entrygraph 0.1.47__tar.gz → 0.1.49__tar.gz

{entrygraph-0.1.47 → entrygraph-0.1.49}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: entrygraph
-Version: 0.1.47
+Version: 0.1.49
 Summary: Language-agnostic code graph: query symbols, entrypoints, and source-to-sink call paths from a SQLite index
 Project-URL: Repository, https://github.com/brettbergin/entrygraph
 Author-email: Brett Bergin <brettberginbc@yahoo.com>

{entrygraph-0.1.47 → entrygraph-0.1.49}/src/entrygraph/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.1.47'
-__version_tuple__ = version_tuple = (0, 1, 47)
+__version__ = version = '0.1.49'
+__version_tuple__ = version_tuple = (0, 1, 49)
 
 __commit_id__ = commit_id = None

{entrygraph-0.1.47 → entrygraph-0.1.49}/src/entrygraph/pipeline/scanner.py

@@ -442,6 +442,15 @@ def _write_symbols(session, extractions, file_id_by_path, alloc, table):
         module_id = alloc.take(Symbol)
         module_ids[path] = module_id
         table.add_module(x.module_path, module_id, x.language)
+        # A module spans its whole file. The extractor doesn't record a line count,
+        # so approximate the last line from the furthest span of any extracted
+        # symbol/import/reference (>= 1 = start_line). This keeps the module row
+        # inside the end_line >= start_line invariant instead of the old end_line=0
+        # (#44), while covering all of the file's content.
+        line_spans = [s.span.end_line for s in x.symbols]
+        line_spans += [r.span.end_line for r in x.references]
+        line_spans += [i.span.end_line for i in x.imports]
+        module_end_line = max(line_spans, default=1)
         symbol_rows.append(
             {
                 "id": module_id,
@@ -451,7 +460,7 @@ def _write_symbols(session, extractions, file_id_by_path, alloc, table):
                 "qname": x.module_path,
                 "parent_id": None,
                 "start_line": 1,
-                "end_line": 0,
+                "end_line": module_end_line,
                 "start_col": 0,
                 "signature": None,
                 "docstring": None,

{entrygraph-0.1.47 → entrygraph-0.1.49}/tests/test_indexer.py

@@ -105,6 +105,20 @@ def test_import_edges(indexed):
         assert "app.services" in imports
 
 
+def test_module_symbols_satisfy_span_invariant(indexed):
+    # Module rows used to be written with end_line=0 < start_line=1 (#44). Every
+    # module symbol must now satisfy end_line >= start_line and span real content.
+    engine, _ = indexed
+    with Session(engine) as s:
+        modules = s.execute(select(Symbol).where(Symbol.kind == SymbolKind.MODULE)).scalars().all()
+        assert modules  # the flask fixture has several modules
+        for m in modules:
+            assert m.start_line == 1
+            assert m.end_line >= m.start_line
+        # at least one module spans past line 1 (files have content)
+        assert any(m.end_line > 1 for m in modules)
+
+
 def test_reindex_is_idempotent(indexed):
     engine, first = indexed
     second = index_repository(FLASK_APP, engine)

{entrygraph-0.1.47 → entrygraph-0.1.49}/tests/test_lang.py

@@ -20,6 +20,17 @@ def test_new_language_extensions():
     assert {"csharp", "php", "rust"} <= EXTRACTABLE
 
 
+def test_not_extracted_group_is_not_in_extractable():
+    # The "recognized but not extracted" extensions must genuinely lack extractors;
+    # cs/rs/php were mis-grouped under that comment despite having extractors (#48).
+    from entrygraph.fs.lang import _EXTENSION_MAP
+
+    not_extracted = {"c", "cpp", "kotlin", "swift", "scala"}
+    assert not (not_extracted & EXTRACTABLE)
+    for ext in (".cs", ".rs", ".php", ".phtml"):
+        assert _EXTENSION_MAP[ext] in EXTRACTABLE
+
+
 def test_filename_detection():
     assert detect_language("Gemfile") == "ruby"
     assert detect_language("subdir/Rakefile") == "ruby"

{entrygraph-0.1.47 → entrygraph-0.1.49}/src/entrygraph/fs/lang.py

@@ -30,18 +30,18 @@ _EXTENSION_MAP = {
     ".rb": "ruby",
     ".rake": "ruby",
     ".gemspec": "ruby",
+    ".cs": "csharp",
+    ".rs": "rust",
+    ".php": "php",
+    ".phtml": "php",
     # recognized but not extracted — kept for honest stats
     ".c": "c",
     ".h": "c",
     ".cpp": "cpp",
     ".cc": "cpp",
     ".hpp": "cpp",
-    ".cs": "csharp",
-    ".rs": "rust",
     ".kt": "kotlin",
     ".swift": "swift",
-    ".php": "php",
-    ".phtml": "php",
     ".scala": "scala",
     ".sh": "shell",
     ".bash": "shell",