Upload 4 files

.gitignore

@@ -0,0 +1,7 @@
+.venv/
+__pycache__/
+*.pyc
+.env
+*.log
+data/
+.cache/

Dockerfile

@@ -0,0 +1,21 @@
+FROM python:3.12-slim
+
+WORKDIR /app
+
+ENV PYTHONUNBUFFERED=1 \
+    PIP_NO_CACHE_DIR=1 \
+    HF_HOME=/data/.cache/huggingface \
+    ICMR_DATA_DIR=/data \
+    PORT=7860
+
+# The bootstrapper uses Hugging Face Hub to download the data/code.
+RUN pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir huggingface_hub
+
+RUN mkdir -p /data
+
+COPY app.py /app/app.py
+
+EXPOSE 7860
+
+CMD ["python", "/app/app.py"]

app.py

@@ -0,0 +1,182 @@
+#!/usr/bin/env python3
+"""
+Hugging Face Docker Space bootstrapper for the ICMR + HITEK demo API.
+
+This is the Python equivalent of the supplied Linux/VPS setup script:
+- Downloads the parquet data files from the source dataset repo.
+- Downloads the prebuilt index parts from the full-db repo.
+- Downloads main.py, build_index.py and requirements.txt.
+- Installs the downloaded Python requirements.
+- Starts main:app with Uvicorn on the Hugging Face $PORT.
+
+Environment variables:
+    DATA_REPO       Default: Kzr0xx/Icmr-and-hitek
+    FULL_REPO       Default: Kzr0xx/icrm-hitek-full-db-mixed
+    HF_TOKEN        Optional token for private/gated repositories.
+    ICMR_DATA_DIR   Default: /data
+    PORT            Default: 7860
+    SKIP_DOWNLOAD   Set to 1/true/yes to skip downloads.
+"""
+
+import os
+import sys
+import subprocess
+from pathlib import Path
+
+from huggingface_hub import hf_hub_download
+
+
+DATA_REPO = os.getenv("DATA_REPO", "Kzr0xx/Icmr-and-hitek")
+FULL_REPO = os.getenv("FULL_REPO", "Kzr0xx/icrm-hitek-full-db-mixed")
+DATA_DIR = Path(os.getenv("ICMR_DATA_DIR", "/data"))
+PORT = int(os.getenv("PORT", "7860"))
+
+FILES = [
+    "part1.parquet",
+    "part2a.parquet",
+    "part2b_new.parquet",
+]
+
+IDX = [
+    "idx_phone.0.parquet",
+    "idx_phone.1.parquet",
+    "idx_phone.2.parquet",
+    "idx_phone.3.parquet",
+    "idx_phone.4.parquet",
+    "idx_phone.5.parquet",
+    "idx_phone.6.parquet",
+    "idx_aadhar.0.parquet",
+    "idx_aadhar.1.parquet",
+    "idx_aadhar.2.parquet",
+    "idx_aadhar.3.parquet",
+    "idx_aadhar.4.parquet",
+    "idx_aadhar.5.parquet",
+    "idx_aadhar.6.parquet",
+]
+
+CODE = [
+    "main.py",
+    "build_index.py",
+    "requirements.txt",
+]
+
+
+def download(repo: str, filename: str, destination: Path):
+    """Download a file from a Hugging Face dataset repository."""
+    destination.parent.mkdir(parents=True, exist_ok=True)
+
+    if destination.exists() and destination.stat().st_size > 0:
+        print(f"[SKIP] {destination} already exists", flush=True)
+        return
+
+    print(f"[DOWNLOAD] {repo}/{filename}", flush=True)
+
+    token = os.getenv("HF_TOKEN") or None
+
+    cached = hf_hub_download(
+        repo_id=repo,
+        filename=filename,
+        repo_type="dataset",
+        token=token,
+        local_dir=str(destination.parent),
+        local_dir_use_symlinks=False,
+    )
+
+    cached_path = Path(cached)
+
+    # hf_hub_download normally places it directly in local_dir.
+    # This fallback handles cases where the returned path differs.
+    if cached_path.resolve() != destination.resolve():
+        destination.parent.mkdir(parents=True, exist_ok=True)
+        if not destination.exists():
+            cached_path.replace(destination)
+
+    print(f"[OK] {destination}", flush=True)
+
+
+def install_requirements(requirements_file: Path):
+    print("[SETUP] Installing API requirements...", flush=True)
+
+    subprocess.check_call([
+        sys.executable,
+        "-m",
+        "pip",
+        "install",
+        "--no-cache-dir",
+        "-r",
+        str(requirements_file),
+    ])
+
+
+def start_api():
+    print("=" * 60, flush=True)
+    print(" ICMR + HITEK Hugging Face Docker Space", flush=True)
+    print("=" * 60, flush=True)
+    print(f"Data directory: {DATA_DIR}", flush=True)
+    print(f"Port: {PORT}", flush=True)
+
+    env = os.environ.copy()
+    env["ICMR_DATA_DIR"] = str(DATA_DIR)
+
+    print(f"Starting API on 0.0.0.0:{PORT}", flush=True)
+
+    os.execvpe(
+        sys.executable,
+        [
+            sys.executable,
+            "-m",
+            "uvicorn",
+            "main:app",
+            "--host",
+            "0.0.0.0",
+            "--port",
+            str(PORT),
+            "--workers",
+            "1",
+        ],
+        env,
+    )
+
+
+def main():
+    DATA_DIR.mkdir(parents=True, exist_ok=True)
+
+    skip = os.getenv("SKIP_DOWNLOAD", "").lower() in {
+        "1", "true", "yes", "on"
+    }
+
+    if not skip:
+        print("[1/3] Downloading data files...", flush=True)
+        for filename in FILES:
+            download(DATA_REPO, filename, DATA_DIR / filename)
+
+        print("[2/3] Downloading index parts...", flush=True)
+        for filename in IDX:
+            download(FULL_REPO, filename, DATA_DIR / filename)
+
+        print("[3/3] Downloading API code...", flush=True)
+        for filename in CODE:
+            download(FULL_REPO, filename, Path("/app") / filename)
+    else:
+        print("[SKIP_DOWNLOAD] Using files already present in /app and /data", flush=True)
+
+    requirements = Path("/app/requirements.txt")
+
+    if requirements.exists():
+        install_requirements(requirements)
+    else:
+        raise FileNotFoundError(
+            "requirements.txt was not downloaded. "
+            "Check FULL_REPO/HF_TOKEN and repository contents."
+        )
+
+    if not Path("/app/main.py").exists():
+        raise FileNotFoundError(
+            "/app/main.py was not downloaded from the configured FULL_REPO."
+        )
+
+    start_api()
+
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -0,0 +1,4 @@
+# Bootstrap dependency.
+# app.py downloads the actual API requirements.txt from FULL_REPO
+# and installs it at Space startup.
+huggingface_hub>=0.30.0