init: archlens — architecture diagram generator

Hex arch + DDD, tree-sitter parsing, Mermaid/ASCII output.
Supports Rust + Python. 92 tests. CI, diff, --check for staleness detection.

CONTEXT.md

@@ -0,0 +1,55 @@
+# Archlens — Domain Glossary
+
+> This file defines the canonical language of the Archlens domain. No implementation details — just terms and their meanings.
+
+## Core Concepts
+
+- **CodeGraph**: The aggregate root. A complete representation of a codebase's architecture — all elements, relationships, and module groupings. Can produce filtered subgraphs of itself by module.
+
+- **CodeElement**: A type-level construct in source code — a class, struct, trait, interface, or enum. Carries a name, kind, location (file + line), visibility, generic parameters, and attributes/decorators.
+
+- **Relationship**: A directed edge between two CodeElements. Has a kind (inheritance or composition) and references source and target elements.
+
+- **Module**: A logical grouping of CodeElements. By default inferred from directory structure, overridable via configuration. Represented as metadata on each CodeElement (`module_path`), not a separate hierarchy.
+
+- **Language**: The programming language of a source file. A closed set: Rust, CSharp, Python. Determines which analysis strategy is used.
+
+- **SourceFile**: A file to be analyzed. Carries a FilePath and its detected Language.
+
+## Analysis
+
+- **AnalysisResult**: The output of analyzing a single file — extracted CodeElements, Relationships, and any AnalysisWarnings.
+
+- **AnalysisWarning**: A non-fatal problem encountered during analysis — unparseable construct, unsupported syntax, etc. Carries file path, line, and message.
+
+- **AnalysisConfig**: Settings that control what gets analyzed — path excludes, granularity level, module-to-directory mappings.
+
+## Output
+
+- **RenderOutput**: The result of rendering a CodeGraph — a collection of RenderedFiles.
+
+- **RenderedFile**: A single output artifact — a named file with content (e.g., `overview.mmd` with Mermaid syntax).
+
+- **OutputConfig**: Settings that control rendering — format, output path, whether to split by module.
+
+## Relationships (kinds)
+
+- **Inheritance**: A type extends or implements another type (class extends class, struct implements trait, class implements interface).
+
+- **Composition**: A type owns or depends on another type via a field, property, or constructor-injected dependency.
+
+## Metadata on CodeElements
+
+- **Visibility**: Whether a type is public, private, internal, or other access level. Used as a filterable property, not shown on diagrams by default.
+
+- **Generics**: Type parameters on a CodeElement. Stored with the element, displayed in simplified form (e.g., `Repository<Order>` not full generic constraints).
+
+- **Attributes**: Decorators, annotations, or attributes on a type (e.g., `[ApiController]`, `#[derive(Clone)]`, `@dataclass`). Stored as metadata, usable for categorization and filtering.
+
+## Diagram Levels
+
+- **Project level**: Dependencies between projects, crates, or packages. Rendered as a flowchart.
+
+- **Module level**: Dependencies between logical modules (directory-based groupings). Rendered as a flowchart with subgraphs.
+
+- **Type level**: Individual types with their inheritance and composition relationships. Rendered as a UML class diagram.