Library metadata generation pipeline.
These functions handle collection, validation, and transformation of library metadata during the generation pipeline.
Pipeline stages: 1. Collection - library_collect_source_files gathers and filters source files 2. Analysis - library_analyze_module (in library_analysis.ts) extracts metadata 3. Validation - library_find_duplicates checks flat namespace constraints 4. Transformation - library_merge_re_exports resolves re-export relationships 5. Output - library_sort_modules prepares deterministic output
@see library_generate.ts for the main generation entry point @see library_analysis.ts for module-level analysis @see library_output.ts for output file generation (JSON/TS wrapper) @see library_gen.ts for Gro-specific integration
6 declarations
library_pipeline.ts view source
CollectedReExport A collected re-export with its source module context.
Used during the two-phase re-export resolution:
1. Phase 1: Collect re-exports from each module during analysis
2. Phase 2: Group by original module and merge into also_exported_from
re_exporting_moduleThe module that re-exports the declaration.
stringre_exportThe re-export info (name and original module).
library_pipeline.ts view source
DuplicateInfo A duplicate declaration with its full metadata and module path.
declarationThe full declaration metadata.
DeclarationJsonmoduleModule path where this declaration is defined.
stringlibrary_pipeline.ts view source
(files: Iterable<SourceFileInfo>, options: ModuleSourceOptions, log?: Logger | undefined): SourceFileInfo[] Collect and filter source files.
Returns source files for TypeScript/JS files and Svelte components, excluding test files. Returns an empty array with a warning if no source files are found.
File types are determined by options.get_analyzer. By default, .ts, .js, and .svelte
files are supported. Customize get_analyzer to support additional file types like .svx.
filesIterable of source file info (from Gro filer, file system, or other source)
Iterable<SourceFileInfo>optionsModule source options for filtering
log?Optional logger for status messages
Logger | undefinedSourceFileInfo[] library_pipeline.ts view source
(source_json: { [x: string]: unknown; name: string; version: string; modules?: { [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; ... 19 more ...; alias_of?: { ...; } | undefined; }[] | undefined; module_comment?: string | undefined; dependencies?: string[] | undefined; dependents?: string[] | undefined; star_exports?: string[] | undefined; }[] | undefined; }): Map<...> Find duplicate declaration names across modules.
Returns a Map of declaration names to their full metadata (only includes duplicates). Callers can decide how to handle duplicates (throw, warn, ignore).
source_json{ [x: string]: unknown; name: string; version: string; modules?: { [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; ... 19 more ...; alias_of?: { ...; } | undefined; }[] | undefine...Map<string, DuplicateInfo[]> library_pipeline.ts view source
(source_json: { [x: string]: unknown; name: string; version: string; modules?: { [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; ... 19 more ...; alias_of?: { ...; } | undefined; }[] | undefined; module_comment?: string | undefined; dependencies?: string[] | undefined; dependents?: string[] | undefined; star_exports?: string[] | undefined; }[] | undefined; }, collected_re_exports: CollectedReExport[]): void Build also_exported_from arrays from collected re-export data.
This function resolves the two-phase re-export problem:
Problem: When module A re-exports from module B, we discover this while analyzing A, but need to update B's declarations. However, B may already be processed or may be processed later.
Solution: Collect all re-exports in phase 1, then merge them in phase 2 after all modules are analyzed.
source_jsonThe source JSON with all modules (will be mutated)
{ [x: string]: unknown; name: string; version: string; modules?: { [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; ... 19 more ...; alias_of?: { ...; } | undefined; }[] | undefine...collected_re_exportsArray of re-exports collected during phase 1
CollectedReExport[]void library_pipeline.ts view source
(modules: { [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; doc_comment?: string | undefined; ... 18 more ...; alias_of?: { ...; } | undefined; }[] | undefined; module_comment?: string | undefined; dependencies?: string[] | undefined; dependents?: string[] | undefined; star_exports?: string[] | undefined; }[]): { ...; }[] Sort modules alphabetically by path for deterministic output and cleaner diffs.
modules{ [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; doc_comment?: string | undefined; ... 18 more ...; alias_of?: { ...; } | undefined; }[] | undefined; module_comment?: string | un...{ [x: string]: unknown; path: string; declarations?: { [x: string]: unknown; name: string; kind: "function" | "json" | "type" | "variable" | "class" | "constructor" | "component" | "css"; doc_comment?: string | undefined; ... 18 more ...; alias_of?: { ...; } | undefined; }[] | undefined; module_comment?: string | un...