> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dataerai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Manage the Zenodo import (admin)

> Configure, trigger, and monitor the automated Zenodo harvest from the system-admin panel.

The **Zenodo import** admin panel lets a system administrator run and monitor the
automated Zenodo harvest — the background job that discovers Zenodo records and
imports their metadata into a project — without using the command line.

<Note>
  This panel is only available to **system administrators**. Open it from the
  account menu (top right) → **Zenodo import**.
</Note>

## Set the parameters

The panel opens with a **Parameters** form, pre-filled from the current schedule
(or system defaults if none exists yet):

* **Project ID** (required) — the UUID of the target project that will own the imported records.
* **Creator ID** — optional; defaults to the **Zenodo** data-source account, so
  imports are attributed to the source rather than to you.
* **Collection ID** — optional; defaults to a public **Zenodo** folder in the
  project (created on first run). The folder and its imported records are
  publicly readable.
* **From date** / **Until date** — optional `YYYY-MM-DD` window to limit which
  records are discovered.
* **OAI set** / **OAI endpoint** / **Metadata prefix** — the Zenodo OAI-PMH source
  settings (defaults are correct for Zenodo; change only if you know you need to).
* **Max records / run** — cap per run (`0` = no cap), plus **Sync page size**,
  **Parallelism**, and **Fetch timeout** for tuning throughput.

## Incremental by default

Each run imports only what's **new since the last successful run** — the panel
shows the watermark ("the next run imports records since …") beneath the
parameters. The first run does a full harvest; later runs pick up only newer
records. Tick **Full re-harvest** to ignore the watermark and re-scan the whole
source (e.g. after changing filters).

## Run it now

Click **Run now** to start a single import immediately with the parameters in the
form. A confirmation shows the started workflow, and the run appears under
**Recent runs** as it progresses.

## Schedule recurring imports

To run the import automatically on a cadence:

1. Set **Schedule enabled (unpaused)** and either an **Interval (minutes)** (at
   least 15) or a **Cron** expression (cron overrides the interval).
2. Click **Save schedule**.

The **Schedule** section shows whether the schedule is enabled or paused, the
cadence, and the next scheduled run. Use **Delete schedule** to remove it.

## Watch progress

**Recent runs** lists each import with its status and processed / failed counts.
Select a run to expand its detail:

* per-outcome counts (imported, updated, unchanged, gone, failed),
* how many records have been **materialized** as assets in the target project,
* the live **Temporal** workflow state for the run, with a link to the in-app
  **Temporal jobs** monitor (and an external Temporal link when your deployment
  is configured for one),
* any recent failed records.

The panel refreshes on its own while a run is in progress.

## Retry failed records

If a managed run finishes as **Partial** or **Failed** and has failed records,
expand the run and click **Retry failed**. Confirm the dialog to start a new
retry workflow for only the failed records in that batch.

Retry does not scan Zenodo again. It reuses the existing batch ledger, resets
failed records so they can run again, and keeps successful and skipped records
unchanged. Use it when a transient Zenodo or network problem left otherwise
valid records in the failed state.