biz/gemini.py

import os, json
from typing import Dict
from google import genai
from google.genai import types

SYSTEM_INSTRUCTIONS = """
You read a cropped Japanese business card image and return strict JSON fields.

Return ONLY this JSON:
{
  "person_kanji": "...", "person_kana": "...", "person_romaji": "...",
  "title": "...", "department": "...", "company": "...",
  "email": ["..."], "phone": ["..."],
  "website": "...", "address_jp": "...", "notes": "..."
}
Use null / [] for unknowns. Romaji: family-name LAST. No extra text.
"""

def make_client():
    use_vertex = os.getenv("GOOGLE_GENAI_USE_VERTEXAI", "false").lower() in ("1","true","yes")
    if use_vertex:
        project = os.getenv("GOOGLE_CLOUD_PROJECT")
        location = os.getenv("GOOGLE_CLOUD_LOCATION", "us-central1")
        if not project:
            raise RuntimeError("Set GOOGLE_CLOUD_PROJECT or use GOOGLE_API_KEY.")
        return genai.Client(vertexai=True, project=project, location=location)
    api_key = os.getenv("GOOGLE_API_KEY")
    if not api_key:
        raise RuntimeError("Missing GOOGLE_API_KEY.")
    return genai.Client(api_key=api_key)

def extract_from_crop(image_bytes: bytes, source_name: str) -> Dict:
    client = make_client()
    img_part = types.Part.from_bytes(data=image_bytes, mime_type="image/png")
    resp = client.models.generate_content(
        model="gemini-2.5-flash",
        contents=[types.Content(role="user", parts=[
            types.Part.from_text(SYSTEM_INSTRUCTIONS), img_part
        ])],
        config=types.GenerateContentConfig(
            temperature=0.2, response_mime_type="application/json"
        )
    )
    try:
        data = json.loads(resp.text)
    except Exception:
        data = {}
    data["source_name"] = source_name
    return data