fix(backend/sdoc_source_code): Fix source node parsing on Windows

For config entries like 'source_nodes = [{ "src/" = ... }]' on Windows,
StrictDoc did not find matches since it compared slash path-strings with
backslash path-strings. This effectively disabled the Source Nodes
feature.

We can use pathlib.Path, which automatically creates a WindowsPath or a
PosixPath and supports matching amongst both.

The fix comes with minor refactoring, most notably representing the
config entries as dataclass.

Author: haxtibal

strictdoc/core/file_traceability_index.py

@@ -43,7 +43,7 @@
 from strictdoc.backend.sdoc_source_code.models.source_node import SourceNode
 from strictdoc.core.constants import GraphLinkType
 from strictdoc.core.document_iterator import SDocDocumentIterator
-from strictdoc.core.project_config import ProjectConfig
+from strictdoc.core.project_config import ProjectConfig, SourceNodesEntry
 from strictdoc.core.source_tree import SourceFile
 from strictdoc.helpers.cast import assert_cast
 from strictdoc.helpers.exception import StrictDocException
@@ -596,9 +596,11 @@ def create_folder_section(
         documents_with_generated_content = set()
 
         section_cache = {}
-        source_nodes_config: List[Dict[str, str]] = project_config.source_nodes
+        source_nodes_config: List[SourceNodesEntry] = (
+            project_config.source_nodes
+        )
         unused_source_node_paths = {
-            config_entry_["path"] for config_entry_ in source_nodes_config
+            config_entry_.path for config_entry_ in source_nodes_config
         }
         for (
             path_to_source_file_,
@@ -610,16 +612,19 @@ def create_folder_section(
             if len(source_nodes_config) == 0:
                 continue
 
-            for config_entry_ in source_nodes_config:
-                config_entry_path = config_entry_["path"]
-                if path_to_source_file_.startswith(config_entry_path):
-                    relevant_source_node_entry = config_entry_
-                    unused_source_node_paths.discard(config_entry_path)
-                    break
+            relevant_source_node_entry = (
+                project_config.get_relevant_source_nodes_entry(
+                    path_to_source_file_
+                )
+            )
+            if relevant_source_node_entry is not None:
+                unused_source_node_paths.discard(
+                    relevant_source_node_entry.path
+                )
             else:
                 continue
 
-            document_uid = relevant_source_node_entry["uid"]
+            document_uid = relevant_source_node_entry.uid
             document = traceability_index.get_node_by_uid(document_uid)
             documents_with_generated_content.add(document)
 
@@ -651,7 +656,10 @@ def create_folder_section(
                 if source_sdoc_node is not None:
                     source_sdoc_node = assert_cast(source_sdoc_node, SDocNode)
                     self.merge_sdoc_node_with_source_node(
-                        source_sdoc_node, source_node_, document, config_entry_
+                        source_sdoc_node,
+                        source_node_,
+                        document,
+                        relevant_source_node_entry,
                     )
                 else:
                     source_sdoc_node = self.create_sdoc_node_from_source_node(
@@ -969,11 +977,11 @@ def create_sdoc_node_from_source_node(
         source_node: SourceNode,
         uid: str,
         parent_document: SDocDocumentIF,
-        relevant_source_node_entry: dict[str, str],
+        relevant_source_node_entry: SourceNodesEntry,
     ) -> SDocNode:
         source_sdoc_node = SDocNode(
             parent=parent_document,
-            node_type=relevant_source_node_entry["node_type"],
+            node_type=relevant_source_node_entry.node_type,
             fields=[],
             relations=[],
             # It is important that this autogenerated node is marked as such.
@@ -1005,10 +1013,10 @@ def merge_sdoc_node_with_source_node(
         sdoc_node: SDocNode,
         source_node: SourceNode,
         parent_document: SDocDocumentIF,
-        source_node_config_entry: dict[str, str],
+        source_node_config_entry: SourceNodesEntry,
     ) -> None:
         # First check if grammar element definitions are compatible.
-        source_node_type = source_node_config_entry["node_type"]
+        source_node_type = source_node_config_entry.node_type
         source_node_grammar = assert_cast(
             parent_document.grammar, DocumentGrammar
         )

strictdoc/core/project_config.py

@@ -8,7 +8,9 @@
 import sys
 import tempfile
 import types
+from dataclasses import dataclass
 from enum import Enum
+from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple
 
 import toml
@@ -40,6 +42,19 @@ def parse_relation_tuple(column_name: str) -> Optional[Tuple[str, str]]:
     return match_result.group(1), match_result.group(4)
 
 
+@dataclass
+class SourceNodesEntry:
+    path: str
+    uid: str
+    node_type: str
+    full_path: Optional[Path]
+
+    @classmethod
+    def from_cfg_data(cls, data: dict[str, str]) -> "SourceNodesEntry":
+        full_path = Path(data["full_path"]) if "full_path" in data else None
+        return cls(data["path"], data["uid"], data["node_type"], full_path)
+
+
 class ProjectFeature(str, Enum):
     # Stable features.
     TABLE_SCREEN = "TABLE_SCREEN"
@@ -193,8 +208,13 @@ def __init__(
         self.test_report_root_dict: Dict[str, str] = (
             test_report_root_dict if test_report_root_dict is not None else {}
         )
-        self.source_nodes: List[Dict[str, str]] = (
-            source_nodes if source_nodes is not None else []
+        self.source_nodes: List[SourceNodesEntry] = (
+            [
+                SourceNodesEntry.from_cfg_data(source_node)
+                for source_node in source_nodes
+            ]
+            if source_nodes is not None
+            else []
         )
 
         # Settings derived from the command-line parameters.
@@ -443,23 +463,39 @@ def get_project_hash(self) -> str:
         assert self.input_paths is not None and len(self.input_paths) > 0
         return get_md5(self.input_paths[0])
 
-    def parse_nodes_type(self, path_to_file: str) -> Optional[tuple[str, str]]:
+    def get_relevant_source_nodes_entry(
+        self, path_to_file: str
+    ) -> Optional[SourceNodesEntry]:
+        """
+        Get relevant source_nodes config item for a given source code file.
+
+        Returns data for the first entry from source_nodes that is a parent path of path_to_file.
+        If path_to_file is absolute, source node config entries are assumed to be in the source_root_path.
+        """
         if self.source_root_path is None:
             return None
 
+        source_file_path = Path(path_to_file)
         for sdoc_source_config_entry_ in self.source_nodes:
             # FIXME: Move the setting of full paths to .finalize() of this config
             #        class when it is implemented.
-            full_path = sdoc_source_config_entry_.setdefault(
-                "full_path",
-                os.path.join(
-                    self.source_root_path, sdoc_source_config_entry_["path"]
-                ),
-            )
-            if path_to_file.startswith(full_path):
-                return sdoc_source_config_entry_[
-                    "uid"
-                ], sdoc_source_config_entry_["node_type"]
+            if sdoc_source_config_entry_.full_path is None:
+                sdoc_source_config_entry_.full_path = Path(
+                    self.source_root_path
+                ) / Path(sdoc_source_config_entry_.path)
+
+            if source_file_path.is_absolute():
+                if (
+                    sdoc_source_config_entry_.full_path
+                    in source_file_path.parents
+                ):
+                    return sdoc_source_config_entry_
+            else:
+                if (
+                    Path(sdoc_source_config_entry_.path)
+                    in source_file_path.parents
+                ):
+                    return sdoc_source_config_entry_
 
         return None
 

strictdoc/core/traceability_index_builder.py

@@ -885,13 +885,18 @@ def source_node_grammar_element(
         project_config: ProjectConfig,
         traceability_index: TraceabilityIndex,
     ) -> Optional[GrammarElement]:
-        maybe_parse_nodes_type = project_config.parse_nodes_type(path_to_file)
-        if maybe_parse_nodes_type is None:
+        source_nodes_cfg_entry = project_config.get_relevant_source_nodes_entry(
+            path_to_file
+        )
+        if source_nodes_cfg_entry is None:
             return None
-        parse_nodes_uid, parse_nodes_type = maybe_parse_nodes_type
         sdoc_document = assert_cast(
-            traceability_index.get_node_by_uid_weak2(parse_nodes_uid),
+            traceability_index.get_node_by_uid_weak2(
+                source_nodes_cfg_entry.uid
+            ),
             SDocDocument,
         )
         assert sdoc_document.grammar is not None
-        return sdoc_document.grammar.elements_by_type.get(parse_nodes_type)
+        return sdoc_document.grammar.elements_by_type.get(
+            source_nodes_cfg_entry.node_type
+        )