codebase-cortex 0.1.0__tar.gz → 0.1.2__tar.gz

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codebase-cortex
-Version: 0.1.0
+Version: 0.1.2
 Summary: AI-powered documentation autopilot — commit code, docs update themselves. Five LangGraph agents analyze diffs, find related code via FAISS embeddings, and sync Notion pages through MCP.
 Project-URL: Homepage, https://github.com/sarupurisailalith/codebase-cortex
 Project-URL: Repository, https://github.com/sarupurisailalith/codebase-cortex
@@ -164,6 +164,22 @@ graph TD
 4. **TaskCreator** — Identifies undocumented areas and creates task pages in Notion
 5. **SprintReporter** — Synthesizes all activity into a weekly sprint summary
 
+## Notion Page Structure
+
+When you run `cortex init`, Cortex creates a parent page in Notion **named after your repository directory**. All documentation pages are created as children of this parent:
+
+```
+your-project/          (repo directory)
+  └── Notion:
+      📄 your-project              ← parent page (named after repo)
+        ├── 🏗️ Architecture Overview
+        ├── 📡 API Reference
+        ├── 📋 Sprint Log
+        └── ✅ Task Board
+```
+
+Each repo gets its own parent page — if you use Cortex in multiple projects, they each get an independent page tree. To bring existing Notion pages under Cortex management, simply move them under the parent page in Notion and run `cortex scan` to discover them.
+
 ## Architecture
 
 For detailed architecture documentation, see [`docs/architecture.md`](docs/architecture.md).

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/README.md

@@ -118,6 +118,22 @@ graph TD
 4. **TaskCreator** — Identifies undocumented areas and creates task pages in Notion
 5. **SprintReporter** — Synthesizes all activity into a weekly sprint summary
 
+## Notion Page Structure
+
+When you run `cortex init`, Cortex creates a parent page in Notion **named after your repository directory**. All documentation pages are created as children of this parent:
+
+```
+your-project/          (repo directory)
+  └── Notion:
+      📄 your-project              ← parent page (named after repo)
+        ├── 🏗️ Architecture Overview
+        ├── 📡 API Reference
+        ├── 📋 Sprint Log
+        └── ✅ Task Board
+```
+
+Each repo gets its own parent page — if you use Cortex in multiple projects, they each get an independent page tree. To bring existing Notion pages under Cortex management, simply move them under the parent page in Notion and run `cortex scan` to discover them.
+
 ## Architecture
 
 For detailed architecture documentation, see [`docs/architecture.md`](docs/architecture.md).

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "codebase-cortex"
-version = "0.1.0"
+version = "0.1.2"
 description = "AI-powered documentation autopilot — commit code, docs update themselves. Five LangGraph agents analyze diffs, find related code via FAISS embeddings, and sync Notion pages through MCP."
 readme = "README.md"
 requires-python = ">=3.11"

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/agents/doc_writer.py

@@ -150,7 +150,7 @@ class DocWriterAgent(BaseAgent):
                 existing_content_section += f"\n### {title}\n```\n{truncated}\n```\n"
 
         # Build dynamic page list from cache
-        doc_pages = cache.find_all_doc_pages()
+        doc_pages = cache.find_all_doc_pages(parent_title=settings.repo_path.name)
         page_list = "\n".join(f"- {p.title}" for p in doc_pages) if doc_pages else "- (no pages yet)"
 
         # Ask LLM to generate doc updates
@@ -240,8 +240,8 @@ Only include pages that genuinely need updating. Respond with ONLY the JSON arra
         logger = get_logger()
         existing: dict[str, str] = {}
 
-        # Fetch all doc pages (skip infrastructure-only pages)
-        doc_pages = cache.find_all_doc_pages()
+        # Fetch all doc pages (skip parent page)
+        doc_pages = cache.find_all_doc_pages(parent_title=settings.repo_path.name)
         # Limit to 10 pages to avoid excessive API calls
         pages_to_fetch = doc_pages[:10]
 
@@ -300,7 +300,8 @@ Only include pages that genuinely need updating. Respond with ONLY the JSON arra
         settings = Settings.from_env()
 
         # Get parent page for new pages
-        parent_page = cache.find_by_title("Codebase Cortex")
+        parent_title = settings.repo_path.name
+        parent_page = cache.find_by_title(parent_title)
         parent_id = parent_page.page_id if parent_page else None
 
         try:

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/agents/sprint_reporter.py

@@ -89,7 +89,7 @@ Write a complete sprint report in markdown."""
         cache = PageCache(cache_path=settings.page_cache_path)
 
         sprint_page = cache.find_by_title("Sprint Log")
-        parent_page = cache.find_by_title("Codebase Cortex")
+        parent_page = cache.find_by_title(settings.repo_path.name)
         parent_id = (sprint_page or parent_page)
         parent_id = parent_id.page_id if parent_id else None
 
@@ -101,17 +101,16 @@ Write a complete sprint report in markdown."""
                 content = f"# Sprint Report — Week of {week_label}\n\n{summary}"
 
                 if sprint_page:
-                    # Append to existing Sprint Log using insert_content_after
+                    # Replace Sprint Log content with latest report
                     await session.call_tool(
                         "notion-update-page",
                         arguments={
                             "page_id": sprint_page.page_id,
-                            "command": "insert_content_after",
-                            "selection_with_ellipsis": "---\n*Auto-gen...by Codebase Cortex*",
-                            "new_str": f"\n\n---\n\n{content}",
+                            "command": "replace_content",
+                            "new_str": content,
                         },
                     )
-                    logger.info(f"Appended to Sprint Log for week of {week_label}")
+                    logger.info(f"Updated Sprint Log for week of {week_label}")
                 else:
                     # Create new sprint report page
                     create_args: dict = {

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/agents/task_creator.py

@@ -91,7 +91,7 @@ Respond with a JSON array of tasks (title, description, priority). Return [] if
         cache = PageCache(cache_path=settings.page_cache_path)
 
         task_board = cache.find_by_title("Task Board")
-        parent_page = cache.find_by_title("Codebase Cortex")
+        parent_page = cache.find_by_title(settings.repo_path.name)
         parent_id = (task_board or parent_page)
         parent_id = parent_id.page_id if parent_id else None
 

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/cli.py

@@ -314,7 +314,7 @@ def run(once: bool, watch: bool, dry_run: bool, full: bool, verbose: bool) -> No
         if arch_page and arch_page.content_hash == "":
             # Pages exist but were never written with real content
             # Check if this looks like a first run after init
-            doc_pages = cache.find_all_doc_pages()
+            doc_pages = cache.find_all_doc_pages(parent_title=settings.repo_path.name)
             all_empty = all(p.content_hash == "" for p in doc_pages)
             if all_empty:
                 console.print("[cyan]First run detected — doing full codebase scan[/cyan]")
@@ -624,7 +624,7 @@ async def _run_prompt(
     from codebase_cortex.utils.section_parser import merge_sections, parse_sections
 
     cache = PageCache(cache_path=settings.page_cache_path)
-    doc_pages = cache.find_all_doc_pages()
+    doc_pages = cache.find_all_doc_pages(parent_title=settings.repo_path.name)
 
     if not doc_pages:
         console.print("[red]No pages in cache. Run 'cortex run --once' first.[/red]")

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/notion/bootstrap.py

@@ -10,7 +10,9 @@ from codebase_cortex.utils.logging import get_logger
 
 logger = get_logger()
 
-PARENT_PAGE_TITLE = "Codebase Cortex"
+def get_parent_page_title(settings: Settings) -> str:
+    """Return the parent page title for a given repo (the repo directory name)."""
+    return settings.repo_path.name
 
 
 def normalize_page_id(raw_id: str) -> str:
@@ -119,7 +121,8 @@ async def discover_child_pages(settings: Settings) -> int:
 
     logger = get_logger()
     cache = PageCache(cache_path=settings.page_cache_path)
-    parent_page = cache.find_by_title("Codebase Cortex")
+    parent_title = get_parent_page_title(settings)
+    parent_page = cache.find_by_title(parent_title)
     if not parent_page:
         return 0
 
@@ -187,7 +190,7 @@ async def discover_child_pages(settings: Settings) -> int:
 async def bootstrap_notion_pages(settings: Settings) -> list[dict]:
     """Create the starter Notion pages via MCP tools.
 
-    Creates a parent "Codebase Cortex" page, then child pages under it.
+    Creates a parent page named after the repo directory, then child pages under it.
     Searches for existing pages first to avoid duplicates.
     Seeds the page cache with all created/found pages.
 
@@ -209,7 +212,7 @@ async def bootstrap_notion_pages(settings: Settings) -> list[dict]:
 
     async with notion_mcp_session(settings) as session:
         # Step 1: Search for existing parent page
-        parent_id = await search_page_by_title(session, PARENT_PAGE_TITLE)
+        parent_id = await search_page_by_title(session, parent_title)
 
         # Step 2: Create parent page if not found
         if not parent_id:
@@ -232,7 +235,7 @@ async def bootstrap_notion_pages(settings: Settings) -> list[dict]:
                 )
                 parent_id = extract_page_id(result)
                 if parent_id:
-                    cache.upsert(parent_id, PARENT_PAGE_TITLE)
+                    cache.upsert(parent_id, parent_title)
                     logger.info(f"Created parent page: {parent_title}")
                 else:
                     logger.error("Failed to extract parent page ID from response")
@@ -241,28 +244,24 @@ async def bootstrap_notion_pages(settings: Settings) -> list[dict]:
                 logger.error(f"Failed to create parent page: {e}")
                 return []
         else:
-            cache.upsert(parent_id, PARENT_PAGE_TITLE)
-            logger.info(f"Found existing parent page: {PARENT_PAGE_TITLE}")
+            cache.upsert(parent_id, parent_title)
+            logger.info(f"Found existing parent page: {parent_title}")
 
         # Step 3: Create child pages under parent
         for page_info in STARTER_PAGES:
             title = page_info["title"]
             display_title = f"{page_info['icon']} {title}"
 
-            # Check cache first, then search Notion
+            # Only check cache — don't search the workspace, as that could
+            # find pages under a different parent from a previous init.
+            # Users adopt existing pages by moving them under the parent
+            # and running `cortex scan`.
             cached = cache.find_by_title(title)
             if cached:
                 pages.append({"title": title, "page_id": cached.page_id})
                 logger.info(f"Already exists (cached): {display_title}")
                 continue
 
-            existing_id = await search_page_by_title(session, title)
-            if existing_id:
-                cache.upsert(existing_id, title)
-                pages.append({"title": title, "page_id": existing_id})
-                logger.info(f"Found existing: {display_title}")
-                continue
-
             # Create new page under parent
             await rate_limiter.acquire()
             try:

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/src/codebase_cortex/notion/page_cache.py

@@ -99,9 +99,9 @@ class PageCache:
                 return page
         return None
 
-    def find_all_doc_pages(self) -> list[CachedPage]:
-        """Return all cached pages except infrastructure pages."""
+    def find_all_doc_pages(self, parent_title: str | None = None) -> list[CachedPage]:
+        """Return all cached pages except the parent page."""
         return [
             p for p in self.pages.values()
-            if p.title != "Codebase Cortex"
+            if parent_title is None or p.title != parent_title
         ]

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/tests/test_bootstrap.py

@@ -41,6 +41,7 @@ async def test_discover_child_pages_no_parent(tmp_path):
 
     settings = MagicMock()
     settings.page_cache_path = cache_path
+    settings.repo_path.name = "my-repo"
 
     count = await discover_child_pages(settings)
     assert count == 0
@@ -53,11 +54,12 @@ async def test_discover_child_pages_finds_new(tmp_path):
 
     cache_path = tmp_path / "cache.json"
     cache = PageCache(cache_path=cache_path)
-    cache.upsert("parent-id-0000-0000-000000000001", "Codebase Cortex")
+    cache.upsert("parent-id-0000-0000-000000000001", "my-repo")
 
     settings = MagicMock()
     settings.page_cache_path = cache_path
     settings.notion_token_path = tmp_path / "tokens.json"
+    settings.repo_path.name = "my-repo"
 
     # Mock the parent page response with a child page UUID in content
     child_id = "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee"
@@ -114,7 +116,7 @@ async def test_discover_child_pages_skips_cached(tmp_path):
 
     cache_path = tmp_path / "cache.json"
     cache = PageCache(cache_path=cache_path)
-    cache.upsert("parent-id-0000-0000-000000000001", "Codebase Cortex")
+    cache.upsert("parent-id-0000-0000-000000000001", "my-repo")
 
     child_id = "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee"
     cache.upsert(child_id, "Already Cached Page")
@@ -122,6 +124,7 @@ async def test_discover_child_pages_skips_cached(tmp_path):
     settings = MagicMock()
     settings.page_cache_path = cache_path
     settings.notion_token_path = tmp_path / "tokens.json"
+    settings.repo_path.name = "my-repo"
 
     parent_response = MagicMock()
     parent_response.isError = False

{codebase_cortex-0.1.0 → codebase_cortex-0.1.2}/tests/test_page_cache.py

@@ -36,13 +36,13 @@ def test_find_by_title_fuzzy_case(tmp_path: Path):
 
 def test_find_all_doc_pages(tmp_path: Path):
     cache = PageCache(cache_path=tmp_path / "cache.json")
-    cache.upsert("parent", "Codebase Cortex")
+    cache.upsert("parent", "my-repo")
     cache.upsert("id1", "Architecture Overview")
     cache.upsert("id2", "API Reference")
 
-    doc_pages = cache.find_all_doc_pages()
+    doc_pages = cache.find_all_doc_pages(parent_title="my-repo")
     titles = [p.title for p in doc_pages]
-    assert "Codebase Cortex" not in titles
+    assert "my-repo" not in titles
     assert "Architecture Overview" in titles
     assert "API Reference" in titles