feat: support nested objects

[Chris1221]

Summary

Adds support for arbitrarily nested YAML objects in yamldoc documentation output.

What changed

• Parser (parser.py): Replaced flat single-level sub-entry handling with a context stack (sub_stack) that routes new entries to the correct parent MetaEntry at any indentation depth.
• Entries (entries.py):
  • MetaEntry now has a type attribute (for schema-derived type info) and a to_table_row() method for inline rendering within a parent table.
  • MetaEntry.to_markdown() takes a depth parameter and uses depth-aware heading levels (##/### at depth 1, ####/##### at depth 2), then recursively renders nested MetaEntries beneath the table.
• Schema resolution (add_type_metadata): Recursive tree search replaces the previous 2-level-only approach, so schema types are applied correctly regardless of nesting depth.
• Tests: Added fixture files (deeper_nesting.yaml, deeper_nesting.schema, deeper_nesting_exclusion.yaml) and full test coverage for 3-level nesting (parse, schema, E2E, markdown output, exclusion).
• README: Removed stale CircleCI badge; updated codecov badge to main branch.

Rendered output for 3-level nesting

Given:

#' This is a three level thing.
three:
    #' This is the second level.
    level_two:
        #' This is the third level.
        level_three: "hello"

Produces:

## `three`

This is a three level thing.

### Member variables:

| Key | Value | Type | Information |
| :-: | :-: | :-: | :-- |
| `level_two` |  | object | This is the second level. |

#### `level_two`

This is the second level.

##### Member variables:

| Key | Value | Type | Information |
| :-: | :-: | :-: | :-- |
| `level_three` | `"hello"` | string | This is the third level. |

Test plan

• All 44 existing tests pass
• TestYAMLs::test_deeper_nesting — parses 3-level structure correctly
• TestSchemas::test_deeper_nesting — schema types applied at depth 3
• TestE2E::test_deeper_nesting — all keys/values present in output
• TestMarkdown::test_deeper_nesting — exact markdown output matches expected
• test_deeper_nesting_exclusion — nested exclusion works correctly

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 97.29730% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 91.61%. Comparing base (f7d9112) to head (a0688b6).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
yamldoc/parser.py	92.68%	3 Missing ⚠️
yamldoc/entries.py	96.55%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #18      +/-   ##
==========================================
+ Coverage   90.52%   91.61%   +1.08%     
==========================================
  Files           7        7              
  Lines         665      787     +122     
==========================================
+ Hits          602      721     +119     
- Misses         63       66       +3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Chris1221]

@claude Please review

[claude]

Claude finished @Chris1221's task in 3m 4s —— View job

---

Review complete

• Gather context (PR diff, changed files, CLAUDE.md)
• Understand the request (code review of nested objects support)
• Review parser.py changes (sub_stack context)
• Review entries.py changes (MetaEntry depth, to_table_row)
• Review test coverage / fixtures
• Check for stray committed artifacts (egg-info)
• Post final review

Note: I could not execute pytest in this sandbox (python invocation was blocked by tooling permissions), so this review is based on static reading of the diff (git diff f7d9112..HEAD) and manual trace-through of the parsing/rendering logic, not a live test run.

The overall approach is solid: replacing the old 2-level-only sub-entry handling with a generic sub_stack in yamldoc/parser.py is a clean way to support arbitrary nesting depth, and the recursive _find_meta_entry search in add_type_metadata is a good simplification over the old flat lookup. Test coverage for the 3-level case (parse/schema/E2E/markdown/exclusion) is thorough.

Two things worth addressing:

1. Markdown heading levels overflow past 2 levels of nested MetaEntrys (yamldoc/entries.py:138-139)

h_section = "#" * (depth * 2)
h_members = "#" * (depth * 2 + 1)

At depth=3 (i.e. a 4th level of key nesting: top -> obj1 -> obj2 -> obj3), h_members becomes 7 # characters, which exceeds CommonMark's max heading level (h6) and will render as plain text instead of a heading. The PR title/commit message ("support arbitrary nesting depth in objects") implies unlimited depth, but only 2 levels of nested MetaEntrys render correctly — this is also called out honestly in the CLAUDE.md diff, but there's no test exercising a 4th level, so this regression-in-waiting isn't caught. Worth either capping heading levels (e.g. min(depth * 2 + 1, 6)) and falling back to bold text beyond that, or adding a test that documents/asserts the current 3-level ceiling if that's an accepted limitation for now.

Fix this →

2. yamldoc.egg-info/ build artifacts committed (introduced in 72ce0f3)

yamldoc.egg-info/PKG-INFO, SOURCES.txt, dependency_links.txt, entry_points.txt, top_level.txt are added by this PR, even though .gitignore already excludes *.egg-info/. These look like accidentally force-added local build artifacts rather than intentional changes.

Fix this →

Minor / non-blocking:

• _find_meta_entry in yamldoc/parser.py does a pre-order search by name and returns the first match. If two different objects in the tree happen to share a key name (e.g. two.entry and three.entry), schema type resolution could apply the wrong type to the wrong node. Not exercised by current fixtures; probably fine for now given typical config-file key naming, but worth a comment or a path-qualified lookup if this becomes an issue in practice.

Nice work overall — the sub_stack design is the right fix for the previous 2-level ceiling, and the test fixtures (deeper_nesting.yaml/.schema, exclusion case) are well-targeted at the new code paths.