cjm-python-decompose-core 0.0.6__tar.gz → 0.0.7__tar.gz

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cjm-python-decompose-core
-Version: 0.0.6
+Version: 0.0.7
 Summary: A Python source-code decomposition core for context graphs: parses .py modules (docstrings, imports, and the symbol tree of functions/classes/methods with their calls) into provenance-carrying graph nodes and edges that co-reside with the markdown source-type. First source = the ecosystem's own code; generalizes to any Python corpus.
 Author-email: "Christian J. Mills" <9126128+cj-mills@users.noreply.github.com>
 License: Apache-2.0
@@ -10,7 +10,7 @@ Classifier: Programming Language :: Python :: 3 :: Only
 Requires-Python: >=3.12
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: cjm-dev-graph-schema>=0.0.8
+Requires-Dist: cjm-dev-graph-schema>=0.0.28
 Requires-Dist: cjm-context-graph-layer>=0.0.7
 Requires-Dist: cjm-context-graph-primitives>=0.0.12
 Dynamic: license-file
@@ -28,12 +28,6 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - **`cjm_python_decompose_core.extract`** — Bind parsed Python onto dev-graph-schema code nodes (the dev-domain binding).
 - **`cjm_python_decompose_core.ingest`** — Flatten a decomposed Python corpus into graph elements (the queue-free half).
 - **`cjm_python_decompose_core.parse`** — Schema-free Python parsing (stdlib `ast`).
-- **`tests.test_extract_ingest`** — Extract (parse -> schema nodes) + ingest (corpus -> graph elements) binding.
-- **`tests.test_imports_emit`** — Imports-as-projection: deriving a module's canonical import block from bindings.
-- **`tests.test_method_attribution`** — First-param capture + monkey-patch-assignment detection (the cross-cell method idioms).
-- **`tests.test_parse`** — Schema-free Python parsing: docstrings, imports, the symbol tree, and calls.
-- **`tests.test_regions_emit`** — Verbatim region decomposition + canonical emit (the authoring / round-trip substrate).
-- **`tests_manual.imports_projection`** — Imports-as-projection faithfulness sweep over the real arc libs (corpus harness).
 
 ## API
 
@@ -76,72 +70,7 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - `parse_module` _function_ — Parse Python source into module docstring + imports + the symbol tree.
 - `parse_regions` _function_ — Decompose a module into ordered verbatim top-level regions (the round-trip substrate).
 
-### `tests.test_extract_ingest`
-
-- `test_corpus_imports_and_calls_resolve_intra_corpus` _function_
-- `test_decompose_text_builds_module_and_symbols` _function_
-- `test_external_imports_not_minted` _function_
-- `test_import_map_and_call_map` _function_
-- `test_ingest_is_idempotent_in_ids` _function_
-- `test_local_edges_about_repo_and_defines_nesting` _function_
-- `test_module_path_and_import_name` _function_
-- `test_resolve_import_relative_and_absolute` _function_
-
-### `tests.test_imports_emit`
-
-- `test_coexisting_submodule_imports_both_survive_round_trip` _function_ — import urllib.request + import urllib.error both bind `urllib` and both are
-- `test_derived_block_is_faithful_and_prunes_dead_imports` _function_
-- `test_docstring_lines_starting_with_from_survive_derive` _function_
-- `test_emit_module_default_is_verbatim_not_derived` _function_
-- `test_emit_module_derive_imports_prunes_dead_and_keeps_bodies` _function_
-- `test_parameter_default_reference_is_not_pruned` _function_
-- `test_rebinding_imports_still_supersede` _function_ — Aliased / from-imports rebinding a name follow Python last-binding-wins —
-- `test_render_binding_forms` _function_
-- `test_render_import_block_orders_dedups_and_groups` _function_
-- `test_render_import_block_wraps_long_from_lines_in_aligned_parens` _function_
-- `test_same_module_imported_twice_dedupes` _function_ — A literally duplicated plain import collapses to one statement.
-
-### `tests.test_method_attribution`
-
-- `test_first_param_and_annotation_captured` _function_
-- `test_first_param_forward_ref_annotation` _function_
-- `test_first_param_unannotated_and_classes_have_none` _function_
-- `test_monkeypatch_assignments_detected` _function_
-- `test_monkeypatch_is_structural_filtering_is_downstream` _function_
-- `test_no_arg_function_has_empty_first_param` _function_
-
-### `tests.test_parse`
-
-- `parsed` _function_
-- `test_class_methods_are_children_with_method_kind` _function_
-- `test_empty_and_docstringless_module` _function_
-- `test_function_calls_are_direct_only` _function_
-- `test_import_bindings_per_symbol_and_module_level` _function_
-- `test_imports_collected_with_relative_and_submodule_forms` _function_
-- `test_iter_symbols_flattens_parents_before_children` _function_
-- `test_local_import_not_hoisted_to_bindings` _function_
-- `test_method_calls` _function_
-- `test_module_docstring_first_line` _function_
-- `test_nested_function_is_its_own_symbol_with_own_calls` _function_
-- `test_refs_are_superset_of_calls_capturing_bases_and_annotations` _function_
-- `test_syntax_error_propagates` _function_
-- `test_top_level_symbols` _function_
-
-### `tests.test_regions_emit`
-
-- `test_codetext_regions_cover_the_non_def_source` _function_
-- `test_decompose_attaches_body_and_order_to_top_level_symbols` _function_
-- `test_decompose_emit_round_trip_from_nodes` _function_
-- `test_emit_is_ast_equal_and_seam_canonical_on_messy_input` _function_
-- `test_emit_regions_round_trips_pep8_byte_exact` _function_
-- `test_parse_regions_orders_symbols_and_text` _function_
-- `test_symbol_body_is_verbatim_with_decorators_and_leading_comments` _function_
-
-### `tests_manual.imports_projection`
-
-- `sweep` _function_ — Decompose every arc-lib module and tally false prunes + genuine dead-import prunes.
-
 ## Dependencies
 
-**Depends on:** `cjm-dev-graph-schema`
+**Depends on:** `cjm-context-graph-layer`, `cjm-context-graph-primitives`, `cjm-dev-graph-schema`
 **Used by:** `cjm-context-graph-projection`, `cjm-notebook-decompose-core`

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/README.md

@@ -11,12 +11,6 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - **`cjm_python_decompose_core.extract`** — Bind parsed Python onto dev-graph-schema code nodes (the dev-domain binding).
 - **`cjm_python_decompose_core.ingest`** — Flatten a decomposed Python corpus into graph elements (the queue-free half).
 - **`cjm_python_decompose_core.parse`** — Schema-free Python parsing (stdlib `ast`).
-- **`tests.test_extract_ingest`** — Extract (parse -> schema nodes) + ingest (corpus -> graph elements) binding.
-- **`tests.test_imports_emit`** — Imports-as-projection: deriving a module's canonical import block from bindings.
-- **`tests.test_method_attribution`** — First-param capture + monkey-patch-assignment detection (the cross-cell method idioms).
-- **`tests.test_parse`** — Schema-free Python parsing: docstrings, imports, the symbol tree, and calls.
-- **`tests.test_regions_emit`** — Verbatim region decomposition + canonical emit (the authoring / round-trip substrate).
-- **`tests_manual.imports_projection`** — Imports-as-projection faithfulness sweep over the real arc libs (corpus harness).
 
 ## API
 
@@ -59,72 +53,7 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - `parse_module` _function_ — Parse Python source into module docstring + imports + the symbol tree.
 - `parse_regions` _function_ — Decompose a module into ordered verbatim top-level regions (the round-trip substrate).
 
-### `tests.test_extract_ingest`
-
-- `test_corpus_imports_and_calls_resolve_intra_corpus` _function_
-- `test_decompose_text_builds_module_and_symbols` _function_
-- `test_external_imports_not_minted` _function_
-- `test_import_map_and_call_map` _function_
-- `test_ingest_is_idempotent_in_ids` _function_
-- `test_local_edges_about_repo_and_defines_nesting` _function_
-- `test_module_path_and_import_name` _function_
-- `test_resolve_import_relative_and_absolute` _function_
-
-### `tests.test_imports_emit`
-
-- `test_coexisting_submodule_imports_both_survive_round_trip` _function_ — import urllib.request + import urllib.error both bind `urllib` and both are
-- `test_derived_block_is_faithful_and_prunes_dead_imports` _function_
-- `test_docstring_lines_starting_with_from_survive_derive` _function_
-- `test_emit_module_default_is_verbatim_not_derived` _function_
-- `test_emit_module_derive_imports_prunes_dead_and_keeps_bodies` _function_
-- `test_parameter_default_reference_is_not_pruned` _function_
-- `test_rebinding_imports_still_supersede` _function_ — Aliased / from-imports rebinding a name follow Python last-binding-wins —
-- `test_render_binding_forms` _function_
-- `test_render_import_block_orders_dedups_and_groups` _function_
-- `test_render_import_block_wraps_long_from_lines_in_aligned_parens` _function_
-- `test_same_module_imported_twice_dedupes` _function_ — A literally duplicated plain import collapses to one statement.
-
-### `tests.test_method_attribution`
-
-- `test_first_param_and_annotation_captured` _function_
-- `test_first_param_forward_ref_annotation` _function_
-- `test_first_param_unannotated_and_classes_have_none` _function_
-- `test_monkeypatch_assignments_detected` _function_
-- `test_monkeypatch_is_structural_filtering_is_downstream` _function_
-- `test_no_arg_function_has_empty_first_param` _function_
-
-### `tests.test_parse`
-
-- `parsed` _function_
-- `test_class_methods_are_children_with_method_kind` _function_
-- `test_empty_and_docstringless_module` _function_
-- `test_function_calls_are_direct_only` _function_
-- `test_import_bindings_per_symbol_and_module_level` _function_
-- `test_imports_collected_with_relative_and_submodule_forms` _function_
-- `test_iter_symbols_flattens_parents_before_children` _function_
-- `test_local_import_not_hoisted_to_bindings` _function_
-- `test_method_calls` _function_
-- `test_module_docstring_first_line` _function_
-- `test_nested_function_is_its_own_symbol_with_own_calls` _function_
-- `test_refs_are_superset_of_calls_capturing_bases_and_annotations` _function_
-- `test_syntax_error_propagates` _function_
-- `test_top_level_symbols` _function_
-
-### `tests.test_regions_emit`
-
-- `test_codetext_regions_cover_the_non_def_source` _function_
-- `test_decompose_attaches_body_and_order_to_top_level_symbols` _function_
-- `test_decompose_emit_round_trip_from_nodes` _function_
-- `test_emit_is_ast_equal_and_seam_canonical_on_messy_input` _function_
-- `test_emit_regions_round_trips_pep8_byte_exact` _function_
-- `test_parse_regions_orders_symbols_and_text` _function_
-- `test_symbol_body_is_verbatim_with_decorators_and_leading_comments` _function_
-
-### `tests_manual.imports_projection`
-
-- `sweep` _function_ — Decompose every arc-lib module and tally false prunes + genuine dead-import prunes.
-
 ## Dependencies
 
-**Depends on:** `cjm-dev-graph-schema`
+**Depends on:** `cjm-context-graph-layer`, `cjm-context-graph-primitives`, `cjm-dev-graph-schema`
 **Used by:** `cjm-context-graph-projection`, `cjm-notebook-decompose-core`

cjm_python_decompose_core-0.0.7/cjm_python_decompose_core/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.0.7"

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/cjm_python_decompose_core/emit.py

@@ -241,8 +241,27 @@ def _label_of(node: Any) -> str:
 
 
 def _looks_like_docstring(text: str) -> bool:
-    """Whether a text region opens with a string literal (a module docstring)."""
-    return text.lstrip().startswith(('"""', "'''", '"', "'"))
+    """Whether a text region is the module PREAMBLE the derived import block must follow.
+
+    True when the region opens with a string literal (the module docstring), when comment
+    lines — a shebang, an encoding cookie, a license header — lead into one, or when the
+    region is comments only. AST-located like `_strip_import_lines`: the first STATEMENT
+    decides, so `#!/usr/bin/env python3` above a docstring no longer reads as "not a
+    docstring" (the cascade_manifests.py capture landed the import block at line 1, above
+    the shebang AND the docstring — the shebang stopped working and the docstring became
+    a bare expression; housekeeping sitting 2026-09-23). A region `ast.parse` rejects
+    falls back to the same test with comment lines removed."""
+    try:
+        tree = ast.parse(text)
+    except SyntaxError:
+        stripped = "\n".join(ln for ln in text.splitlines()
+                             if not ln.lstrip().startswith("#")).lstrip()
+        return (not stripped) or stripped.startswith(('"""', "'''", '"', "'"))
+    if not tree.body:
+        return bool(text.strip())  # comments only (a shebang / header with no statements)
+    first = tree.body[0]
+    return (isinstance(first, ast.Expr) and isinstance(first.value, ast.Constant)
+            and isinstance(first.value.value, str))
 
 
 _IMPORT_LINE = re.compile(r"\s*(from|import)\s")

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/cjm_python_decompose_core/extract.py

@@ -77,21 +77,31 @@ def _build_symbols(
     parsed: ParsedModule,     # Its parse
     content_hash: str,        # The module file's content hash (symbols share their file's source)
     path: str,                # The file path (provenance locator)
+    symbol_identity: Any = None,  # Optional identity hook: qualname -> birth fields ({birth_repo_key, birth_module_path, birth_qualname, generation}) or None = born at this address
 ) -> Tuple[List[CodeSymbolNode], List[Dict[str, Any]]]:  # (flat symbol nodes, DEFINES edges)
     """Flatten the parse tree into symbol nodes + structural DEFINES edges.
 
     DEFINES runs module->top-level symbol and parent-symbol->nested-symbol, so the
-    nesting (class -> its methods) is preserved as edges, not just qualname text."""
+    nesting (class -> its methods) is preserved as edges, not just qualname text.
+
+    `symbol_identity` is the container-independent identity hook (36f649d3): asked per
+    qualname, it names the BIRTH address of a symbol that was renamed or re-homed since
+    it was born (a keep-identity op in the source journal), so the node keeps the id it
+    was born with and every journaled edge onto it survives the move; None means the
+    symbol is born at this address (the plain derivation). It runs BEFORE the DEFINES
+    edges are built, because those edges name the ids."""
     symbols: List[CodeSymbolNode] = []
     edges: List[Dict[str, Any]] = []
 
     def make(ps) -> CodeSymbolNode:
+        birth = (symbol_identity(ps.qualname) or {}) if symbol_identity else {}
         node = CodeSymbolNode(
             module_id=module.id, qualname=ps.qualname, symbol_kind=ps.kind,
             path=path, content_hash=content_hash, lineno=ps.lineno,
             docstring=ps.docstring, calls=list(ps.calls), refs=list(ps.refs),
             import_bindings=list(ps.import_bindings),
             properties={"decorators": list(ps.decorators)} if ps.decorators else {},
+            **birth,
         )
         symbols.append(node)
         children = [make(c) for c in ps.children]
@@ -111,6 +121,7 @@ def decompose_text(
     text: str,                        # The module source text
     content_hash: Optional[str] = None,  # Precomputed file hash (else hashed over the UTF-8 text)
     import_name: Optional[str] = None,   # Override the derived dotted import name
+    symbol_identity: Any = None,      # Optional identity hook (qualname -> birth fields; see `_build_symbols`) — container-independent ids (36f649d3)
 ) -> DecomposedModule:  # The decomposed module
     """Parse + bind in one step from in-memory source text."""
     ch = content_hash if content_hash is not None else SourceRef.compute_hash(text.encode("utf-8"))
@@ -121,7 +132,7 @@ def decompose_text(
         docstring=parsed.docstring, imports=list(parsed.imports),
         import_bindings=list(parsed.module_used_bindings),
     )
-    symbols, defines = _build_symbols(module, parsed, ch, path)
+    symbols, defines = _build_symbols(module, parsed, ch, path, symbol_identity=symbol_identity)
 
     # Verbatim-region overlay (the authoring / round-trip substrate): attach each
     # top-level symbol's VERBATIM body + order, mint CodeText nodes for the non-def
@@ -155,12 +166,14 @@ def decompose_file(
     repo_key: str,    # The repo's durable conceptual slug
     path: str,        # Path to the .py file
     repo_root: str,   # Repo root (for the repo-relative module path)
+    symbol_identity_for: Any = None,  # Optional factory: module_path -> the module's identity hook (see `decompose_text`)
 ) -> DecomposedModule:  # The decomposed module
     """Read a `.py` file and decompose it (hash over the raw file bytes)."""
     raw = Path(path).read_bytes()
     mp = module_path_for(path, repo_root)
     return decompose_text(repo_key, mp, str(path), raw.decode("utf-8"),
-                          content_hash=SourceRef.compute_hash(raw))
+                          content_hash=SourceRef.compute_hash(raw),
+                          symbol_identity=symbol_identity_for(mp) if symbol_identity_for else None)
 
 
 def iter_py_files(
@@ -178,12 +191,14 @@ def decompose_paths(
     repo_key: str,           # The repo's durable conceptual slug
     paths: Iterable[str],    # The .py files to decompose
     repo_root: str,          # Repo root (for repo-relative module paths)
+    symbol_identity_for: Any = None,  # Optional factory: module_path -> the module's identity hook (see `decompose_text`)
 ) -> List[DecomposedModule]:  # One DecomposedModule per parseable file
     """Decompose an explicit set of files; unparseable files are skipped (recorded by the caller)."""
     out: List[DecomposedModule] = []
     for path in paths:
         try:
-            out.append(decompose_file(repo_key, path, repo_root))
+            out.append(decompose_file(repo_key, path, repo_root,
+                                      symbol_identity_for=symbol_identity_for))
         except SyntaxError:
             continue
     return out
@@ -193,10 +208,12 @@ def decompose_package(
     repo_key: str,                    # The repo's durable conceptual slug
     package_dir: str,                 # The importable package directory (e.g. ".../cjm_dev_graph_schema")
     repo_root: Optional[str] = None,  # Repo root for relative module paths (default = package_dir's parent)
+    symbol_identity_for: Any = None,  # Optional factory: module_path -> the module's identity hook (see `decompose_text`)
 ) -> List[DecomposedModule]:  # The decomposed package modules
     """Decompose every `.py` under a package dir (the lib's own importable source).
 
     `repo_root` defaults to the package dir's parent, so module paths read like
     `cjm_dev_graph_schema/nodes.py` (the importable form)."""
     root = repo_root if repo_root is not None else str(Path(package_dir).parent)
-    return decompose_paths(repo_key, iter_py_files(package_dir), root)
+    return decompose_paths(repo_key, iter_py_files(package_dir), root,
+                           symbol_identity_for=symbol_identity_for)

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/cjm_python_decompose_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cjm-python-decompose-core
-Version: 0.0.6
+Version: 0.0.7
 Summary: A Python source-code decomposition core for context graphs: parses .py modules (docstrings, imports, and the symbol tree of functions/classes/methods with their calls) into provenance-carrying graph nodes and edges that co-reside with the markdown source-type. First source = the ecosystem's own code; generalizes to any Python corpus.
 Author-email: "Christian J. Mills" <9126128+cj-mills@users.noreply.github.com>
 License: Apache-2.0
@@ -10,7 +10,7 @@ Classifier: Programming Language :: Python :: 3 :: Only
 Requires-Python: >=3.12
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: cjm-dev-graph-schema>=0.0.8
+Requires-Dist: cjm-dev-graph-schema>=0.0.28
 Requires-Dist: cjm-context-graph-layer>=0.0.7
 Requires-Dist: cjm-context-graph-primitives>=0.0.12
 Dynamic: license-file
@@ -28,12 +28,6 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - **`cjm_python_decompose_core.extract`** — Bind parsed Python onto dev-graph-schema code nodes (the dev-domain binding).
 - **`cjm_python_decompose_core.ingest`** — Flatten a decomposed Python corpus into graph elements (the queue-free half).
 - **`cjm_python_decompose_core.parse`** — Schema-free Python parsing (stdlib `ast`).
-- **`tests.test_extract_ingest`** — Extract (parse -> schema nodes) + ingest (corpus -> graph elements) binding.
-- **`tests.test_imports_emit`** — Imports-as-projection: deriving a module's canonical import block from bindings.
-- **`tests.test_method_attribution`** — First-param capture + monkey-patch-assignment detection (the cross-cell method idioms).
-- **`tests.test_parse`** — Schema-free Python parsing: docstrings, imports, the symbol tree, and calls.
-- **`tests.test_regions_emit`** — Verbatim region decomposition + canonical emit (the authoring / round-trip substrate).
-- **`tests_manual.imports_projection`** — Imports-as-projection faithfulness sweep over the real arc libs (corpus harness).
 
 ## API
 
@@ -76,72 +70,7 @@ A Python source-code decomposition core for context graphs: parses .py modules (
 - `parse_module` _function_ — Parse Python source into module docstring + imports + the symbol tree.
 - `parse_regions` _function_ — Decompose a module into ordered verbatim top-level regions (the round-trip substrate).
 
-### `tests.test_extract_ingest`
-
-- `test_corpus_imports_and_calls_resolve_intra_corpus` _function_
-- `test_decompose_text_builds_module_and_symbols` _function_
-- `test_external_imports_not_minted` _function_
-- `test_import_map_and_call_map` _function_
-- `test_ingest_is_idempotent_in_ids` _function_
-- `test_local_edges_about_repo_and_defines_nesting` _function_
-- `test_module_path_and_import_name` _function_
-- `test_resolve_import_relative_and_absolute` _function_
-
-### `tests.test_imports_emit`
-
-- `test_coexisting_submodule_imports_both_survive_round_trip` _function_ — import urllib.request + import urllib.error both bind `urllib` and both are
-- `test_derived_block_is_faithful_and_prunes_dead_imports` _function_
-- `test_docstring_lines_starting_with_from_survive_derive` _function_
-- `test_emit_module_default_is_verbatim_not_derived` _function_
-- `test_emit_module_derive_imports_prunes_dead_and_keeps_bodies` _function_
-- `test_parameter_default_reference_is_not_pruned` _function_
-- `test_rebinding_imports_still_supersede` _function_ — Aliased / from-imports rebinding a name follow Python last-binding-wins —
-- `test_render_binding_forms` _function_
-- `test_render_import_block_orders_dedups_and_groups` _function_
-- `test_render_import_block_wraps_long_from_lines_in_aligned_parens` _function_
-- `test_same_module_imported_twice_dedupes` _function_ — A literally duplicated plain import collapses to one statement.
-
-### `tests.test_method_attribution`
-
-- `test_first_param_and_annotation_captured` _function_
-- `test_first_param_forward_ref_annotation` _function_
-- `test_first_param_unannotated_and_classes_have_none` _function_
-- `test_monkeypatch_assignments_detected` _function_
-- `test_monkeypatch_is_structural_filtering_is_downstream` _function_
-- `test_no_arg_function_has_empty_first_param` _function_
-
-### `tests.test_parse`
-
-- `parsed` _function_
-- `test_class_methods_are_children_with_method_kind` _function_
-- `test_empty_and_docstringless_module` _function_
-- `test_function_calls_are_direct_only` _function_
-- `test_import_bindings_per_symbol_and_module_level` _function_
-- `test_imports_collected_with_relative_and_submodule_forms` _function_
-- `test_iter_symbols_flattens_parents_before_children` _function_
-- `test_local_import_not_hoisted_to_bindings` _function_
-- `test_method_calls` _function_
-- `test_module_docstring_first_line` _function_
-- `test_nested_function_is_its_own_symbol_with_own_calls` _function_
-- `test_refs_are_superset_of_calls_capturing_bases_and_annotations` _function_
-- `test_syntax_error_propagates` _function_
-- `test_top_level_symbols` _function_
-
-### `tests.test_regions_emit`
-
-- `test_codetext_regions_cover_the_non_def_source` _function_
-- `test_decompose_attaches_body_and_order_to_top_level_symbols` _function_
-- `test_decompose_emit_round_trip_from_nodes` _function_
-- `test_emit_is_ast_equal_and_seam_canonical_on_messy_input` _function_
-- `test_emit_regions_round_trips_pep8_byte_exact` _function_
-- `test_parse_regions_orders_symbols_and_text` _function_
-- `test_symbol_body_is_verbatim_with_decorators_and_leading_comments` _function_
-
-### `tests_manual.imports_projection`
-
-- `sweep` _function_ — Decompose every arc-lib module and tally false prunes + genuine dead-import prunes.
-
 ## Dependencies
 
-**Depends on:** `cjm-dev-graph-schema`
+**Depends on:** `cjm-context-graph-layer`, `cjm-context-graph-primitives`, `cjm-dev-graph-schema`
 **Used by:** `cjm-context-graph-projection`, `cjm-notebook-decompose-core`

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/cjm_python_decompose_core.egg-info/requires.txt

@@ -1,3 +1,3 @@
-cjm-dev-graph-schema>=0.0.8
+cjm-dev-graph-schema>=0.0.28
 cjm-context-graph-layer>=0.0.7
 cjm-context-graph-primitives>=0.0.12

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/pyproject.toml

@@ -14,7 +14,7 @@ classifiers = [
     "Programming Language :: Python :: 3",
     "Programming Language :: Python :: 3 :: Only",
 ]
-dependencies = ['cjm-dev-graph-schema>=0.0.8', 'cjm-context-graph-layer>=0.0.7', 'cjm-context-graph-primitives>=0.0.12']
+dependencies = ['cjm-dev-graph-schema>=0.0.28', 'cjm-context-graph-layer>=0.0.7', 'cjm-context-graph-primitives>=0.0.12']
 
 [project.urls]
 Repository = "https://github.com/cj-mills/cjm-python-decompose-core"

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/tests/test_extract_ingest.py

@@ -134,3 +134,36 @@ def test_external_imports_not_minted():
     d = decompose_text("demo-repo", "demo/c.py", "/abs/c.py", src, import_name="demo.c")
     _, edges = corpus_graph_elements([d])
     assert [e for e in edges if e["relation_type"] == DevRelations.IMPORTS] == []
+
+
+def test_symbol_identity_hook_keeps_the_birth_id_and_covers_nested_symbols():
+    # The container-independent identity hook (36f649d3): a class re-homed from a.py keeps
+    # a.py's id in b.py, its methods derive from the same birth address, and a symbol the
+    # hook does not name is simply born here.
+    born = decompose_text("cjm-demo", "demo/a.py", "/tmp/a.py",
+                          "class Thing:\n    def m(self):\n        return 1\n")
+    born_ids = {s.qualname: s.id for s in born.symbols}
+
+    def hook(qualname):
+        if qualname.split(".")[0] == "Thing":
+            return {"birth_repo_key": "cjm-demo", "birth_module_path": "demo/a.py",
+                    "birth_qualname": "Thing" + qualname[len("Thing"):], "generation": 0}
+        return None
+
+    moved = decompose_text("cjm-demo", "demo/b.py", "/tmp/b.py",
+                           "class Thing:\n    def m(self):\n        return 1\n\n\n"
+                           "def fresh():\n    return 2\n",
+                           symbol_identity=hook)
+    ids = {s.qualname: s.id for s in moved.symbols}
+    assert ids["Thing"] == born_ids["Thing"] and ids["Thing.m"] == born_ids["Thing.m"]
+    assert ids["fresh"] not in born_ids.values()
+    assert moved.module.id != born.module.id
+    # The DEFINES edges name the kept ids (the hook ran BEFORE the edges were built).
+    defines = {(e["source_id"], e["target_id"]) for e in moved.local_edges
+               if e["relation_type"] == DevRelations.DEFINES}
+    assert (moved.module.id, ids["Thing"]) in defines and (ids["Thing"], ids["Thing.m"]) in defines
+    # The birth fields ride the wire; a symbol born here carries none.
+    thing = next(s.to_graph_node() for s in moved.symbols if s.qualname == "Thing")
+    assert thing["properties"]["birth_module_path"] == "demo/a.py"
+    fresh = next(s.to_graph_node() for s in moved.symbols if s.qualname == "fresh")
+    assert "birth_module_path" not in fresh["properties"]

{cjm_python_decompose_core-0.0.6 → cjm_python_decompose_core-0.0.7}/tests/test_imports_emit.py

@@ -119,6 +119,37 @@ def test_docstring_lines_starting_with_from_survive_derive():
     compile(emitted, "m.py", "exec")
 
 
+def test_derived_import_block_lands_below_a_shebang_and_docstring():
+    # A shebang (or any comment header) ABOVE the module docstring is still the preamble:
+    # the derived import block must land after it, never at line 1 (regression: the
+    # cascade_manifests.py capture put the imports above `#!/usr/bin/env python3` and the
+    # docstring — the shebang died and the docstring became a bare expression).
+    src = ("#!/usr/bin/env python3\n"
+           '"""A script with a shebang.\n'
+           '"""\n'
+           "\n"
+           "import os\n"
+           "\n\n"
+           "def f():\n"
+           "    return os.getpid()\n")
+    d = decompose_text("pkg", "pkg/m.py", "/tmp/m.py", src)
+    nodes = [s.to_graph_node() for s in d.symbols] + [t.to_graph_node() for t in d.texts]
+    emitted = emit_module_from_nodes(nodes, module_node=d.module.to_graph_node(), derive_imports=True)
+    lines = emitted.splitlines()
+    assert lines[0] == "#!/usr/bin/env python3"
+    assert lines[1] == '"""A script with a shebang.'
+    assert lines.index("import os") > lines.index('"""', 2)
+    compile(emitted, "m.py", "exec")
+    # A comments-only header (no docstring) is a preamble too.
+    src2 = "# -*- coding: utf-8 -*-\n# license header\n\nimport os\n\n\ndef g():\n    return os.sep\n"
+    d2 = decompose_text("pkg", "pkg/n.py", "/tmp/n.py", src2)
+    nodes2 = [s.to_graph_node() for s in d2.symbols] + [t.to_graph_node() for t in d2.texts]
+    emitted2 = emit_module_from_nodes(nodes2, module_node=d2.module.to_graph_node(), derive_imports=True)
+    assert emitted2.splitlines()[0] == "# -*- coding: utf-8 -*-"
+    assert "import os" in emitted2
+    compile(emitted2, "n.py", "exec")
+
+
 def test_coexisting_submodule_imports_both_survive_round_trip():
     """import urllib.request + import urllib.error both bind `urllib` and both are
     live — the emit must carry BOTH (regression: the one-descriptor-per-name table

cjm_python_decompose_core-0.0.6/cjm_python_decompose_core/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.0.6"