ORCESTRA data browser

This repository contains the ORCESTRA data browser.

The project contains several packages.

• utils: common utility functions
• cli: command line tools, mainly for index handling
• ui: the browser user interface as Vue 3 based web application

local development

We use node and npm. Use the following command to install all dependencies:

npm install

You might also want to (re-) build the utils package, as this is required for the others to work.

cd packages/utils
npm run build

Web UI

You can then run your own local development server for the UI using:

cd packages/ui
npm run dev

indexing

In order to display the main page quickly, we have to generate an index of the existing dataset. There's a tool to create such an index based on an IPFS unixfs folder.

collecting the index

The indexing tool expects a CID to build the index. You may want to find it using:

ipfs resolve /ipns/latest.orcestra-campaign.org/products

Which returns /ipfs/<CID>. Using only the <CID> part, you can now build an index like so:

cd packages/cli
npm run scan -- --cid <CID> -o products.json -C cache

(-C is optional and specifies a directory for caching items)

NOTE: the tool will try to figure out if you have an IPFS gateway configured loally an will use it if present. In case it doesn't find the gateway, it will run it's own IPFS node, but that might be slower than using a well-connected long-running node as gateway.

compress the index

Currently, we use the array of stac items as index for the main page (as produced by the scanning tool). This can quickly become large, especially when including trajectory data. A hacky way to compress the index can be done using jq:

cat products.json | jq 'map({properties, assets})' > products_short.json

displaying DOIs

To correctly render DOI landing pages, CIDs for which DOIs are registered must be listed in packages/ui/src/components/data/dois.json.