fix(extract): resolve bash calls into sourced-file functions (#2141)

graphify/extract.py

@@ -133,6 +133,8 @@
     _workspace_globs,
 )
 
+from graphify.symbol_resolution import resolve_bash_source_edges  # noqa: E402
+
 from graphify.extractors.engine import REFERENCE_CONTEXTS, _CSHARP_TYPE_PARAMETER_SCOPE_DECLARATIONS, _C_PRIMITIVE_TYPE_NODES, _JAVA_BUILTIN_TYPES, _JAVA_TYPE_PARAMETER_SCOPE_DECLARATIONS, _JS_FUNCTION_VALUE_TYPES, _JS_SCOPE_BOUNDARY, _PYTHON_ANNOTATION_NOISE, _PYTHON_TYPE_CONTAINERS, _RUBY_CLASS_FACTORIES, _c_collect_type_refs, _cpp_collect_type_refs, _cpp_declarator_name, _cpp_local_var_types, _csharp_attribute_names, _csharp_classify_base, _csharp_collect_type_refs, _csharp_extra_walk, _csharp_member_type_table, _csharp_namespace_id, _csharp_namespace_name, _csharp_pre_scan_interfaces, _csharp_type_parameters_in_scope, _dynamic_import_js, _extract_generic, _find_body, _find_require_call, _get_cpp_func_name, _java_annotation_names, _java_collect_type_refs, _java_extra_walk, _java_type_parameters_in_scope, _js_collect_pattern_idents, _js_dispatch_value_idents, _js_extra_walk, _js_local_bound_names, _js_member_assignment_target, _js_module_bound_names, _kotlin_collect_type_refs, _kotlin_function_return_type_node, _kotlin_property_type_node, _kotlin_user_type_name, _php_collect_type_refs, _php_method_return_type_node, _php_name_text, _python_collect_assignment_targets, _python_collect_param_refs, _python_collect_type_refs, _python_local_bound_names, _python_module_bound_names, _python_param_names, _read_csharp_type_name, _require_imports_js, _ruby_const_last_name, _ruby_extra_walk, _ruby_local_class_bindings, _ruby_new_class_name, _scala_collect_type_refs, _semantic_reference_edge, _source_location, _swift_classify_base, _swift_collect_type_refs, _swift_constructor_type, _swift_declaration_keyword, _swift_extra_walk, _swift_local_var_types, _swift_pre_scan, _swift_property_name, _swift_property_type_node, _swift_receiver_name, _swift_user_type_name, _ts_decorator_name, _ts_descendant_decorators, _ts_emit_decorator_edges, _ts_extra_walk, _ts_method_name, _ts_receiver_type_table  # noqa: E402,F401
 
 from graphify.extractors.pascal import _PAS_BEGIN_END_TOKEN_RE, _PAS_CALL_RE, _PAS_END_SEMI_RE, _PAS_IMPL_HEADER_RE, _PAS_KEYWORDS, _PAS_METHOD_DECL_RE, _PAS_MODULE_RE, _PAS_TOKEN_RE, _PAS_TYPE_HEADER_RE, _PAS_USES_RE, _extract_pascal_regex, _pascal_find_body, _pascal_split_bases, _pascal_split_sections, _pascal_split_uses, _pascal_strip_comments, extract_pascal  # noqa: E402,F401
@@ -4937,6 +4939,25 @@ def _learn(e: dict) -> None:
             import logging
             logging.getLogger(__name__).warning("C# cross-file import resolution failed, skipping: %s", exc)
 
+    # Cross-file Bash source-backed call resolution: a call to a function defined
+    # in a file this one `source`s is left unresolved by the per-file extractor
+    # (it only links calls to same-file functions, #2141). Match each bash raw_call
+    # against functions in the sourced files and emit the calls edge — scoped to
+    # the source relationship, so a call to an external command never binds to a
+    # same-named function in an unsourced file. Runs after the id-remap passes
+    # above so caller_nids and function node ids are final; dedups the source
+    # edge the extractor already emitted via existing_edges.
+    sh_paths = [p for p in paths if p.suffix in (".sh", ".bash")]
+    if sh_paths:
+        sh_results = [r for r, p in zip(per_file, paths) if p.suffix in (".sh", ".bash")]
+        try:
+            all_edges.extend(
+                resolve_bash_source_edges(sh_results, sh_paths, root, existing_edges=all_edges)
+            )
+        except Exception as exc:
+            import logging
+            logging.getLogger(__name__).warning("Bash cross-file call resolution failed, skipping: %s", exc)
+
     # Cross-file call resolution for all languages
     # Each extractor saved unresolved calls in raw_calls. Now that we have all
     # nodes from all files, resolve any callee that exists in another file.
@@ -5047,6 +5068,14 @@ def _learn(e: dict) -> None:
         # both mislabel the relation and block the mixes_in emit as a dup (#1668).
         if rc.get("is_mixin"):
             continue
+        # Bash calls are resolved only by resolve_bash_source_edges (run above),
+        # which scopes resolution to the files a script actually `source`s. The
+        # global name match here would bind a bash call to any same-named function
+        # in an unsourced file (an INFERRED phantom edge) and would resolve calls
+        # to external commands that merely share a name with a function elsewhere
+        # in the corpus — exactly what #2141 must not do.
+        if rc.get("language") == "bash":
+            continue
         # Exact-case match first (case is semantic). Fold only when the CALLING
         # file's language is case-insensitive, and only against the folded index of
         # case-insensitive-language definitions — so a Python `Path()` call can never

graphify/extractors/bash.py

@@ -28,6 +28,14 @@ def extract_bash(path: Path) -> dict:
     str_path = str(path)
     nodes: list[dict] = []
     edges: list[dict] = []
+    # Cross-file resolution scaffolding consumed by resolve_bash_source_edges in
+    # the extract pipeline: `bash_sources` records which files this one `source`s,
+    # `raw_calls` records calls whose callee isn't defined in this file (candidate
+    # calls into a sourced library). The extractor sees one file at a time and so
+    # can't resolve these itself (#2141).
+    raw_calls: list[dict] = []
+    bash_sources: list[dict] = []
+    raw_seen: set[tuple[str, str]] = set()
     seen_ids: set[str] = set()
     function_bodies: list[tuple[str, Any]] = []
     defined_functions: set[str] = set()
@@ -129,6 +137,24 @@ def walk_calls(body_node, func_nid: str, seen_calls: set) -> None:
                             add_edge(func_nid, tgt, "calls",
                                      child.start_point[0] + 1,
                                      confidence="EXTRACTED", context="call")
+                    elif (name and name not in _BASH_SOURCE_COMMANDS
+                          and name not in _BASH_SCRIPT_RUNNERS):
+                        # Callee isn't defined in this file — it may be a function
+                        # from a sourced library. Record an unresolved raw_call for
+                        # resolve_bash_source_edges to bind against the files this
+                        # script `source`s. A callee that is not a sourced function
+                        # (a genuine external command) matches nothing there and
+                        # yields no edge, so this can't over-connect the graph (#2141).
+                        raw_key = (func_nid, name)
+                        if raw_key not in raw_seen:
+                            raw_seen.add(raw_key)
+                            raw_calls.append({
+                                "language": "bash",
+                                "callee": name,
+                                "caller_nid": func_nid,
+                                "source_file": str_path,
+                                "source_location": f"L{child.start_point[0] + 1}",
+                            })
             walk_calls(child, func_nid, seen_calls)
 
     def walk(node, parent_nid: str) -> None:
@@ -180,6 +206,15 @@ def walk(node, parent_nid: str) -> None:
                                 tgt_nid = _make_id(str(resolved))
                                 add_edge(file_nid, tgt_nid, "imports_from", line,
                                          context="import")
+                                # Record the sourced file so resolve_bash_source_edges
+                                # can bind calls into its functions (#2141). Gated on
+                                # existence like the edge above, so crafted traversal
+                                # paths never enter the resolver's data.
+                                bash_sources.append({
+                                    "target_path": raw,
+                                    "source_file": str_path,
+                                    "source_location": f"L{line}",
+                                })
                         else:
                             tgt_nid = _make_id(raw)
                             if tgt_nid:
@@ -245,4 +280,5 @@ def _prescan_functions(node) -> None:
     for fn_nid, body in function_bodies:
         walk_calls(body, fn_nid, set())
 
-    return {"nodes": nodes, "edges": edges}
+    return {"nodes": nodes, "edges": edges,
+            "raw_calls": raw_calls, "bash_sources": bash_sources}

tests/test_extract.py

@@ -1842,6 +1842,80 @@ def test_extract_bash_source_user_defined_emits_calls_not_imports_from(tmp_path)
     )
 
 
+def test_extract_bash_emits_raw_calls_and_bash_sources_for_sourced_calls(tmp_path):
+    """extract_bash must surface the data cross-file resolution needs: a
+    ``bash_sources`` entry per sourced file and a ``raw_calls`` entry for each
+    call whose callee is not defined in the same file. Without these,
+    resolve_bash_source_edges has nothing to resolve a sourced-function call
+    from (#2141)."""
+    (tmp_path / "b.sh").write_text("#!/usr/bin/env bash\nb_func() { echo ok; }\n")
+    a = tmp_path / "a.sh"
+    a.write_text(
+        "#!/usr/bin/env bash\n"
+        "source ./b.sh\n"
+        "main() { b_func; }\n"
+    )
+    result = extract_bash(a)
+
+    sources = result.get("bash_sources", [])
+    assert any(str(s.get("target_path", "")).endswith("b.sh") for s in sources), sources
+
+    main_nid = next(n["id"] for n in result["nodes"] if n.get("label") == "main()")
+    raw_calls = result.get("raw_calls", [])
+    assert any(
+        rc.get("language") == "bash"
+        and rc.get("callee") == "b_func"
+        and rc.get("caller_nid") == main_nid
+        for rc in raw_calls
+    ), raw_calls
+
+
+def test_extract_bash_call_to_sourced_function_resolves(tmp_path):
+    """#2141 repro: a call to a function defined in a sourced file must produce a
+    real ``calls`` edge through the full extract() pipeline, so ``path`` and
+    ``callers`` can traverse it."""
+    (tmp_path / "b.sh").write_text("#!/usr/bin/env bash\nb_func() { echo ok; }\n")
+    (tmp_path / "a.sh").write_text(
+        "#!/usr/bin/env bash\n"
+        "source ./b.sh\n"
+        "main() { b_func; }\n"
+    )
+    result = extract([tmp_path / "a.sh", tmp_path / "b.sh"], cache_root=tmp_path)
+    calls = {(e["source"], e["target"]) for e in result["edges"] if e["relation"] == "calls"}
+    assert ("a_main", "b_b_func") in calls, sorted(calls)
+
+
+def test_extract_bash_sourced_call_does_not_duplicate_source_edge(tmp_path):
+    """Wiring the source-backed call resolver must not re-emit the ``imports_from``
+    source edge the extractor already resolved for ``source ./b.sh`` (#2141)."""
+    (tmp_path / "b.sh").write_text("#!/usr/bin/env bash\nb_func() { echo ok; }\n")
+    (tmp_path / "a.sh").write_text(
+        "#!/usr/bin/env bash\n"
+        "source ./b.sh\n"
+        "main() { b_func; }\n"
+    )
+    result = extract([tmp_path / "a.sh", tmp_path / "b.sh"], cache_root=tmp_path)
+    imports = [(e["source"], e["target"]) for e in result["edges"]
+               if e["relation"] == "imports_from"]
+    assert imports.count(("a", "b")) == 1, imports
+
+
+def test_extract_bash_call_to_external_command_stays_unlinked(tmp_path):
+    """A call to a command that is not a function in any sourced file (an external
+    binary) must not gain a cross-file ``calls`` edge — even when a same-named
+    function exists in an *unsourced* file. Source-scoped resolution is what keeps
+    #2141 from over-connecting the graph (acceptance criterion)."""
+    # b.sh is NOT sourced by a.sh, yet defines a function named `deploy`.
+    (tmp_path / "b.sh").write_text("#!/usr/bin/env bash\ndeploy() { echo ok; }\n")
+    (tmp_path / "a.sh").write_text(
+        "#!/usr/bin/env bash\n"
+        "main() { deploy; }\n"
+    )
+    result = extract([tmp_path / "a.sh", tmp_path / "b.sh"], cache_root=tmp_path)
+    calls = {(e["source"], e["target"]) for e in result["edges"] if e["relation"] == "calls"}
+    assert ("a_main", "b_deploy") not in calls, sorted(calls)
+
+
 # ---------------------------------------------------------------------------
 # JSON extractor tests (#866)
 # ---------------------------------------------------------------------------