agent-codinglanguage-mapper 1.1.2__tar.gz → 1.1.4__tar.gz

{agent_codinglanguage_mapper-1.1.2/src/agent_codinglanguage_mapper.egg-info → agent_codinglanguage_mapper-1.1.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-codinglanguage-mapper
-Version: 1.1.2
+Version: 1.1.4
 Summary: Agent-focused semantic mapping and language services for mixed-language codebases.
 Author: Dark Light
 License-Expression: MPL-2.0
@@ -106,6 +106,12 @@ Parser-native nodes are intentionally private. Ranges are zero-based and half-op
 
 `CodebaseMapper.open()` creates an outline index first. `inspect`, `focus`, and outgoing `callees`/`uses` traces load detail only for the selected source file. Incoming `references`/`callers`/`used_by` traces load the remaining workspace detail because their evidence can originate in any indexed file. Parsed detail and focus remain resident for the mapper session. The same optimized path is used for small and large files.
 
+Batch consumers that only need the complete outline should use
+`snapshot = CodebaseMapper.index(".")`. This one-shot path returns the same deterministic
+`WorkspaceSnapshot` without retaining source buffers, parser outlines, or the interactive refresh
+cache after indexing. It is intended for inventory and documentation pipelines over very large
+repositories.
+
 The private session index retains each parser-native outline tree, its source buffer, a source digest, and a lightweight file fingerprint. A detail upgrade therefore queries the existing tree instead of rereading or reparsing every indexed source. When an indexed source changes during a worker session, the mapper rebuilds the outline snapshot, updates `workspace_revision`, clears stale detail/focus state, and invalidates old cursors before answering the next request. Parser-native nodes never appear in the public result types.
 
 ## CLI

{agent_codinglanguage_mapper-1.1.2 → agent_codinglanguage_mapper-1.1.4}/README.md

@@ -59,6 +59,12 @@ Parser-native nodes are intentionally private. Ranges are zero-based and half-op
 
 `CodebaseMapper.open()` creates an outline index first. `inspect`, `focus`, and outgoing `callees`/`uses` traces load detail only for the selected source file. Incoming `references`/`callers`/`used_by` traces load the remaining workspace detail because their evidence can originate in any indexed file. Parsed detail and focus remain resident for the mapper session. The same optimized path is used for small and large files.
 
+Batch consumers that only need the complete outline should use
+`snapshot = CodebaseMapper.index(".")`. This one-shot path returns the same deterministic
+`WorkspaceSnapshot` without retaining source buffers, parser outlines, or the interactive refresh
+cache after indexing. It is intended for inventory and documentation pipelines over very large
+repositories.
+
 The private session index retains each parser-native outline tree, its source buffer, a source digest, and a lightweight file fingerprint. A detail upgrade therefore queries the existing tree instead of rereading or reparsing every indexed source. When an indexed source changes during a worker session, the mapper rebuilds the outline snapshot, updates `workspace_revision`, clears stale detail/focus state, and invalidates old cursors before answering the next request. Parser-native nodes never appear in the public result types.
 
 ## CLI

{agent_codinglanguage_mapper-1.1.2 → agent_codinglanguage_mapper-1.1.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "agent-codinglanguage-mapper"
-version = "1.1.2"
+version = "1.1.4"
 description = "Agent-focused semantic mapping and language services for mixed-language codebases."
 readme = "README.md"
 requires-python = ">=3.10,<3.15"

agent_codinglanguage_mapper-1.1.4/src/agent_codinglanguage_mapper/_version.py

@@ -0,0 +1 @@
+__version__ = "1.1.4"

{agent_codinglanguage_mapper-1.1.2 → agent_codinglanguage_mapper-1.1.4}/src/agent_codinglanguage_mapper/indexer.py

@@ -54,10 +54,7 @@ class WorkspaceIndexCache:
     directories: tuple[tuple[str, int], ...] = ()
 
     def sources_changed(self, root: Path) -> bool:
-        if any(
-            not _dependencies_unchanged(cached.dependency_stamps)
-            for cached in self.entries.values()
-        ):
+        if any(not _dependencies_unchanged(cached.dependency_stamps) for cached in self.entries.values()):
             return True
         for relative, cached in self.entries.items():
             path = root / relative
@@ -83,6 +80,67 @@ class WorkspaceIndexCache:
         return new_inventory != self.inventory
 
 
+def _deduplicate_parser_symbols(items: tuple[Symbol, ...]) -> tuple[Symbol, ...]:
+    """Collapse identical symbols and C/C++ aliases for one concrete declaration."""
+
+    def declaration_key(item: Symbol) -> tuple[str, str, int, int]:
+        anchor = item.name_range or item.range
+        return (
+            item.qualified_name,
+            anchor.path,
+            anchor.start_line,
+            anchor.start_character,
+        )
+
+    groups: dict[tuple[str, str, int, int], list[Symbol]] = defaultdict(list)
+    for item in items:
+        groups[declaration_key(item)].append(item)
+    emitted: set[tuple[str, str, int, int]] = set()
+    selected: list[Symbol] = []
+    concrete_kinds = {"class", "enum", "struct", "union"}
+    for item in items:
+        key = declaration_key(item)
+        if key in emitted:
+            continue
+        emitted.add(key)
+        group: list[Symbol] = []
+        identities: set[tuple[object, ...]] = set()
+        for candidate in groups[key]:
+            identity = (
+                candidate.name,
+                candidate.qualified_name,
+                candidate.kind,
+                candidate.language,
+                candidate.range,
+                candidate.project_id,
+                candidate.file_id,
+                candidate.signature,
+                candidate.native_kind,
+                candidate.parent_id,
+                candidate.name_range,
+                candidate.role,
+            )
+            if identity in identities:
+                continue
+            identities.add(identity)
+            group.append(candidate)
+        if group and group[0].language is Language.DELPHI:
+            logical: dict[tuple[str, str, str], Symbol] = {}
+            for candidate in group:
+                logical.setdefault(
+                    (candidate.kind, candidate.role, candidate.parent_id),
+                    candidate,
+                )
+            group = list(logical.values())
+        concrete = [candidate for candidate in group if candidate.kind in concrete_kinds]
+        aliases = [candidate for candidate in group if candidate.kind == "type_alias"]
+        if len(concrete) == 1 and aliases:
+            selected.append(concrete[0])
+        else:
+            selected.extend(group)
+    return tuple(selected)
+
+
 def build_workspace_snapshot(
     root: str | Path,
     languages: Iterable[Language] | None = None,
@@ -108,9 +166,7 @@ def build_workspace_snapshot(
     selected_detail_files = None if detail_files is None else {Path(item).as_posix() for item in detail_files}
     incremental_detail = base_snapshot is not None and selected_detail_files is not None
     base_files_by_path = (
-        {item.path: item for item in base_snapshot.files}
-        if incremental_detail and base_snapshot is not None
-        else {}
+        {item.path: item for item in base_snapshot.files} if incremental_detail and base_snapshot is not None else {}
     )
     base_references_by_path: dict[str, list[Reference]] = defaultdict(list)
     base_relations_by_path: dict[str, list[Relation]] = defaultdict(list)
@@ -137,9 +193,8 @@ def build_workspace_snapshot(
     for found in discovery.files:
         relative = found.path.relative_to(workspace).as_posix()
         include_detail = (
-            (selected_detail_files is None or relative in selected_detail_files)
-            and found.path.suffix.casefold() != ".inc"
-        )
+            selected_detail_files is None or relative in selected_detail_files
+        ) and found.path.suffix.casefold() != ".inc"
         file_id = stable_id("file", found.project_id, relative)
         project = projects_by_id[found.project_id]
         delphi_config: DelphiProjectDiscovery | None = None
@@ -271,11 +326,7 @@ def build_workspace_snapshot(
             else:
                 adapter = pooled_adapter
             outline = adapter.parse_outline(relative, source)
-        result = (
-            adapter.resolve(adapter.parse_detail(relative, source, outline))
-            if include_detail
-            else outline
-        )
+        result = adapter.resolve(adapter.parse_detail(relative, source, outline)) if include_detail else outline
         if index_cache is not None:
             index_cache.entries[relative] = _CachedParse(
                 found.project_id,
@@ -292,8 +343,9 @@ def build_workspace_snapshot(
             )
             active_cache_paths.add(relative)
         source_digests.append((relative, source_digest))
+        parsed_symbols = _deduplicate_parser_symbols(result.symbols)
         normalized_symbols_list: list[Symbol] = []
-        for parsed_symbol in result.symbols:
+        for parsed_symbol in parsed_symbols:
             normalized_range = _workspace_required_range(
                 workspace,
                 parsed_symbol.range,
@@ -304,10 +356,7 @@ def build_workspace_snapshot(
                 parsed_symbol.name_range,
                 workspace_path_cache,
             )
-            if (
-                normalized_range is parsed_symbol.range
-                and normalized_name_range is parsed_symbol.name_range
-            ):
+            if normalized_range is parsed_symbol.range and normalized_name_range is parsed_symbol.name_range:
                 normalized_symbols_list.append(parsed_symbol)
             else:
                 normalized_symbols_list.append(
@@ -320,7 +369,7 @@ def build_workspace_snapshot(
         normalized_symbols = tuple(normalized_symbols_list)
         ids = {
             parsed_symbol.target_id: _workspace_symbol_id(found.project_id, file_id, normalized)
-            for parsed_symbol, normalized in zip(result.symbols, normalized_symbols)
+            for parsed_symbol, normalized in zip(parsed_symbols, normalized_symbols)
         }
         symbols.extend(
             Symbol(
@@ -338,7 +387,7 @@ def build_workspace_snapshot(
                 normalized.name_range,
                 parsed_symbol.role,
             )
-            for parsed_symbol, normalized in zip(result.symbols, normalized_symbols)
+            for parsed_symbol, normalized in zip(parsed_symbols, normalized_symbols)
         )
         for reference in result.references:
             source_id = ids.get(reference.source_id, "")
@@ -363,7 +412,7 @@ def build_workspace_snapshot(
                         range=normalized_range,
                     )
                 )
-        local_symbols_by_id = {parsed_symbol.target_id: parsed_symbol for parsed_symbol in result.symbols}
+        local_symbols_by_id = {parsed_symbol.target_id: parsed_symbol for parsed_symbol in parsed_symbols}
         for relation in result.relations:
             relation_library = relation.target_library
             relation_source = local_symbols_by_id.get(relation.source_id)
@@ -382,14 +431,12 @@ def build_workspace_snapshot(
                     target_id=ids.get(relation.target_id, ""),
                     target_library=relation_library,
                     evidence=tuple(
-                        _workspace_required_range(workspace, value, workspace_path_cache)
-                        for value in relation.evidence
+                        _workspace_required_range(workspace, value, workspace_path_cache) for value in relation.evidence
                     ),
                 )
             )
         problems.extend(
-            _workspace_problem(workspace, problem, relative, workspace_path_cache)
-            for problem in result.problems
+            _workspace_problem(workspace, problem, relative, workspace_path_cache) for problem in result.problems
         )
 
     inventory = workspace_inventory(workspace, scan=scan)
@@ -415,11 +462,7 @@ def build_workspace_snapshot(
         relations=resolved_relations,
         problems=tuple(problems),
         source_digests=tuple(source_digests),
-        metadata_digests=tuple(
-            (relative, digest)
-            for relative, _modified, _size, digest in inventory
-            if digest
-        ),
+        metadata_digests=tuple((relative, digest) for relative, _modified, _size, digest in inventory if digest),
     )
 
 
@@ -554,24 +597,16 @@ def _discard_shadowed_fragment_results(
             and owner_path == anchor.path
             and (
                 source_key(symbol) in contextual_keys
-                or (
-                    symbol.language is Language.DELPHI
-                    and symbol.kind == "unit"
-                    and symbol.name_range is None
-                )
+                or (symbol.language is Language.DELPHI and symbol.kind == "unit" and symbol.name_range is None)
             )
         ):
             discarded_ids.add(symbol.target_id)
     kept_symbols = [item for item in symbols if item.target_id not in discarded_ids]
     kept_references = [
-        item
-        for item in references
-        if item.source_id not in discarded_ids and item.target_id not in discarded_ids
+        item for item in references if item.source_id not in discarded_ids and item.target_id not in discarded_ids
     ]
     kept_relations = [
-        item
-        for item in relations
-        if item.source_id not in discarded_ids and item.target_id not in discarded_ids
+        item for item in relations if item.source_id not in discarded_ids and item.target_id not in discarded_ids
     ]
     return kept_symbols, kept_references, kept_relations
 
@@ -624,6 +659,7 @@ def _workspace_symbol_id(project_id: str, file_id: str, symbol: Symbol) -> str:
         "symbol",
         project_id,
         file_id,
+        anchor.path,
         symbol.qualified_name,
         str(anchor.start_line),
         str(anchor.start_character),
@@ -768,21 +804,15 @@ def _resolve_workspace(
             and _is_abi_export_symbol(source)
             and all(
                 item.target_id != source.target_id
-                for item in abi_exports[
-                    (_normalize_library(relation.target_library), relation.target_name)
-                ]
+                for item in abi_exports[(_normalize_library(relation.target_library), relation.target_name)]
             )
         ):
             library = _normalize_library(relation.target_library)
             if library:
                 abi_exports[(library, relation.target_name)].append(source)
 
-    resolved_references = tuple(
-        _resolve_reference(item, by_name, by_id, visibility) for item in references
-    )
-    resolved_relations = tuple(
-        _resolve_relation(item, by_name, abi_exports, by_id, visibility) for item in relations
-    )
+    resolved_references = tuple(_resolve_reference(item, by_name, by_id, visibility) for item in references)
+    resolved_relations = tuple(_resolve_relation(item, by_name, abi_exports, by_id, visibility) for item in relations)
     return resolved_references, resolved_relations
 
 
@@ -924,11 +954,7 @@ def _workspace_visibility(
     visibility: list[_ImportRule] = []
     for reference in references:
         owner = by_id.get(reference.source_id)
-        if (
-            owner is None
-            or reference.reference_kind != "import"
-            or owner.kind not in {"module", "namespace", "unit"}
-        ):
+        if owner is None or reference.reference_kind != "import" or owner.kind not in {"module", "namespace", "unit"}:
             continue
         visibility.append(
             _ImportRule(
@@ -1009,7 +1035,7 @@ def _csharp_namespace(symbol: Symbol, by_id: dict[str, Symbol]) -> str:
         if current.kind == "namespace":
             qualified = current.qualified_name
             module_prefix = f"{PurePosixPath(current.range.path).stem}."
-            return qualified[len(module_prefix):] if qualified.startswith(module_prefix) else qualified
+            return qualified[len(module_prefix) :] if qualified.startswith(module_prefix) else qualified
         current = by_id.get(current.parent_id)
     return ""
 
@@ -1053,7 +1079,7 @@ def _rule_exposes_candidate(
         qualifier_parts = normalized_qualifier.split(".")
         alias_parts = normalized_alias.split(".")
         if qualifier_parts[: len(alias_parts)] == alias_parts and module_matches:
-            remainder = ".".join(qualifier_parts[len(alias_parts):])
+            remainder = ".".join(qualifier_parts[len(alias_parts) :])
             return not remainder or _candidate_parent_scope(candidate, by_id).endswith(remainder)
 
     if not rule.wildcard or not module_matches:
@@ -1065,9 +1091,7 @@ def _rule_exposes_candidate(
 
     candidate_parent = _candidate_parent_scope(candidate, by_id)
     if normalized_qualifier:
-        expected_parent = ".".join(
-            part for part in (rule.import_path, normalized_qualifier) if part
-        )
+        expected_parent = ".".join(part for part in (rule.import_path, normalized_qualifier) if part)
         return candidate_parent == expected_parent
     if rule.module_alias:
         return candidate_parent == rule.import_path

{agent_codinglanguage_mapper-1.1.2 → agent_codinglanguage_mapper-1.1.4}/src/agent_codinglanguage_mapper/mapper.py

@@ -141,7 +141,11 @@ class CodebaseMapper:
 
     @classmethod
     def open(cls, root: str | Path, languages: str | tuple[Language, ...] = "auto") -> "CodebaseMapper":
-        selected = None if languages == "auto" else tuple(Language(item) if isinstance(item, str) else item for item in languages)
+        selected = (
+            None
+            if languages == "auto"
+            else tuple(Language(item) if isinstance(item, str) else item for item in languages)
+        )
         index_cache = WorkspaceIndexCache()
         return cls(
             root,
@@ -151,6 +155,26 @@ class CodebaseMapper:
             index_cache=index_cache,
         )
 
+    @classmethod
+    def index(
+        cls,
+        root: str | Path,
+        languages: str | tuple[Language, ...] = "auto",
+    ) -> WorkspaceSnapshot:
+        """Build a complete outline snapshot without retaining an interactive cache."""
+
+        selected = (
+            None
+            if languages == "auto"
+            else tuple(Language(item) if isinstance(item, str) else item for item in languages)
+        )
+        return build_workspace_snapshot(
+            root,
+            selected,
+            detail_files=(),
+            index_cache=None,
+        )
+
     def __enter__(self) -> "CodebaseMapper":
         return self
 
@@ -189,10 +213,7 @@ class CodebaseMapper:
                 if (
                     request.file_id
                     and candidate.file_id != request.file_id
-                    and (
-                        requested_source is None
-                        or Path(candidate.range.path).as_posix() != requested_source.path
-                    )
+                    and (requested_source is None or Path(candidate.range.path).as_posix() != requested_source.path)
                 ):
                     continue
                 if request.languages and candidate.language not in request.languages:
@@ -273,11 +294,7 @@ class CodebaseMapper:
         )
 
     def _refresh_changed_sources(self) -> None:
-        if (
-            not self._lazy_details
-            or self._index_cache is None
-            or not self._index_cache.sources_changed(self.root)
-        ):
+        if not self._lazy_details or self._index_cache is None or not self._index_cache.sources_changed(self.root):
             return
         self.snapshot = build_workspace_snapshot(
             self.root,
@@ -351,7 +368,7 @@ class CodebaseMapper:
             text_offset = self._decode_cursor(request)
             if text_offset > len(full_text) or (request.cursor and text_offset == len(full_text)):
                 raise ProtocolError("cursor offset is invalid")
-            item[text_key] = full_text[text_offset:text_offset + _SOURCE_CHUNK_CHARS]
+            item[text_key] = full_text[text_offset : text_offset + _SOURCE_CHUNK_CHARS]
             if text_key == "source":
                 line_offset = full_text[:text_offset].count("\n")
                 item["citation"] = f"{symbol.range.path}:{int(str(item['citation']).rsplit(':', 1)[1]) + line_offset}"
@@ -450,7 +467,7 @@ class CodebaseMapper:
                     low = middle + 1
                 else:
                     high = middle - 1
-            bounded[detail_key] = text[:max(0, high)]
+            bounded[detail_key] = text[: max(0, high)]
             return bounded
 
         for key in ("kind", "role", "project_id", "file_id", "qualified_name", "signature"):
@@ -487,9 +504,7 @@ class CodebaseMapper:
             selected = [item for item in references if item.target_id in request_target_id_set]
             return self._page_response(request, [_reference_mapping(item) for item in selected])
         if relation in {"callers", "used_by"}:
-            target_ids = {
-                item.source_id for item in references if item.target_id in request_target_id_set
-            }
+            target_ids = {item.source_id for item in references if item.target_id in request_target_id_set}
             target_ids.update(
                 item.source_id
                 for item in self.snapshot.relations
@@ -501,16 +516,12 @@ class CodebaseMapper:
             return self._page_response(request, [_symbol_mapping(item) for item in caller_symbols])
         if relation in {"callees", "uses"}:
             target_ids = {
-                item.target_id
-                for item in references
-                if item.source_id in request_target_id_set and item.target_id
+                item.target_id for item in references if item.source_id in request_target_id_set and item.target_id
             }
             target_ids.update(
                 item.target_id
                 for item in self.snapshot.relations
-                if relation == "uses"
-                and item.source_id in request_target_id_set
-                and item.target_id
+                if relation == "uses" and item.source_id in request_target_id_set and item.target_id
             )
             callee_symbols = [
                 self.snapshot.symbols_by_id[item] for item in target_ids if item in self.snapshot.symbols_by_id
@@ -545,7 +556,7 @@ class CodebaseMapper:
         start = self._decode_cursor(request)
         selected: list[dict[str, object]] = []
         used = 0
-        for item in items[start:start + request.max_items]:
+        for item in items[start : start + request.max_items]:
             size = len(json.dumps(item, sort_keys=True, separators=(",", ":")))
             if used + size > request.max_chars:
                 if not selected:

{agent_codinglanguage_mapper-1.1.2 → agent_codinglanguage_mapper-1.1.4/src/agent_codinglanguage_mapper.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-codinglanguage-mapper
-Version: 1.1.2
+Version: 1.1.4
 Summary: Agent-focused semantic mapping and language services for mixed-language codebases.
 Author: Dark Light
 License-Expression: MPL-2.0
@@ -106,6 +106,12 @@ Parser-native nodes are intentionally private. Ranges are zero-based and half-op
 
 `CodebaseMapper.open()` creates an outline index first. `inspect`, `focus`, and outgoing `callees`/`uses` traces load detail only for the selected source file. Incoming `references`/`callers`/`used_by` traces load the remaining workspace detail because their evidence can originate in any indexed file. Parsed detail and focus remain resident for the mapper session. The same optimized path is used for small and large files.
 
+Batch consumers that only need the complete outline should use
+`snapshot = CodebaseMapper.index(".")`. This one-shot path returns the same deterministic
+`WorkspaceSnapshot` without retaining source buffers, parser outlines, or the interactive refresh
+cache after indexing. It is intended for inventory and documentation pipelines over very large
+repositories.
+
 The private session index retains each parser-native outline tree, its source buffer, a source digest, and a lightweight file fingerprint. A detail upgrade therefore queries the existing tree instead of rereading or reparsing every indexed source. When an indexed source changes during a worker session, the mapper rebuilds the outline snapshot, updates `workspace_revision`, clears stale detail/focus state, and invalidates old cursors before answering the next request. Parser-native nodes never appear in the public result types.
 
 ## CLI