feat(ml): include top3 predictions in /classify-image response

feat(ml): integrate new 120-class dog breed classifier model (ViT-Large) via HF Hub download

Update app.py

fix: formatar nome de raca no /classify-image (remover prefixo Stanford Dogs)

Update app.py

Improve breed label resolution via id2label with name normalization

Fix undefined Path in vision loader

Update vision model to load fine-tuned ViT breed classifier from HF Hub

feat(ml): support HEAD requests on root and health routes to fix 405 errors

feat(ml): update root route to return interactive status dashboard

feat(ml): return 200 OK HTML on root for health check compatibility

feat(ml): redirect root route to /docs to fix Hugging Face Space starting overlay

chore(ml): add Hugging Face YAML metadata to README.md

feat(api): add POST /classify-image endpoint + tests + docs (v1.4.0)

feat(ml): add ViT dual-head species/breed classifier training script

Merge pull request #3 from firstoff23/railway/fix-deploy-cb3023

config: set N8N_WEBHOOK_URL for production and align ports in fly.toml

feat: adiciona skeleton global e testes e2e

repo: remover backups and contexts de debug do controle de versao

critico: remover fallback aleatorio e proteger rotas sensiveis

critico: remover fallback aleatorio e proteger rotas sensiveis

fix: update tests and requirements to support tensorflow-cpu and secure supabase rls

fix: keep 1 machine always running to avoid 50s cold-start timeout

backend: update Dockerfile port to 8000

backend: add missing dependencies to requirements.txt

backend: add fly.toml configuration file

feat(task2): add /detect-species endpoint (google/vit-base-patch16-224) and auto-species badge in RecordingPage

feat(health, i18n, vision): implement health records bulletin, PT/EN multi-language locales, and YOLOv8 camera frame posture detection

feat: implement photo breed identification via transformers pipeline and update frontend to pass animal_type

fix: use huggingface_hub InferenceClient instead of httpx for HF Space compatibility

feat: add POST /identify-breed using HuggingFace Inference API (ViT dog/cat breed models)

feat: prepara backend ML para Hugging Face Spaces

fix(railway): add nixpacks.toml with ffmpeg — more robust than railway.toml phases

fix(railway): add ffmpeg and libsndfile1 to nixpacks build

fix: correct indentation of classify route DB/Redis block

fix: fix shutdown if db_pool indentation (4 spaces not 8)

Refactor database and Redis connefix: correct indentation of classify route DB/Redis blockction checks

Refactor classification result handlifix: fix indentation of DB/Redis block in classify routeng and storage

fix: correct indentation in classify route after DB/Redis integration

Implement DB and Redis connectiofeat: integrate PostgreSQL persistence and Redis caching in /classifyns on startup

feat: add asyncpg, psycopg2-binary and redis dependencies

Add health check endpoint and root status endpoint

chore: add railway.toml for FastAPI backend deploy

fix: usa runtime Python 3.11 no Railway

chore: add Codex permanent instructions

chore: add Codex permanent instructions

fix: atualiza runtime Python do backend

fix: use Python 3.11 minor version in ml_backend/runtime.txt

feat: adiciona runtime Python ao backend FastAPI

feat: prepara backend FastAPI para deploy