skaffold.yaml

# ── KP (Kamerplanter Core) ──────────────────────────────────────────
# Backend, Frontend, Celery, ArangoDB, TimescaleDB, Valkey
# Usage (core only):  skaffold dev -m kp
#
# FULL STACK (one command): skaffold dev        # no -m → runs kp + ki + recognition
#   equivalent to:          skaffold dev -m kp -m ki -m recognition
# Image recognition (REQ-029-A) AND pest detection (REQ-044) BOTH need the
# `recognition` module (the kamerplanter-recognition inference-service), which in
# turn needs the `ki` module's pgvector store. Running only `-m kp` leaves the
# inference-service undeployed → both admin cards show "service unreachable".
apiVersion: skaffold/v4beta13
kind: Config
metadata:
  name: kp

build:
  local:
    push: false
    useBuildkit: true
  artifacts:
    - image: kamerplanter-backend
      context: src/backend
      docker:
        dockerfile: Dockerfile.dev
      sync:
        manual:
          - src: "app/**/*.py"
            dest: /app
    - image: kamerplanter-frontend
      context: src/frontend
      docker:
        dockerfile: Dockerfile.dev
      sync:
        manual:
          - src: "src/**/*.ts"
            dest: /app
          - src: "src/**/*.tsx"
            dest: /app
          - src: "src/**/*.css"
            dest: /app

deploy:
  statusCheckDeadlineSeconds: 1800
  helm:
    releases:
      - name: kamerplanter
        chartPath: helm/kamerplanter
        valuesFiles:
          - helm/kamerplanter/values-dev.yaml
        namespace: default
        setValueTemplates:
          controllers.backend.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_backend}}"
          controllers.backend.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_backend}}"
          controllers.frontend.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_frontend}}"
          controllers.frontend.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_frontend}}"
          controllers.celery-worker.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_backend}}"
          controllers.celery-worker.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_backend}}"
          controllers.celery-beat.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_backend}}"
          controllers.celery-beat.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_backend}}"
          valkey.dataStorage.enabled: true
          valkey.dataStorage.size: 1Gi
          valkey.dataStorage.keepPvc: true

portForward:
  - resourceType: service
    resourceName: kamerplanter-backend
    namespace: default
    port: 8000
    localPort: 8000
  - resourceType: service
    resourceName: kamerplanter-frontend
    namespace: default
    port: 5173
    localPort: 3000
  - resourceType: service
    resourceName: kamerplanter-arangodb
    namespace: default
    port: 8529
    localPort: 8529
  - resourceType: service
    resourceName: kamerplanter-timescaledb
    namespace: default
    port: 5432
    localPort: 5432

profiles:
  - name: backend-only
    patches:
      # Remove frontend build artifact (index 1)
      - op: remove
        path: /build/artifacts/1
      # Remove frontend port forward (index 1: backend=0, frontend=1)
      - op: remove
        path: /portForward/1

  - name: frontend-only
    patches:
      # Remove backend build artifact (index 0)
      - op: remove
        path: /build/artifacts/0
      # Remove backend port forward (index 0: backend=0)
      - op: remove
        path: /portForward/0

  - name: debug
    patches:
      - op: add
        path: /build/artifacts/0/docker/buildArgs
        value:
          DEBUGPY_ENABLED: "true"
    activation:
      - command: debug
---
# ── KI (Knowledge/AI Stack) ─────────────────────────────────────────
# Knowledge-Service, Embedding-Service, VectorDB
# Usage: skaffold dev -m ki
# Services: kamerplanter-ki-knowledge-service, kamerplanter-ki-vectordb, kamerplanter-ki-embedding-service
apiVersion: skaffold/v4beta13
kind: Config
metadata:
  name: ki

build:
  local:
    push: false
    useBuildkit: true
  artifacts:
    - image: kamerplanter-vectordb
      context: docker/vectordb
      docker:
        dockerfile: Dockerfile
    - image: kamerplanter-knowledge
      context: .
      docker:
        dockerfile: docker/knowledge/Dockerfile
    - image: kamerplanter-knowledge-service
      context: src/knowledge-service
      docker:
        dockerfile: Dockerfile.dev
      sync:
        manual:
          - src: "app/**/*.py"
            dest: /app
    - image: kamerplanter-embedding-service
      context: docker/embedding-service
      docker:
        dockerfile: Dockerfile.dev
    - image: kamerplanter-reranker-service
      context: docker/reranker-service
      docker:
        dockerfile: Dockerfile.dev

deploy:
  statusCheckDeadlineSeconds: 1800
  helm:
    releases:
      - name: kamerplanter-ki
        chartPath: helm/kamerplanter
        valuesFiles:
          - helm/kamerplanter/values-dev-ki.yaml
        namespace: default
        setValueTemplates:
          controllers.vectordb.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_vectordb}}"
          controllers.vectordb.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_vectordb}}"
          controllers.knowledge-service.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_knowledge_service}}"
          controllers.knowledge-service.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_knowledge_service}}"
          controllers.knowledge-service.initContainers.knowledge.image.repository: "{{.IMAGE_REPO_kamerplanter_knowledge}}"
          controllers.knowledge-service.initContainers.knowledge.image.tag: "{{.IMAGE_TAG_kamerplanter_knowledge}}"
          controllers.embedding-service.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_embedding_service}}"
          controllers.embedding-service.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_embedding_service}}"
          controllers.reranker-service.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_reranker_service}}"
          controllers.reranker-service.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_reranker_service}}"

portForward:
  - resourceType: service
    resourceName: kamerplanter-ki-vectordb
    namespace: default
    port: 5432
    localPort: 5433
  - resourceType: service
    resourceName: kamerplanter-ki-knowledge-service
    namespace: default
    port: 8000
    localPort: 8090
  - resourceType: service
    resourceName: kamerplanter-ki-embedding-service
    namespace: default
    port: 8080
    localPort: 8080
  - resourceType: service
    resourceName: kamerplanter-ki-reranker-service
    namespace: default
    port: 8081
    localPort: 8081
---
# ── Recognition (DINOv2 inference-service) ──────────────────────────
# Inference-service only — serves BOTH plant identification (REQ-029-A, /match)
# and pest detection (REQ-044, /pest/*). Shares the KI stack's pgvector store
# (kamerplanter-ki-vectordb), so run alongside `-m ki` (or at least its
# vectordb). Omit this module to run without any image recognition; then the
# plant- and pest-recognition admin cards report "service unreachable".
# Usage: skaffold dev -m ki -m recognition   (or just `skaffold dev` for all)
apiVersion: skaffold/v4beta13
kind: Config
metadata:
  name: recognition

build:
  local:
    push: false
    useBuildkit: true
  artifacts:
    - image: kamerplanter-inference-service
      context: src/inference-service
      docker:
        dockerfile: Dockerfile.dev
      sync:
        manual:
          - src: "app/**/*.py"
            dest: /app

deploy:
  statusCheckDeadlineSeconds: 1800
  helm:
    releases:
      - name: kamerplanter-recognition
        chartPath: helm/kamerplanter
        valuesFiles:
          - helm/kamerplanter/values-dev-recognition.yaml
        namespace: default
        setValueTemplates:
          controllers.inference-service.containers.main.image.repository: "{{.IMAGE_REPO_kamerplanter_inference_service}}"
          controllers.inference-service.containers.main.image.tag: "{{.IMAGE_TAG_kamerplanter_inference_service}}"

portForward:
  - resourceType: service
    resourceName: kamerplanter-recognition
    namespace: default
    port: 8000
    localPort: 8091