scripts/deploy_hf.sh

#!/usr/bin/env bash
# ──────────────────────────────────────────────────────────────────────────
# scripts/deploy_hf.sh — fully-automated deploy to Hugging Face Spaces
# ──────────────────────────────────────────────────────────────────────────
#
# Usage:
#   HF_TOKEN="<your-hf-write-token>" ./scripts/deploy_hf.sh <hf-user>/<space-name>
#
# Example (token redacted — get yours at https://huggingface.co/settings/tokens):
#   HF_TOKEN="$HF_TOKEN" ./scripts/deploy_hf.sh W1nd5pac/microclimate-x
#
# What it does (no manual steps):
#   1. Ensures huggingface_hub CLI is installed in .venv/
#   2. Authenticates with HF_TOKEN
#   3. Creates the Space (Docker SDK) if it doesn't exist yet
#   4. Uploads the whole repo (server-side LFS handles the 217 MB model)
#   5. Prints the live URL when the build is queued
#
# Skips:
#   data/   figures/   tests/   .venv/   .git/   *.sqlite3   __pycache__/
# ──────────────────────────────────────────────────────────────────────────
set -euo pipefail

if [[ $# -lt 1 ]]; then
  echo "Usage: HF_TOKEN=hf_xxx $0 <hf-user>/<space-name>"
  echo "Example: HF_TOKEN=hf_xxx $0 W1nd5pac/microclimate-x"
  exit 2
fi

REPO_ID="$1"
ROOT="$(cd "$(dirname "$0")/.." && pwd)"
cd "$ROOT"

# Clean env so other venvs / PYTHONPATH leaks don't break us.
unset PYTHONPATH VIRTUAL_ENV PYTHONHOME

# ── 1. ensure .venv has huggingface_hub ──────────────────────────────────
if [[ ! -x ".venv/bin/hf" ]]; then
  echo "▶ Installing huggingface_hub CLI into .venv/ …"
  .venv/bin/pip install -q -U "huggingface_hub[cli,hf_transfer]"
fi
HF=".venv/bin/hf"

# Speed-boost for the 217 MB model upload.
export HF_HUB_ENABLE_HF_TRANSFER=1

# ── 2. authenticate ──────────────────────────────────────────────────────
if [[ -z "${HF_TOKEN:-}" ]]; then
  if ! $HF auth whoami >/dev/null 2>&1; then
    echo "❌ HF_TOKEN env not set and not already logged in."
    echo "   Get a Write token at https://huggingface.co/settings/tokens and run:"
    echo "     HF_TOKEN=hf_xxx $0 $REPO_ID"
    exit 1
  fi
fi

if [[ -n "${HF_TOKEN:-}" ]]; then
  # Re-login non-interactively so we use the supplied token (idempotent).
  echo "▶ Authenticating as the token's owner …"
  echo "$HF_TOKEN" | $HF auth login --token "$HF_TOKEN" --add-to-git-credential >/dev/null 2>&1 || true
fi

WHOAMI=$($HF auth whoami 2>/dev/null | head -1 || echo "?")
echo "  Logged in as: $WHOAMI"

# ── 3. create the Space if missing (idempotent — 409 means "exists") ─────
echo "▶ Ensuring Space $REPO_ID exists (Docker SDK) …"
CREATE_OUTPUT=$($HF repos create "$REPO_ID" --repo-type space --space-sdk docker 2>&1 || true)
if echo "$CREATE_OUTPUT" | grep -q "Successfully created"; then
  echo "  Created fresh Space."
elif echo "$CREATE_OUTPUT" | grep -qi "already created\|409"; then
  echo "  Space already exists — will push to it."
else
  echo "$CREATE_OUTPUT"
  echo "❌ Unexpected response from 'hf repos create'. Aborting."
  exit 1
fi

# ── 4. sanity-check the model exists locally ─────────────────────────────
MODEL="models/rf_model.pkl"
if [[ ! -f "$MODEL" ]]; then
  echo "⚠️  $MODEL not found — the Space will fall back to a heuristic predictor."
  read -r -p "Continue without the trained model? [y/N] " ans
  [[ "$ans" =~ ^[Yy]$ ]] || exit 1
fi

# ── 5. upload everything ─────────────────────────────────────────────────
echo "▶ Uploading repo → spaces/$REPO_ID …"
echo "  (217 MB rf_model.pkl uses HF's server-side Xet/LFS — no local LFS needed)"

DEPLOY_MSG="Deploy $(date -u +%Y-%m-%dT%H:%M:%SZ) — $(git rev-parse --short HEAD 2>/dev/null || echo local)"

# Pass 1: bulk-upload everything except the model (default: respects .gitignore
# which already excludes *.pkl, so the big file won't go in this pass).
$HF upload \
  "$REPO_ID" \
  . \
  . \
  --repo-type=space \
  --commit-message="$DEPLOY_MSG (code)" \
  --exclude "data/*" \
  --exclude "figures/*" \
  --exclude "tests/*" \
  --exclude ".venv/*" \
  --exclude ".local/*" \
  --exclude ".pytest_cache/*" \
  --exclude ".ruff_cache/*" \
  --exclude ".mypy_cache/*" \
  --exclude "**/__pycache__/*" \
  --exclude "*.sqlite3" \
  --exclude "*.sqlite3-*" \
  --exclude "*.pyc" \
  --exclude ".DS_Store" \
  --exclude ".git/*" \
  --exclude ".github/*"

# Pass 2: explicitly push the 217 MB Random Forest model. An explicit
# single-file path bypasses .gitignore filtering — without this step the Space
# falls back to the heuristic predictor and the AUC=0.871 claim won't reproduce.
if [[ -f "$MODEL" ]]; then
  echo "▶ Uploading models/rf_model.pkl (217 MB) — bypassing .gitignore …"
  $HF upload \
    "$REPO_ID" \
    "$MODEL" \
    "models/rf_model.pkl" \
    --repo-type=space \
    --commit-message="$DEPLOY_MSG (model)"
fi

echo
echo "✅ Upload complete. Space is rebuilding now."
echo "   Status:    https://huggingface.co/spaces/$REPO_ID"
echo "   Live URL:  https://huggingface.co/spaces/$REPO_ID  (≈ 3-5 min for first build)"
echo
echo "Tip: once the green Running badge shows, send the Live URL to your supervisor."