Class: Woods::Obsidian::VaultExporter

Inherits:
Object
  • Object
show all
Defined in:
lib/woods/obsidian/vault_exporter.rb

Overview

Orchestrates exporting Woods extraction output to a self-contained Obsidian vault.

Reads the flat extraction directory via IndexReader (using raw_graph_data, so string keys + persisted pagerank), then writes:

- one Markdown note per exported unit (NoteBuilder),
- per-type `_index.md` MOCs + a root `_Overview.md`,
- a `_woods/` machine sidecar (id<->path manifest + verbatim graph copies),
- `.obsidian/` config + `Units.base` (only into a woods-owned vault),
- a `.woods-vault` ownership sentinel.

Finally it sweeps stale woods-managed notes — guarded so it can never delete a user's own notes or run against a foreign vault.

All dependency edges come from one in-memory edge_map (the graph's edges/reverse); per-unit JSON is read only for note bodies.

Examples:

Woods::Obsidian::VaultExporter.new(
  index_dir: "tmp/woods", vault_path: "tmp/woods/obsidian_vault"
).export_all
# => { exported: 412, indexes: 19, swept: 3, skipped: 0, errors: [] }

Constant Summary collapse

MAX_ERRORS = 
100
PURGE_GUARD_FRACTION = 
0.30
PURGE_GUARD_MIN_DOCS = 
10
SENTINEL = 
'.woods-vault'
SIDECAR_DIR = 
'_woods'
FRAMEWORK_TYPES = 
%w[rails_source gem_source].freeze
MAX_FRONTMATTER_LINES =

Upper bound on frontmatter lines scanned when identifying a managed note during the sweep — our frontmatter is ~12 lines; this just bounds work on a malformed file that opens with "---" but never closes the fence.

200

Instance Method Summary collapse

Constructor Details

#initialize(index_dir:, vault_path:, config: Woods.configuration, reader: nil, include_source: false, include_framework: false, force_purge: false, output: $stdout) ⇒ VaultExporter

Returns a new instance of VaultExporter.

Parameters:

  • index_dir (String) —

    extraction output directory to read

  • vault_path (String) —

    directory to write the vault into

  • config (Configuration) (defaults to: Woods.configuration) —

    Woods configuration (default: global)

  • reader (Object, nil) (defaults to: nil) —

    IndexReader (auto-created if nil)

  • include_source (Boolean) (defaults to: false) —

    embed credential-scrubbed source in notes

  • include_framework (Boolean) (defaults to: false) —

    include rails_source units

  • force_purge (Boolean) (defaults to: false) —

    bypass the mass-deletion sweep guard

  • output (IO) (defaults to: $stdout) —

    progress stream



57
58
59
60
61
62
63
64
65
66
 
# File 'lib/woods/obsidian/vault_exporter.rb', line 57

def initialize(index_dir:, vault_path:, config: Woods.configuration, reader: nil,
               include_source: false, include_framework: false, force_purge: false, output: $stdout)
  @config = config
  @vault = Pathname.new(vault_path.to_s)
  @reader = reader || build_reader(index_dir)
  @include_source = include_source
  @include_framework = include_framework
  @force_purge = force_purge
  @output = output
end

Instance Method Details

#export_all ⇒ Hash

Generate the full vault. Idempotent: re-running against an unchanged extraction produces byte-identical output.

Returns:

  • (Hash) —

    { exported:, indexes:, swept:, skipped:, errors: }



72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/woods/obsidian/vault_exporter.rb', line 72

def export_all
  graph = @reader.raw_graph_data || {}
  @nodes = graph['nodes'] || {}
  @pagerank = graph['pagerank'] || {}
  analysis = safe_graph_analysis
  owned = vault_owned_at_start?

  emitted, skipped = partition_emitted
  mapper = build_mapper(emitted)
  edge_map = build_edge_map(emitted.to_set, graph)
  builder = build_note_builder(mapper, analysis)

  written = Set.new
  errors = []
  exported = write_notes(emitted, mapper, edge_map, builder, written, errors)
  indexes = write_indexes(emitted, mapper, written)
  write_sidecar(emitted, mapper, graph, analysis, written)
  write_owned_assets(emitted) if owned

  swept = owned && errors.empty? ? sweep(written) : 0
  warn_foreign unless owned
  { exported: exported, indexes: indexes, swept: swept, skipped: skipped, errors: cap_errors(errors) }
end