Add crate2tables plugin: export RO-Crate tables via roctable - #16
Merged
Conversation
Flattens a built crate into one CSV per entity type using roctable (github:ptsefton/roctable, WIP/not on npm), config-driven the same way its own inspect/csv CLI works: discovers types/properties into crate2tables-config.json on every build (non-destructive to existing choices), extracts+writes CSV for whatever the config already selects into crate2tables-output/. load_text is disabled with a warning rather than crashing, since roctable's implementation reads files via Node's fs, not available in chaos2crate's browser build. See chaos2crate/docs/crate2tables-spec.md for the full design, config schema, and the Phase 2 UI plan for interactive type/field selection.
roctable's extractTables() now takes an optional fileReader instead of reading load_text files via Node fs (ptsefton/roctable#1). This plugin passes browserFileReader(dirHandle), wrapping readFileTextFromDirectory — dropping the stripUnsupportedFeatures workaround that disabled load_text with a warning. Depends on roctable's file:../roctable sibling checkout while both are under active development — package.json points there for now instead of the github: dependency, to be switched back once ptsefton/roctable#1 merges. Closes #15
…lace it Reads ro-crate-metadata.json from the folder (if present) and passes it through to buildCrate() as opts.existingJson, so a build against an already-curated crate updates it instead of throwing it away — see chaos2crate SPEC.md §6.1a and its crate.js change for what reconciling actually does.
A file the scan found with no matching entity in the existing crate (reconciliation, chaos2crate SPEC.md §6.1a) was previously added automatically, attached to the root dataset with only a warning after the fact — a silent guess about a decision a curator should actually make. buildCrate(ctx) now diffs the scan against the existing crate's @ids first and, if anything's new, blocks on a confirmation modal (new-files-confirm.js): a checkbox file tree, everything checked by default, folders reflecting all/some/none of their contents. Only the confirmed subset reaches buildCrate() at all — an unchecked file simply isn't part of this build, same as one the scan never saw, and gets asked about again next time unless it's removed from the folder. Dismissing the modal outright (as opposed to confirming even an empty selection) cancels the whole build, since that made no explicit choice. Verified the orchestration (fresh build/no-new-files/confirmed/skipped/ cancelled) with a mocked openModal, and the tree itself (build, render, checkbox↔indeterminate propagation on toggle, confirmed-result shape) with a minimal fake DOM, since this repo has no jsdom dependency.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
crate2tablesbuild plugin: exports a built RO-Crate as one CSV per entity type usingroctable(config-driven the same way its owninspect/csvCLI works — discovers types/properties intocrate2tables-config.jsonon every build, extracts+writes CSV for whatever the config already selects).load_textsupport wired to roctable's injectable file reader (Make load_text's file read injectable instead of hardcoded to Node fs ptsefton/roctable#2) viabrowserFileReader(dirHandle), wrappingreadFileTextFromDirectory— no more Node-fslimitation.generic-inputnow reconciles against an existing crate instead of replacing it: readsro-crate-metadata.jsonfrom the folder (if present) and passes it through tobuildCrate()(chaos2crate PR pending) asopts.existingJson. This fixes real data loss — building against an already-curated crate was silently discarding almost everything (reproduced against a real 764-entity collection: rebuild → 465 entities, everything but bare file stubs gone).new-files-confirm.js) before it's added — never a silent guess.Dependencies
roctableis installed as"roctable": "file:../roctable"(sibling checkout) while both repos are under active development — should switch to a pinnedgithub:ptsefton/roctablegit dependency once Make load_text's file read injectable instead of hardcoded to Node fs ptsefton/roctable#2 merges.buildCrate()'sopts.existingJson/reconciliation support and thedeletePluginOutputs()fix that lets reconciliation actually see the existing file.Test plan
crate2tables'scrate:built/output:writehooks end-to-end against an in-memory crate, including a realload_textread through the injected readergeneric-input's reconciliation end-to-end against a real 764-entity collection: no-op reconcile (0 new/0 missing), a synthetic delta (1 new/1 missing), and a genuinely new file confirmed via "Add selected" — all preserve every existing entityvite build(single-plugin and all-plugins) in chaos2crate stays clean with this registeredCloses #15