refactor manage columns to use vault by thehabes · Pull Request #450 · CenterForDigitalHumanities/TPEN-interfaces

thehabes · 2026-02-09T21:35:56Z

No description provided.

Copilot

Pull request overview

Refactors the “manage columns” and “create column” interfaces to load IIIF resources via the shared vault utility instead of ad-hoc fetch calls.

Changes:

Replaces direct fetch(...).json() calls with vault.get(...) / vault.getWithFallback(...) for annotation pages, canvases, and annotations.
Removes local URL validation + “fetch specific type” helper methods from both components.
Adds clearer error handling when page/canvas resources fail to load.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
interfaces/manage-columns/index.js	Switches annotationpage/canvas/annotation loading to `vault`, removes bespoke fetch helpers.
components/create-column/index.js	Mirrors the same `vault`-based IIIF fetching refactor as manage-columns.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-09T21:42:31Z

interfaces/manage-columns/index.js

-                const data = await res.json()
+                let data = anno
+                if (!data?.target) {
+                    data = await vault.get(anno.id ?? anno, 'annotation', true)


vault.get(anno.id ?? anno, ...) can end up passing a non-string object into vault.get when anno is an object without an id (since anno.id is undefined, the ?? falls back to the whole object). If vault.get expects an ID/URI, this will fail at runtime. Consider normalizing to an ID first (e.g., only call vault.get when you have a string ID), and return null (or throw) when no ID is available.

Suggested change

data = await vault.get(anno.id ?? anno, 'annotation', true)

const annotationId =

typeof anno === 'string'

? anno

: (typeof anno?.id === 'string' ? anno.id : null)

if (!annotationId) return null

data = await vault.get(annotationId, 'annotation', true)

components/create-column/index.js

Copilot · 2026-02-09T21:42:32Z

interfaces/manage-columns/index.js

+        if (!pageID) throw new Error("No page ID provided")
+        const annotationPageData = await vault.getWithFallback(pageID, 'annotationpage', TPEN.activeProject?.manifest, true)
+        if (!annotationPageData) throw new Error("Failed to load annotation page")
+        const canvasData = await vault.getWithFallback(
+            annotationPageData.target, 'canvas', TPEN.activeProject?.manifest
+        )
+        if (!canvasData) throw new Error("Failed to load canvas data")


The same vault.getWithFallback page/canvas loading flow appears in both interfaces/manage-columns/index.js and components/create-column/index.js. To prevent drift (e.g., different error handling, fallback behavior, or resource typing over time), consider extracting this into a shared utility (e.g., loadAnnotationPageAndCanvas(pageID, manifest)), used by both components.

cubap

ok

@deprecated

* Hydrate from seed when fetch fails Introduce a seed-based hydration path and helper (hydrateFromObject) so an already-provided item object can be used if the network fetch fails. Reorder logic to define hydration first, use the seed fallback on non-OK responses and exceptions, and return the hydrated object. Also adjust calls to this.get and this.set to pass the embedded object and the correct type. Local caching to localStorage is preserved. * runVaulttest * vault all fetches * Prefetch manifests before resolving canvases Add Vault.prefetchDocuments to eagerly cache manifest documents and use it when a canvas or annotation page cannot be resolved from the vault. Components updated (continue-working, transcription-block, transcription interface) now attempt to hydrate all project manifests and retry vault lookups before falling back to placeholders. getImage was converted to async and now uses vault.get (with prefetch) instead of direct fetch, with explicit fallbacks to 404/no-image assets. These changes reduce failed network fetches and improve resolution of IIIF v2/v3 canvases and pages. * Use vault and prefetch manifests; enhance Vault Replace ad-hoc fetches with the centralized vault API across multiple components and add a manifest prefetch fallback when an item is not immediately resolvable. Components updated: annotorious-annotator/line-parser, column-selector, continue-working, default-transcribe, legacy-annotator/plain, line-image, simple-transcription — they now import and use the shared vault, attempt vault.prefetchDocuments(TPEN.activeProject.manifest) when a resource is missing, and retry resolving. Default-transcribe and others now use vault.get for annotation/page/contentresource resolution and show clearer failure messages. Major Vault improvements (js/vault.js): - Normalizes IDs (removes fragment, expands 24-char Mongo hex IDs to TPEN.RERUMURL/id/...). - Deduplicates concurrent fetches with inFlightPromises to avoid duplicate network requests. - Improved hydration: walks embedded objects, queues and caches known IIIF resource types, skips common non-resource properties, normalizes and caches discovered URLs/IDs, and stores data keyed by detected resource type when available. - More resilient fetch handling (returns null on failure if a seed exists) and safer localStorage writes keyed to resource type. - PrefetchDocuments now calls vault.get with a best-effort type and collects errors. - Removed the debug runVaultTest export. These changes make resource resolution more consistent and robust, reduce duplicate network loads, and allow components to recover by hydrating manifests when resources are initially missing. * Process IIIF resources and simplify ID normalization Treat canvas/manifest/collection IDs as external/embedded (return empty URL) and simplify _normalizeId to only strip fragment identifiers. Add a _processIIIFResource helper to detect IIIF resources, call get() for them, and return a minimal {id,type,label} representation while avoiding duplicate fetches via the visited set. Update _fetchAndHydrate to invoke this helper for array items and non-array object values. Remove previous behavior that auto-prefixed Mongo hex IDs with TPEN.RERUMURL and the older string-URL detection/fetch logic to better handle embedded IIIF manifests and avoid incorrect fetching. * Update interfaces/transcription/index.js Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Update components/read-only-transcribe/index.js Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * Guard against empty URI in vault fetch to prevent unnecessary requests (#437) * Initial plan * Add guard to prevent fetch with empty URI Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Add prefetchManifests and prefetchCollections aliases with explicit type parameters (#438) * Initial plan * Add prefetchManifests and prefetchCollections aliases and update all usages Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Refactor prefetch methods to use docType parameter instead of duplicating code Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Eliminate N+1 fetches for embedded IIIF resources in vault (#439) * Initial plan * Fix N+1 fetch issue by caching embedded IIIF resources directly Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Improve comments explaining embedded object detection logic Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Improve embedded object detection with specific content properties Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Allow annotations with body OR target to be cached as embedded Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Add explicit IIIF v2 prefixed types to resource set (#440) * Initial plan * Fix IIIF v2 prefix handling in type normalization Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Move prefix pattern to class constant and document resource types Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Revert prefix stripping; explicitly add IIIF v2 prefixed types Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Clarify documentation for agent and annotationlist types Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Update vault.js * deprecate these * strip out TPEN.js and vault.js This component should not include these dependencies. * Import vault in read-only-transcribe component Add an import for the vault module (../../js/vault.js) in components/read-only-transcribe/index.js so the component can access vault-related functionality. * This will work here (#443) * Early guard to avoid NPEs when this.#transcriptions cannot be used (#444) * Address Issues 13-17 from static review: Fix vault consistency, error handling, and performance issues (#442) * Initial plan * Address Issues 13-17: Fix vault consistency, error handling, and performance issues Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> * Fix vault.js issues 8-11: Hoist constants, fix noCache, clone data, add getWithFallback helper (#445) Issues Fixed: - Issue 8: Move SKIP_PROPERTIES and IIIF_RESOURCE_TYPES to module scope to prevent recreation on every fetch (28 + 13 = 41 object allocations eliminated per fetch) - Issue 9: Extract vault.getWithFallback() helper method to consolidate repeated 5-line fallback pattern used in 10+ locations - Issue 10: Fix noCache parameter to bypass in-memory store, not just localStorage. When noCache=true, caller gets fresh fetch, not stale cached copy - Issue 11: Clone data with structuredClone() before BFS mutation to prevent corrupting caller's original object via shared reference Performance Impact: - Eliminates unnecessary Set allocations (Issue 8) - Reduces code duplication and maintenance burden (Issue 9) - Fixes cache bypass behavior (Issue 10) - Prevents data corruption bugs (Issue 11) * Update Conditionals to avoid NPEs (#447) * update conditions to fix pre existing errors * update conditions to fix pre existing errors * Use vault.getWithFallback for fetching (#449) Replace repeated vault.get + prefetchManifests patterns with a single vault.getWithFallback utility to simplify resource resolution and avoid duplicated manifest-hydration logic. Affected components: annotorious-annotator/line-parser.js, column-selector/index.js, continue-working/index.js, default-transcribe/index.js, legacy-annotator/plain.js, simple-transcription/index.js, transcription-block/index.js, interfaces/transcription/index.js. Preserves existing behavior (including the optional strict flag) while centralizing the fallback to project manifests. * refactor manage columns to use vault (#450) * refactor manage columns to use vault * refactor manage columns to use vault * hotfix for error messaging * No 'Auto Parse' for now * No 'Auto Parse' for now * Fix so CONTRIBUTOR users see the link into the /annotator interface * See transcription text in .transcription-input on the transcription interface. * Don't change @deprecated components * Changes while reviewing * Changes during review * Changes during review * zooming on view transcription interface * Don't let view transcription interface be larger than the viewport and cause scrolling * double right-click to zoom out * double right-click to zoom out * small change for accessibility * small change for accessibility. Add some consistency around how these buttons show or not. * small change for accessibility. Add some consistency around how these buttons show or not. * whoops make it prettier * changes while reviewing * remove references to elements that are no longer used * Extra defensive for an extra chance to find the image * some cleanup * Changes while reviewing * Changes while reviewing * Changes while reviewing * Changes while reviewing --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com> Co-authored-by: Bryan Haberberger <bryan.j.haberberger@slu.edu>

@deprecated

* Hydrate from seed when fetch fails Introduce a seed-based hydration path and helper (hydrateFromObject) so an already-provided item object can be used if the network fetch fails. Reorder logic to define hydration first, use the seed fallback on non-OK responses and exceptions, and return the hydrated object. Also adjust calls to this.get and this.set to pass the embedded object and the correct type. Local caching to localStorage is preserved. * runVaulttest * vault all fetches * Prefetch manifests before resolving canvases Add Vault.prefetchDocuments to eagerly cache manifest documents and use it when a canvas or annotation page cannot be resolved from the vault. Components updated (continue-working, transcription-block, transcription interface) now attempt to hydrate all project manifests and retry vault lookups before falling back to placeholders. getImage was converted to async and now uses vault.get (with prefetch) instead of direct fetch, with explicit fallbacks to 404/no-image assets. These changes reduce failed network fetches and improve resolution of IIIF v2/v3 canvases and pages. * Use vault and prefetch manifests; enhance Vault Replace ad-hoc fetches with the centralized vault API across multiple components and add a manifest prefetch fallback when an item is not immediately resolvable. Components updated: annotorious-annotator/line-parser, column-selector, continue-working, default-transcribe, legacy-annotator/plain, line-image, simple-transcription — they now import and use the shared vault, attempt vault.prefetchDocuments(TPEN.activeProject.manifest) when a resource is missing, and retry resolving. Default-transcribe and others now use vault.get for annotation/page/contentresource resolution and show clearer failure messages. Major Vault improvements (js/vault.js): - Normalizes IDs (removes fragment, expands 24-char Mongo hex IDs to TPEN.RERUMURL/id/...). - Deduplicates concurrent fetches with inFlightPromises to avoid duplicate network requests. - Improved hydration: walks embedded objects, queues and caches known IIIF resource types, skips common non-resource properties, normalizes and caches discovered URLs/IDs, and stores data keyed by detected resource type when available. - More resilient fetch handling (returns null on failure if a seed exists) and safer localStorage writes keyed to resource type. - PrefetchDocuments now calls vault.get with a best-effort type and collects errors. - Removed the debug runVaultTest export. These changes make resource resolution more consistent and robust, reduce duplicate network loads, and allow components to recover by hydrating manifests when resources are initially missing. * Process IIIF resources and simplify ID normalization Treat canvas/manifest/collection IDs as external/embedded (return empty URL) and simplify _normalizeId to only strip fragment identifiers. Add a _processIIIFResource helper to detect IIIF resources, call get() for them, and return a minimal {id,type,label} representation while avoiding duplicate fetches via the visited set. Update _fetchAndHydrate to invoke this helper for array items and non-array object values. Remove previous behavior that auto-prefixed Mongo hex IDs with TPEN.RERUMURL and the older string-URL detection/fetch logic to better handle embedded IIIF manifests and avoid incorrect fetching. * Update interfaces/transcription/index.js * Update components/read-only-transcribe/index.js * Guard against empty URI in vault fetch to prevent unnecessary requests (#437) * Initial plan * Add guard to prevent fetch with empty URI --------- * Add prefetchManifests and prefetchCollections aliases with explicit type parameters (#438) * Initial plan * Add prefetchManifests and prefetchCollections aliases and update all usages * Refactor prefetch methods to use docType parameter instead of duplicating code --------- * Eliminate N+1 fetches for embedded IIIF resources in vault (#439) * Initial plan * Fix N+1 fetch issue by caching embedded IIIF resources directly * Improve comments explaining embedded object detection logic * Improve embedded object detection with specific content properties * Allow annotations with body OR target to be cached as embedded --------- * Add explicit IIIF v2 prefixed types to resource set (#440) * Initial plan * Fix IIIF v2 prefix handling in type normalization * Move prefix pattern to class constant and document resource types * Revert prefix stripping; explicitly add IIIF v2 prefixed types * Clarify documentation for agent and annotationlist types --------- * Update vault.js * deprecate these * strip out TPEN.js and vault.js This component should not include these dependencies. * Import vault in read-only-transcribe component Add an import for the vault module (../../js/vault.js) in components/read-only-transcribe/index.js so the component can access vault-related functionality. * This will work here (#443) * Early guard to avoid NPEs when this.#transcriptions cannot be used (#444) * Address Issues 13-17 from static review: Fix vault consistency, error handling, and performance issues (#442) * Initial plan * Address Issues 13-17: Fix vault consistency, error handling, and performance issues --------- * Fix vault.js issues 8-11: Hoist constants, fix noCache, clone data, add getWithFallback helper (#445) Issues Fixed: - Issue 8: Move SKIP_PROPERTIES and IIIF_RESOURCE_TYPES to module scope to prevent recreation on every fetch (28 + 13 = 41 object allocations eliminated per fetch) - Issue 9: Extract vault.getWithFallback() helper method to consolidate repeated 5-line fallback pattern used in 10+ locations - Issue 10: Fix noCache parameter to bypass in-memory store, not just localStorage. When noCache=true, caller gets fresh fetch, not stale cached copy - Issue 11: Clone data with structuredClone() before BFS mutation to prevent corrupting caller's original object via shared reference Performance Impact: - Eliminates unnecessary Set allocations (Issue 8) - Reduces code duplication and maintenance burden (Issue 9) - Fixes cache bypass behavior (Issue 10) - Prevents data corruption bugs (Issue 11) * Update Conditionals to avoid NPEs (#447) * update conditions to fix pre existing errors * update conditions to fix pre existing errors * Use vault.getWithFallback for fetching (#449) Replace repeated vault.get + prefetchManifests patterns with a single vault.getWithFallback utility to simplify resource resolution and avoid duplicated manifest-hydration logic. Affected components: annotorious-annotator/line-parser.js, column-selector/index.js, continue-working/index.js, default-transcribe/index.js, legacy-annotator/plain.js, simple-transcription/index.js, transcription-block/index.js, interfaces/transcription/index.js. Preserves existing behavior (including the optional strict flag) while centralizing the fallback to project manifests. * refactor manage columns to use vault (#450) * refactor manage columns to use vault * refactor manage columns to use vault * hotfix for error messaging * No 'Auto Parse' for now * No 'Auto Parse' for now * Fix so CONTRIBUTOR users see the link into the /annotator interface * See transcription text in .transcription-input on the transcription interface. * Don't change @deprecated components * Changes while reviewing * Changes during review * Changes during review * zooming on view transcription interface * Don't let view transcription interface be larger than the viewport and cause scrolling * double right-click to zoom out * double right-click to zoom out * small change for accessibility * small change for accessibility. Add some consistency around how these buttons show or not. * small change for accessibility. Add some consistency around how these buttons show or not. * whoops make it prettier * changes while reviewing * remove references to elements that are no longer used * Extra defensive for an extra chance to find the image * some cleanup * Changes while reviewing * Changes while reviewing * Changes while reviewing * Changes while reviewing --------- Co-authored-by: Patrick Cuba <cubap@slu.edu> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com> Co-authored-by: cubap <1119165+cubap@users.noreply.github.com>

thehabes added 2 commits February 9, 2026 15:34

refactor manage columns to use vault

12e75a7

refactor manage columns to use vault

f1a55dd

thehabes requested a review from Copilot February 9, 2026 21:40

Copilot started reviewing on behalf of thehabes February 9, 2026 21:40 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

thehabes self-assigned this Feb 9, 2026

thehabes marked this pull request as ready for review February 9, 2026 21:43

thehabes requested a review from cubap February 9, 2026 21:43

cubap approved these changes Feb 9, 2026

View reviewed changes

cubap merged commit 3be1a6e into vaulting Feb 9, 2026
7 checks passed

cubap deleted the manage-columns-vaulting branch February 9, 2026 21:52

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

refactor manage columns to use vault#450

refactor manage columns to use vault#450
cubap merged 2 commits intovaultingfrom
manage-columns-vaulting

thehabes commented Feb 9, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

cubap left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-                    data = await vault.get(anno.id ?? anno, 'annotation', true)
+                    const annotationId =
+                        typeof anno === 'string'
+                            ? anno
+                            : (typeof anno?.id === 'string' ? anno.id : null)
+                    if (!annotationId) return null
+                    data = await vault.get(annotationId, 'annotation', true)

Conversation

thehabes commented Feb 9, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

cubap left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants