init__(self):\n"
+                                  "        pass\n"
+                                  "    def __call__(self):\n"
+                                  "        return 1\n"
+                                  "def use():\n"
+                                  "    f = Foo()\n"
+                                  "    return f()\n");
+    ASSERT_NOT_NULL(r);
+    int init_rows = 0;
+    int class_rows = 0;
+    for (int i = 0; i < r->resolved_calls.count; i++) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[i];
+        if (rc->kind != CBM_RESOLVED_INVOCATION || !rc->caller_qn || !rc->callee_qn ||
+            !strstr(rc->caller_qn, "use"))
+            continue;
+        if (strstr(rc->callee_qn, "__init__")) {
+            init_rows++;
+            ASSERT_STR_EQ(rc->strategy, "lsp_constructor_init");
+            ASSERT_STR_EQ(rc->reason, "Foo");
+            ASSERT(strncmp(rc->callee_qn, "builtins.", 9) != 0);
+        } else if (strstr(rc->callee_qn, "Foo")) {
+            class_rows++;
+            ASSERT_STR_EQ(rc->strategy, "lsp_constructor");
+            ASSERT_NULL(rc->reason);
+        }
+    }
+    ASSERT_EQ(init_rows, 1);
+    ASSERT_EQ(class_rows, 1);
+    cbm_free_result(r);
+    PASS();
+}
+
+/* builtins.object.__init__ is in the stdlib registry. Construction must not
+ * record a resolution to it. */
+TEST(pylsp_constructor_skips_builtin_init) {
+    CBMFileResult *r = extract_py("from builtins import object\n"
+                                  "def build():\n"
+                                  "    return object()\n");
+    ASSERT_NOT_NULL(r);
+    for (int i = 0; i < r->resolved_calls.count; i++) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[i];
+        if (!rc->callee_qn)
+            continue;
+        ASSERT(strncmp(rc->callee_qn, "builtins.", 9) != 0 ||
+               strstr(rc->callee_qn, ".__init__") == NULL);
+    }
+    int idx = find_resolved(r, "build", "builtins.object");
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0) {
+        const CBMResolvedCall *rc = &r->resolved_calls.items[idx];
+        ASSERT_STR_EQ(rc->strategy, "lsp_constructor");
+        ASSERT(strstr(rc->callee_qn, "__init__") == NULL);
+    }
+    cbm_free_result(r);
+    PASS();
+}
+
 TEST(pylsp_method_via_inheritance) {
     CBMFileResult *r = extract_py(
         "class Base:\n"
@@ -924,6 +1082,103 @@ TEST(pylsp_crossfile_inheritance) {
     PASS();
 }
 
+/* Imported ClassName() resolves to that class's __init__. Bare names are
+ * module-qualified during per-file registration, so the imported class has
+ * to come from the cross-file registry. */
+TEST(pylsp_crossfile_constructor_targets_init) {
+    const char *source = "from models import Embedder\n"
+                         "def build():\n"
+                         "    return Embedder(model=\"y\")\n";
+
+    CBMLSPDef defs[2];
+    memset(defs, 0, sizeof(defs));
+    defs[0].qualified_name = "models.Embedder";
+    defs[0].short_name = "Embedder";
+    defs[0].label = "Class";
+    defs[0].def_module_qn = "models";
+    defs[0].lang = CBM_LANG_PYTHON;
+    defs[1].qualified_name = "models.Embedder.__init__";
+    defs[1].short_name = "__init__";
+    defs[1].label = "Method";
+    defs[1].receiver_type = "models.Embedder";
+    defs[1].def_module_qn = "models";
+    defs[1].lang = CBM_LANG_PYTHON;
+
+    const char *imp_names[] = {"Embedder"};
+    const char *imp_qns[] = {"models.Embedder"};
+
+    CBMArena arena;
+    cbm_arena_init(&arena);
+    CBMTypeRegistry *reg = cbm_py_build_cross_registry(&arena, defs, 2);
+    ASSERT_NOT_NULL(reg);
+    CBMResolvedCallArray out = {0};
+    cbm_run_py_lsp_cross_with_registry(&arena, source, (int)strlen(source), "test.main", reg,
+                                       imp_names, imp_qns, 1, NULL, &out, NULL);
+
+    int idx = find_resolved_arr(&out, "build", "models.Embedder.__init__");
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0) {
+        ASSERT_STR_EQ(out.items[idx].strategy, "lsp_constructor_init");
+        ASSERT_STR_EQ(out.items[idx].reason, "Embedder");
+        ASSERT(strncmp(out.items[idx].callee_qn, "builtins.", 9) != 0);
+    }
+    cbm_arena_destroy(&arena);
+    PASS();
+}
+
+/* class B(A) where A's __init__ lives in another module. The cross registry
+ * records the real base QN; per-file registration would have stored
+ * test.main.A and missed the base. */
+TEST(pylsp_crossfile_inherited_init) {
+    const char *source = "from base_mod import A\n"
+                         "class B(A):\n"
+                         "    def go(self):\n"
+                         "        return 1\n"
+                         "def build():\n"
+                         "    return B()\n";
+
+    CBMLSPDef defs[3];
+    memset(defs, 0, sizeof(defs));
+    defs[0].qualified_name = "base_mod.A";
+    defs[0].short_name = "A";
+    defs[0].label = "Class";
+    defs[0].def_module_qn = "base_mod";
+    defs[0].lang = CBM_LANG_PYTHON;
+    defs[1].qualified_name = "base_mod.A.__init__";
+    defs[1].short_name = "__init__";
+    defs[1].label = "Method";
+    defs[1].receiver_type = "base_mod.A";
+    defs[1].def_module_qn = "base_mod";
+    defs[1].lang = CBM_LANG_PYTHON;
+    defs[2].qualified_name = "test.main.B";
+    defs[2].short_name = "B";
+    defs[2].label = "Class";
+    defs[2].def_module_qn = "test.main";
+    defs[2].embedded_types = "base_mod.A";
+    defs[2].lang = CBM_LANG_PYTHON;
+
+    const char *imp_names[] = {"A"};
+    const char *imp_qns[] = {"base_mod.A"};
+
+    CBMArena arena;
+    cbm_arena_init(&arena);
+    CBMTypeRegistry *reg = cbm_py_build_cross_registry(&arena, defs, 3);
+    ASSERT_NOT_NULL(reg);
+    CBMResolvedCallArray out = {0};
+    cbm_run_py_lsp_cross_with_registry(&arena, source, (int)strlen(source), "test.main", reg,
+                                       imp_names, imp_qns, 1, NULL, &out, NULL);
+
+    int idx = find_resolved_arr(&out, "build", "base_mod.A.__init__");
+    ASSERT_GTE(idx, 0);
+    if (idx >= 0) {
+        ASSERT_STR_EQ(out.items[idx].strategy, "lsp_constructor_init");
+        ASSERT_STR_EQ(out.items[idx].reason, "B");
+        ASSERT(strstr(out.items[idx].callee_qn, "B.__init__") == NULL);
+    }
+    cbm_arena_destroy(&arena);
+    PASS();
+}
+
 TEST(pylsp_batch_two_files) {
     const char *src_a =
         "def helper():\n"
@@ -5010,6 +5265,11 @@ SUITE(py_lsp) {
     RUN_TEST(pylsp_method_call_simple);
     RUN_TEST(pylsp_method_via_self);
     RUN_TEST(pylsp_constructor_call_returns_instance);
+    RUN_TEST(pylsp_constructor_call_targets_init);
+    RUN_TEST(pylsp_constructor_inherited_init);
+    RUN_TEST(pylsp_constructor_without_init_stays_on_class);
+    RUN_TEST(pylsp_callable_instance_call_skips_init);
+    RUN_TEST(pylsp_constructor_skips_builtin_init);
     RUN_TEST(pylsp_method_via_inheritance);
     RUN_TEST(pylsp_no_false_positive_on_unknown_method);
     /* Phases 7-8 — decorators, super(), multi-inheritance */
@@ -5028,6 +5288,8 @@ SUITE(py_lsp) {
     RUN_TEST(pylsp_cross_registry_interleaved_order);
     RUN_TEST(pylsp_crossfile_classmethod_on_class_issue228);
     RUN_TEST(pylsp_crossfile_inheritance);
+    RUN_TEST(pylsp_crossfile_constructor_targets_init);
+    RUN_TEST(pylsp_crossfile_inherited_init);
     RUN_TEST(pylsp_batch_two_files);
     RUN_TEST(pylsp_from_import_alias_equal_module_leaf_targets_imported_member);
     RUN_TEST(pylsp_project_prefixed_direct_alias_same_tail_is_not_from_import_reference);