Scanning a Configuration Export

The config scan is the Discovery input source that answers questionnaire questions from your system's own configuration. You drop a configuration export into the Cycle project, choose config scan at the hub, and the scan settles every question the export can settle. What is left is the set of questions that need a person.

Blue Yonder focus

The config scan reads a Blue Yonder Warehouse Management System (WMS) configuration export in this iteration.

Why scan a configuration export?

The questionnaire asks how the system under test is set up, and much of that is already recorded in the system itself.

• Less to answer by hand: Questions the configuration can settle are settled before anyone is asked.
• Answers grounded in the configuration: The value comes from what the system is actually set to, rather than from someone's recollection of it.
• A reviewable trail: Every scan answer records that the scan supplied it, along with the configuration it read, so you can check the reasoning rather than trust it.
• Repeatable: Correct or extend the export and scan again, as often as the export improves.

How the scan works

Most questionnaire questions carry a query written against the configuration. The scan runs those queries against your export and records each answer it can settle confidently. A question it cannot settle confidently is left outstanding for you.

Three behaviors are worth knowing before you run it.

The scan only asks about questions that currently apply. Questions sitting behind an unanswered gate are skipped rather than guessed at. The scan repeats its pass as answers accumulate, so a question opened up by an answer the scan just recorded is still reached.

Answering a gate can settle questions behind it. When you answer a question that other questions depend on, the scan fills in behind it from the export, and the agent tells you which questions that settled. Answering one question can therefore complete several. This pass only touches questions you have not answered.

Some questions are always yours. A handful carry no query at all, because no configuration value can settle them. The scan reports these as skipped and never attempts them.

Preparing the configuration export

1. At the input-source hub, tell the agent you want to run the config scan.
2. The agent creates a Config-Export folder in your Cycle project and reports its location. Because the folder is new and still empty, the agent asks you to populate it. This is the normal first pass, not an error.
3. Copy your configuration export into that folder. The export is one CSV file per configuration table, each named after the table it holds.
4. Choose config scan again. The agent runs the scan and reports what it resolved.

A populated folder looks like this:

Config-Export/
  _model.csv
  aisle.csv
  building.csv
  carrier.csv
  carrier_detail.csv
  location.csv
  putaway_rule.csv.ignore
  shipment_status.csv
  ...

The names above are illustrative. Real table names come from the system under test, and a real export holds far more files than this.

_model.csv is the export's own metadata rather than a configuration table, and the scan leaves it alone.

Parking a file without removing it

A file whose name ends in .ignore is skipped. Rename a CSV you want the scan to leave alone rather than deleting it.

Running the scan

The export replaces answers you have already given

The configuration export is authoritative. The scan overwrites every answer it can settle, including answers you entered by hand, and it lists each replacement with the previous value and who supplied it. Run the scan before working the questionnaire where you can, and read the list of replacements when you run it afterward.

The scan reports four counts.

Count	What it means
answered	The query ran and settled the question. The answer is recorded.
unresolved	The query ran and the export could not answer it. These questions need a person.
deferred	The query never ran, because the question did not apply at the time.
skipped	No query exists for the question. These questions are always answered by a person.

Deferred questions are not a shortfall in the export

A question is deferred because it was gated, not because the export fell short. A question gated behind something nobody has answered yet fills itself once that gate is answered. A question gated against by an answer does not apply at all unless that answer changes.

To scan again after adding or correcting CSVs, choose config scan at the hub a second time. The scan re-runs against the folder's current contents and reports what changed.

Reviewing what the scan answered

Ask the agent to review a questionnaire section. Each applicable question comes back with its current value, whether you or the scan supplied it, and the configuration the scan read to reach it. Change any of them from that review, the same way you would change an answer you gave yourself. See Changing an answer.

Run this review after any scan over a questionnaire you had already worked on.

Example

You say:

I want to run the config scan.

The agent:

1. Creates the Config-Export folder in your project and asks you to copy the configuration export into it.

You say:

The export is in. Go ahead.

The agent:

1. Runs each applicable question's query against the export.
2. Records the answers it settled confidently, attributing each one to the scan.
3. Reports its counts, and lists any answer it replaced along with the previous value.
4. Returns you to the hub with the config scan marked as done and the questionnaire partly answered.

Answering the questions the scan left

• Answering Questions: Work through what the export could not settle, and check what the scan recorded.
• Ingesting Documents: The third input source, which produces test specifications rather than answers.