User Journey: Seed a fresh environment from ORE sample directories

"I have a fresh party and no trades. I want the ORE sample directories loaded so I can book, price and report without building data by hand." The journey is done when the sample trades read in books under a named portfolio, the sample conventions are stored, and the person has the counts of what arrived.

The actor is an operations user or tenant administrator who is signed in to the new party and holds the import permissions. The trigger is a new installation, a reset environment, or a new party that should start from known data. The journey starts with the ORE sample directories on disk and ends when the person has read the import counts and run one report.

This is not the same thing as the 2016 ORE Samples synthetic collection. That collection generates market data from a named feed. This journey imports the sample trade and configuration files.

1. Choose the directories. The person picks one ORE sample tree, or several, in the layout ORE ships: Example_1/Input/portfolio.xml, conventions.xml, curveconfig.xml, todaysmarket.xml, pricingengine.xml and ore.xml. The person may skip Output and other folders that hold results rather than inputs.
2. Upload the trade tree. The person runs ore upload <src_dir> --request-id <uuid>. The system packs and uploads it and prints the request id and the object key.
3. Import the trades. The person runs ore import <request_id> --parent-portfolio-name "<name>" --official or targets the open sample sandbox. The system scans every XML file, classifies it by root element, builds portfolios and books from the folder names, creates missing currencies, and books the trades. The shell waits for the workflow.
4. Import the configuration. For each sample the person runs ore import-run <report_definition_id> <src_dir>. The system creates the binding under the report definition and stores the conventions, curves, today's market and pricing engines. Repeat for each sample that carries a run.
5. Load the market. The person either runs marketdata import with the sample's market.txt and fixings.txt, or starts the synthetic collection the party is seeded with. Only one source of the same series should run at a time.
6. Read the result. The person reads the counts and runs one report to prove the data is usable.

The shell prints the saved currency, portfolio, book and trade counts, the run import's stored files, and the market data counts. The trade list shows the imported trades in their books under the named portfolio. The conventions screens show the stored families. A report runs against the seeded data.

A directory with no portfolio.xml imports no trades and says so through the counts. A file with an unexpected root element is ignored by the scanner, and the person sees fewer trades than expected, not an error. A run import refuses when the run document names a file the input does not hold, or when a file does not parse, before it writes. A trade whose counterparty cannot be resolved is reported per item and left out. A missing engine package stops the report, not the import.

• A second run of the same seed must never double the trades. Today nothing matches an imported trade by its ORE id, so a retry does duplicate it.
• An import must never overwrite a live portfolio's trades. The default re-import mode attaches new trades and leaves existing books alone.
• A currency the tenant already holds must never be replaced.
• A sample's conventions must never be stored as another party's rows.

There is no screen. The entry points are the shell verbs ore upload, ore import, ore import-run and marketdata import. The web UI has a conventions screen and a market data screen, but neither has a seed or import action.

1. No batch seed. Each sample is two or three separate commands. Nothing takes a root folder of samples and imports the whole set in one run.
2. The scan is invisible. The person cannot list the files the scanner found, or the ones it ignored, before the import. The existing portfolio import journey carries the same gap.
3. No trade identity check. A re-import creates new trade ids, so there is no way to say "this ORE trade is already here".
4. No dry run. The counts are available only after the writes.

• Import a portfolio from ORE - the detailed trade import.
• Retry a failed or interrupted import - what to do when a seed stops.
• Onboard a party with its conventions and portfolio - the smaller version for one party.

• Trade journeys
• Conventions journeys

• Data exchange journeys - the group hub.