LUNA Lab
EN
Help
Navigation
Choose language
Print viewPDFMarkdown (.md)Text (.txt)Export this page’s content and sources.

Collection and synchronization

Find a task, follow the steps, and check the result. Guidance for researchers and administrators, in both languages.

Reviewed 2026-10-12 · 50 topics

Search these guides in Korean or English. This does not search private research materials.

Choose another topic

Distinguish source registration, authorization and completed collection.

Reviewed · 2026-10-12

On this page · 9

1. Registration versus collection

  1. Open Grants → grant or child project → External source settings. Consider a grant-level connection for shared sources.
  2. Choose Add connection and a provider. Follow the provider-specific GitHub, Notion, Google Drive or Zotero chapter for authentication.
  3. Saving a connection does not complete collection. Click Collect now on its card and refresh the status.
  4. Check the latest time, count and status, then open a collected item and compare its title and body with the original.
Back to search and contents ↑

2. Keep collecting changes

  1. Open connection settings, enable the connection and periodic collection, and set an interval such as 60 minutes.
  2. A lookback window can exclude older documents. Consider an unrestricted initial window for historical plans.
  3. Edit the source and wait for collection or click Collect now. Collection time and source modification time are different.
  4. Inspect the new version and body. Recollecting unchanged content is not a new research result. Periodic collection does not create AI report schedules.
Back to search and contents ↑

3. Failed, partial or stopped collection

  1. Expand Collection history and details on the connection card and read its error code.
  2. For authentication_or_configuration, check source permissions, token expiry and Google authorization. Entering a folder URL alone is not Google authentication.
  3. Partial collection can reflect format, body-size, request or descendant limits. Verify required documents rather than treating it as complete.
  4. To pause, disable the connection or periodic collection and save. Deleting a connection does not delete the external repository, Notion page or Drive file.
Back to search and contents ↑

4. Grant-level connections and personal accounts

  1. Open a grant's External source connections. A connection manager can add a provider here for shared sources, or at project level for project-specific material.
  2. Use a recognizable label such as Shared Zotero references or Project design Notion. Never put tokens or internal secrets in the label.
  3. During Google authorization choose your account with access to the actual folder/calendar. Manage separate accounts through separate cards rather than borrowing another researcher's authorization.
  4. GitHub organization administration does not replace repository/Projects token permissions, organization approval or SSO authorization. Save, collect and open the resulting material.
  5. For a shared GitHub connection choose GitHub → Repository or Projects board. Enter https://github.com/OWNER/REPO for a repository, or https://github.com/orgs/ORG/projects/NUMBER and the repository to collect for a board. Personal boards use a users URL. Use a recognizable label and put the token only in the token field.
  6. Save → Collect now → the grant research work page to inspect boards and material. Child projects also show the shared connection without copying it. Grants without child projects work too. Grant owners, delegated grant managers and owners or delegated managers of active child projects can add, edit, authorize and collect shared connections. Ordinary members cannot manage connections. Delegation is an actual management permission, not a role label. Grant settings, membership administration and common board field/work changes still require grant management rights. Remote writes also require token permissions and explicit confirmation.
Back to search and contents ↑

5. Overleaf and OneDrive source links

  1. Choose Overleaf or OneDrive under Add external connection and enter a recognizable label and source URL.
  2. Save and open the link in the external service. Viewers also need access there.
  3. Direct connections currently store source links. They do not automatically collect every file/PDF or turn them into AI input.
  4. To read an Overleaf GitHub copy, arrange authorized synchronization in the source service, then connect and collect the GitHub repository separately. Check mirror freshness at the source.
Back to search and contents ↑

Root-page and nested child documents

Direct child pages on the connected root and nested child documents are collected too. Non-database child documents appear as reports by default; database entries retain their configured database roles. The root itself remains an overview material. Reports published by the platform retain their existing report IDs without duplicate registration. Automatic collection preserves manual classification. An explicitly requested administrator repair compares the current preview and changes classification while preserving notes, tags, researchers, dates, original text and history. Source permissions and connection scope do not expand. Large trees continue across batches: distinguish the last check from a completed scan. No additional AI call is made.

Back to search and contents ↑

Resume collection and map databases

Notion and OneNote save unfinished scans and continue them later. Last check and last complete scan are separate. Open Notion database classification and coverage, select roles for discovered database IDs and preview changes to existing materials. Manual edits and material IDs remain intact. Denied source access restricts reading; a single missing inventory entry does not delete history. Optional attachment preservation uses the existing private archive within size limits and makes no AI or OCR calls.

Back to search and contents ↑

Connect a Slack channel

Choose Slack and a channel URL in the grant or project source settings, then save. Ask a workspace administrator to approve the internal app and invite it to that channel before connecting Slack. Missing server app settings remain pending. Only messages and threads in the selected public or private channel become materials. DMs, message sending and automatic AI reports are excluded. Attachments retain source references and file metadata; this does not mean their bytes have been preserved. Collection makes no AI calls. Slack plan, retention and API limits still apply.

Back to search and contents ↑

Connection recovery and progress

Full traversal, text collection and attachment preservation are separate states. Rate limits and temporary outages resume at the displayed retry time, with longer intervals after repeated failures. Reconnect the affected account when consent has expired or been revoked. Google source reading and Drive preservation have separate grants. Google Testing status or administrator restrictions cannot be repaired by server retries alone. Check the connection code and app administrator settings. Collection runs separately from AI work and never enables disabled automation.

Back to search and contents ↑

Still need help?

Include the page, time, steps and error message. Leave out passwords, tokens and private research text.

Contact support →