feat(extract): add Pascal/Delphi and Lazarus IDE support

[simeonbodurov]

Summary

Adds full knowledge-graph extraction for Pascal/Delphi codebases and
Lazarus IDE project files. Developed with Claude Sonnet 4.6.

---

1 — Pascal/Delphi extractor (extract_pascal)

New extractor for .pas, .pp, .dpr, .dpk, .inc files via
tree-sitter-pascal.

Extracted nodes: file, unit/program/library, class/interface/helper,
procedure/function implementations.

Extracted edges:

• file → contains → module
• module → imports → dependency (uses clause, resolved to path-based IDs)
• class → inherits → base class / interface
• class/module → contains/method → procedure or function
• procedure → calls → procedure (in-file call resolution, including bare
  Reset;-style calls without parentheses)

Key design — import resolution: Uses clause targets are resolved to
path-based node IDs by scanning all Pascal files under the project root
(_pascal_project_root + _pascal_resolve_unit helpers with caching).
Without this, bare unit names like SysUtils resolve to IDs that never
match any file node, making the entire import graph invisible.

Optional dependency: pip install tree-sitter-pascal
If not installed, extract_pascal returns {"nodes":[], "edges":[], "error":...}
and the rest of the pipeline is unaffected.

---

2 — Inherits-edge fix

The original declType/typeref handler built inherits edge targets
with _make_id(_read(child)) — just the bare class name. But class nodes
use _make_id(stem, type_name), so targets never matched, making the
entire class hierarchy invisible in the graph.

Fix adds _pascal_class_stem_cache and _pascal_resolve_class(): strips
the conventional T/I prefix, locates the defining file by stem lookup
(same cache mechanism as _pascal_resolve_unit), and returns the correct
_make_id(file_stem, class_name) ID. RTL/unresolvable bases (e.g.
TObject) fall back to _make_id(bare_name) with an explicit stub node,
following the same pattern as the Python extractor.

Also removes the break that stopped after the first typeref, so all
parents of multi-inheritance declarations are captured.

---

3 — Lazarus form extractor (extract_lazarus_form, .lfm)

.lfm files are text-based UI component trees. The extractor parses
object Name: TClassName ... end blocks into a containment graph and
captures OnXxx = HandlerName event bindings as references edges
(context: "event") linking each component to its handler procedure.

---

4 — Lazarus package extractor (extract_lazarus_package, .lpk)

.lpk files are XML package definitions. The extractor reads the package
name, required package dependencies (→ imports edges), and listed unit
files (→ contains edges). Unit names are resolved to path-based node IDs
via _pascal_resolve_unit so they link to the same nodes produced by
extract_pascal on .pas files.

---

Files changed

File	Change
graphify/extract.py	extract_pascal, extract_lazarus_form, extract_lazarus_package, resolution helpers
graphify/detect.py	Add .pas .pp .dpr .dpk .inc .lfm .lpk to CODE_EXTENSIONS
tests/test_pascal.py	28 new tests covering all three extractors
tests/fixtures/sample.pas	Representative Delphi fixture
tests/fixtures/sample.lfm	Lazarus form fixture
tests/fixtures/sample.lpk	Lazarus package fixture

---

Prerequisites

pip install tree-sitter-pascal

Wheel for Windows (cp38 abi3): available at
https://github.com/Isopod/tree-sitter-pascal

Co-Authored-By: Claude Sonnet 4.6 noreply@anthropic.com

[safishamsi]

Great PR overall — follows extractor conventions, solid test coverage (28 tests), and the inherits-edge fix is a genuine improvement. Just one thing needed before merge.

Required: add .lpr extension

The PR title promises Lazarus IDE support but .lpr (the Lazarus program file — the entry point of every Lazarus project, identical Pascal syntax to .dpr) is missing from both CODE_EXTENSIONS in detect.py and _DISPATCH in extract.py. This is a one-line fix in each file, mapping .lpr to extract_pascal.

.lpi (Lazarus project info, XML format) can be skipped or added in a follow-up — that's a separate parser.

Minor (non-blocking)

• rstrip("()") strips any combination of ( and ) chars rather than the literal suffix "()" — removesuffix("()") would be more precise (though benign in practice since labels are always "Name()")
• Module-level _pascal_unit_cache / _pascal_class_stem_cache persist across extract() calls in the same process — could leak stale data if files move between runs. Low priority but worth a note.

Once .lpr is added this is a clean merge.

[safishamsi]

Merged - thank you for your first OSS contribution! Pascal/Delphi/Lazarus support is now in v7. Great to hear graphify is useful for navigating your Delphi project.