fc-ttsreader: ship cluster-native fc-speech-align (faster-whisper) + bump web

- New ttsreader-align Deployment + Service + 5Gi PVC under
  apps/fc-ttsreader/. Wraps SYSTRAN/faster-whisper in a small FastAPI app
  exposing POST /align (fc-align contract used by Shared.Speech) AND
  POST /transcribe (audio-in feature consumed by ttsreader-web Lane G).
  Source: apps/fc-ttsreader/speech-align/ (Dockerfile + app.py +
  requirements.txt). Built locally (apt-get RUN steps need BLUEJAY-WS,
  not noc1) and ctr-imported to all 3 RKE2 nodes.
- ttsreader-web env: flip Speech__Alignment__Enabled=true and point
  BaseUrl at http://ttsreader-align.fc-ttsreader.svc.cluster.local.:9200.
  Add new TtsReader__Transcription__* env triplet pointing at the same
  service (same /transcribe endpoint).
- Bump ttsreader-web image to v202604251046 (carries the
  TranscriptionController + MCP tool + Quick.razor InputFile UI).
This commit is contained in:
Andrew Stoltz
2026-04-25 10:50:45 -05:00
parent 9df26620b8
commit df115e4d1e
4 changed files with 344 additions and 12 deletions

View File

@@ -112,6 +112,109 @@ spec:
persistentVolumeClaim:
claimName: ttsreader-piper-data
---
# fc-speech-align — cluster-native faster-whisper wrapper.
# Exposes POST /align (fc-align contract used by FlowerCore.Shared.Speech) AND
# POST /transcribe (audio-file-in feature). CPU model = base.en, int8 compute.
# Source: bluejay-infra/apps/fc-ttsreader/speech-align/ (Dockerfile + app.py).
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: ttsreader-align-models
namespace: fc-ttsreader
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 5Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: ttsreader-align
namespace: fc-ttsreader
labels:
app.kubernetes.io/name: ttsreader-align
app.kubernetes.io/part-of: flowercore
spec:
replicas: 1
strategy:
type: Recreate
selector:
matchLabels:
app.kubernetes.io/name: ttsreader-align
template:
metadata:
labels:
app.kubernetes.io/name: ttsreader-align
app.kubernetes.io/part-of: flowercore
spec:
securityContext:
fsGroup: 1654
runAsNonRoot: true
runAsUser: 1654
containers:
- name: align
image: localhost/fc-speech-align:v1
imagePullPolicy: Never
ports:
- containerPort: 9200
name: http
env:
- name: WHISPER_MODEL
value: "Systran/faster-whisper-base.en"
- name: WHISPER_DEVICE
value: "cpu"
- name: WHISPER_COMPUTE_TYPE
value: "int8"
- name: WHISPER_CACHE_DIR
value: "/models"
- name: DEFAULT_LANGUAGE
value: "en"
resources:
requests:
cpu: 250m
memory: 512Mi
limits:
cpu: 2000m
memory: 2Gi
volumeMounts:
- name: models
mountPath: /models
readinessProbe:
httpGet:
path: /health
port: 9200
initialDelaySeconds: 30
periodSeconds: 10
timeoutSeconds: 5
failureThreshold: 18
livenessProbe:
httpGet:
path: /health
port: 9200
initialDelaySeconds: 180
periodSeconds: 30
timeoutSeconds: 5
failureThreshold: 3
volumes:
- name: models
persistentVolumeClaim:
claimName: ttsreader-align-models
---
apiVersion: v1
kind: Service
metadata:
name: ttsreader-align
namespace: fc-ttsreader
spec:
selector:
app.kubernetes.io/name: ttsreader-align
ports:
- port: 9200
targetPort: 9200
name: http
---
apiVersion: apps/v1
kind: Deployment
metadata:
@@ -142,7 +245,7 @@ spec:
fsGroupChangePolicy: OnRootMismatch
containers:
- name: web
image: localhost/fc-ttsreader-web:v202604251018
image: localhost/fc-ttsreader-web:v202604251046
imagePullPolicy: Never
ports:
- containerPort: 5217
@@ -173,20 +276,24 @@ spec:
- name: TtsReader__Kokoro__TimeoutSeconds
value: "120"
- name: Speech__Alignment__Enabled
# Off until either:
# (a) a native /align backend is deployed inside the cluster, or
# (b) the BLUEJAY-WS host exposes the speaches container on the
# LAN-routable bind (10.0.56.20:9200, not just 127.0.0.1)
# AND Common ships the openai-compatible Backend support
# (currently on feat/shared-indexing, not on master).
# While disabled, /preview-with-timings still returns word timings
# via EstimatedAlignmentClient — slightly less accurate, but the
# UI can still drive word-level highlight playback.
value: "false"
# Cluster-native faster-whisper (Lane F, 2026-04-25). The
# ttsreader-align deployment in this manifest wraps
# SYSTRAN/faster-whisper with a /align endpoint matching the
# FlowerCore.Shared.Speech master contract.
value: "true"
- name: Speech__Alignment__BaseUrl
value: "http://10.0.56.20:9200"
value: "http://ttsreader-align.fc-ttsreader.svc.cluster.local.:9200"
- name: Speech__Alignment__TimeoutSeconds
value: "120"
# Cluster-native transcription endpoint shares the same pod
# (POST /transcribe). Lane G consumes this from the
# FlowerCore.TtsReader.Web AudioImport feature.
- name: TtsReader__Transcription__Enabled
value: "true"
- name: TtsReader__Transcription__BaseUrl
value: "http://ttsreader-align.fc-ttsreader.svc.cluster.local.:9200"
- name: TtsReader__Transcription__TimeoutSeconds
value: "300"
- name: TtsReader__Ollama__BaseUrl
value: "http://10.0.57.17:11434"
- name: TtsReader__Ollama__DefaultModel