Build a contextual knowledge collection · 5 / 5

Bring in a CSV of press observations

Grow the collection with two reviewable proposals from a shift handover CSV.

YOUR TASK

Make this happen

3 steps

Expand notes.csv with for_each = csv({ path = "notes.csv", key = "code" }).
Map each.value.entity_id and each.value.body into ENTITY notes.
Keep both imported proposals TENANT/PENDING/CHAT_DISTILLED and preserve all earlier resources.

THE IDEA BEHIND IT

A real CSV expansion

The CLI reads CSV locally and expands one resource per stable code. press_b and press_c become addressed instances. Keep these keys stable across later imports; changing one changes resource identity. Quoted cells preserve commas. Headers and keys must be unique and nonempty.

Import is not publication

These rows are distilled shift observations, so retain CHAT_DISTILLED provenance and PENDING review. Use MANUAL for genuinely human-authored imports in your own projects. Checking here reads and expands the CSV but does not apply a bulk write, confirm notes or create embeddings.

Take it to the platform deliberately

Replace fixture TypeIDs with intended resources, review the CLI plan, then apply through your normal authenticated workflow and confirm appropriate notes. Do not import a personal collection under a shared service identity: the submitting user becomes its author. Do not assume missing CSV rows merely hide notes; review planned removals.

The real CLI validates declarations and expands CSV locally. Authentication, graph bindings, note confirmation, ranking and retrieval are explained, not simulated. No resources are created.

What you’re building: Help an operator find general guidance, their own qualifications and advice specific to the press they are using.

This step starts with the earlier solution and one new challenge. Each lesson keeps its own draft.

Read the starter files

concept.cococo

concept "roller" {
  label = "Roller Jamming"
  conceptKind = "failure-mode"
}

main.cococo

knowledge_note "setup" {
  kind = "CONCEPT"
  body = "In high humidity, warm the shop before a long run to reduce roller jams."
  scope = "TENANT"
  status = "PENDING"
  provenance = "CHAT_DISTILLED"
  conceptId = concept.roller.id
}

personal.cococo

knowledge_note "my_shift" {
  kind = "CONCEPT"
  body = "For my morning shift, ask the shift lead to check humidity before warming the shop."
  scope = "PERSONAL"
  status = "PENDING"
  provenance = "CHAT_DISTILLED"
  conceptId = concept.roller.id
}

press.cococo

knowledge_note "press_a" {
  kind = "ENTITY"
  body = "On Press A, ask the shift lead to inspect the red roller before the first long run."
  scope = "TENANT"
  status = "PENDING"
  provenance = "CHAT_DISTILLED"
  entityId = "wcr_01k00000000000000000000001"
}

import.cococo

knowledge_note "imported" {
  # TODO: add for_each — see the task above.
  kind = "ENTITY"
  # Supplied entityId: each.value.entity_id
  # TODO: assign entityId using the value above.
  # Supplied body: each.value.body
  # TODO: assign body using the value above.
  # TODO: add scope — see the task above.
  # TODO: add status — see the task above.
  # TODO: add provenance — see the task above.
}

notes.csv

code,entity_id,body
press_b,wcr_01k00000000000000000000002,"On Press B, check the feeder alignment before the first run."
press_c,wcr_01k00000000000000000000003,"On Press C, ask the shift lead to review the humidity log."

Reference: knowledge · authoring