SMART_KYC_OCR / app.py
gopichandra's picture
Update app.py
873e9ae verified
raw
history blame
7.45 kB
import os
os.environ["OMP_NUM_THREADS"] = "1"
import re
import datetime as dt
import gradio as gr
from utils import extract_kyc_fields
# ------------------------------------------------------------------
# πŸ” CREDENTIALS STRATEGY
# 1) Preferred: read from environment (HF Space Secrets / .env).
# 2) Fallback (NOT RECOMMENDED): hardcoded defaults below.
# If you insist, replace the REPLACE_ME values locally.
# ------------------------------------------------------------------
SF_DEFAULT_USERNAME = "[email protected]" # e.g., "[email protected]"
SF_DEFAULT_PASSWORD = "Lic@2025" # e.g., "Lic@2025"
SF_DEFAULT_TOKEN = "AmmfRcd6IiYaRtSGntBnzNMQU" # e.g., "AmmfRcd6IiYaRtSGntBnzNMQU"
SF_DEFAULT_DOMAIN = "login" # "login" (prod) or "test" (sandbox)
# ------------------------------------------------------------------
# simple-salesforce is required for SF push
try:
from simple_salesforce import Salesforce
SF_AVAILABLE = True
except Exception:
SF_AVAILABLE = False
# ---------- helpers ----------
def _parse_birthdate(dob_text: str):
"""
Normalize DOB to YYYY-MM-DD for Salesforce Date fields.
Supports dd/mm/yyyy, dd-mm-yyyy, dd.mm.yyyy, yyyy-mm-dd, or just YYYY (mapped to mid-year).
"""
if not dob_text or dob_text == "Not found":
return None
s = dob_text.strip()
m = re.fullmatch(r"(\d{4})-(\d{2})-(\d{2})", s)
if m:
y, mo, d = map(int, m.groups())
try:
return dt.date(y, mo, d).isoformat()
except ValueError:
return None
m = re.fullmatch(r"(\d{2})[./-](\d{2})[./-](\d{4})", s)
if m:
d, mo, y = map(int, m.groups())
try:
return dt.date(y, mo, d).isoformat()
except ValueError:
return None
m = re.fullmatch(r"(19|20)\d{2}", s)
if m:
y = int(s)
try:
return dt.date(y, 6, 15).isoformat()
except ValueError:
return None
return None
def sf_connect_from_env_or_defaults():
"""
Tries env vars first (HF Space Secrets / .env), then hardcoded defaults.
Env vars: SF_USERNAME, SF_PASSWORD, SF_SECURITY_TOKEN, SF_DOMAIN
"""
if not SF_AVAILABLE:
raise RuntimeError("simple-salesforce is not installed.")
username = os.getenv("SF_USERNAME", SF_DEFAULT_USERNAME)
password = os.getenv("SF_PASSWORD", SF_DEFAULT_PASSWORD)
token = os.getenv("SF_SECURITY_TOKEN", SF_DEFAULT_TOKEN)
domain = os.getenv("SF_DOMAIN", SF_DEFAULT_DOMAIN or "login")
# Basic validation: block if still placeholders
if any(v.startswith("REPLACE_ME") for v in [username, password, token]):
raise ValueError(
"Salesforce credentials missing. "
"Either set Space Secrets/ENV (SF_USERNAME, SF_PASSWORD, SF_SECURITY_TOKEN, SF_DOMAIN) "
"or replace the REPLACE_ME_* constants in app.py."
)
return Salesforce(username=username, password=password, security_token=token, domain=domain)
def sf_push_kyc_record(sf, ocr_results):
"""
Create one KYC_Record__c combining Aadhaar + PAN.
Fields:
Aadhaar_Number__c, Aadhaar_Name__c, Aadhaar_DOB__c (Date)
PAN_Number__c, Pan_Name__c, Pan_DOB__c (Date)
"""
a = ocr_results.get("aadhaar") or {}
p = ocr_results.get("pan") or {}
aadhaar_number = a.get("aadhaar_number") if (a.get("card_type") == "AADHAAR") else None
aadhaar_name = a.get("name") if (a.get("card_type") == "AADHAAR") else None
aadhaar_dob = _parse_birthdate(a.get("dob")) if (a.get("card_type") == "AADHAAR") else None
pan_number = p.get("pan_number") if (p.get("card_type") == "PAN") else None
pan_name = p.get("name") if (p.get("card_type") == "PAN") else None
pan_dob = _parse_birthdate(p.get("dob")) if (p.get("card_type") == "PAN") else None
payload = {
"Aadhaar_Number__c": aadhaar_number,
"Aadhaar_Name__c": aadhaar_name,
"Aadhaar_DOB__c": aadhaar_dob,
"PAN_Number__c": pan_number,
"Pan_Name__c": pan_name,
"Pan_DOB__c": pan_dob,
}
payload = {k: v for k, v in payload.items() if v is not None}
try:
result = sf.KYC_Record__c.create(payload)
return {"success": True, "id": result.get("id"), "payload": payload}
except Exception as e:
return {"success": False, "error": str(e), "payload": payload}
# ---------- gradio callback ----------
def process_documents(aadhaar_file, pan_file, push_to_sf):
"""
- Runs OCR on Aadhaar and PAN separately.
- Optionally pushes a single KYC_Record__c to Salesforce.
"""
results = {"aadhaar": None, "pan": None}
if not aadhaar_file and not pan_file:
return {"error": "Please upload at least one file (Aadhaar and/or PAN)."}
# OCR Aadhaar
if aadhaar_file:
try:
res = extract_kyc_fields(aadhaar_file.name)
res["source_file"] = os.path.basename(aadhaar_file.name)
results["aadhaar"] = res
except Exception as e:
results["aadhaar"] = {"error": f"Aadhaar OCR failed: {str(e)}", "card_type": "UNKNOWN"}
# OCR PAN
if pan_file:
try:
res = extract_kyc_fields(pan_file.name)
res["source_file"] = os.path.basename(pan_file.name)
results["pan"] = res
except Exception as e:
results["pan"] = {"error": f"PAN OCR failed: {str(e)}", "card_type": "UNKNOWN"}
output = {"ocr": results}
if push_to_sf:
try:
sf = sf_connect_from_env_or_defaults()
created = sf_push_kyc_record(sf, results)
output["salesforce"] = {"pushed": created.get("success", False), **created}
except Exception as e:
output["salesforce"] = {"pushed": False, "error": str(e)}
return output
# ---------- UI ----------
with gr.Blocks(title="Smart KYC OCR β†’ Salesforce (KYC_Record__c)") as demo:
gr.Markdown(
"""
# 🧾 Smart KYC OCR β†’ Salesforce
Upload **Aadhaar** and **PAN** in separate boxes, then (optional) push a single **KYC_Record__c**.
- Preferred: configure `SF_USERNAME`, `SF_PASSWORD`, `SF_SECURITY_TOKEN`, `SF_DOMAIN` as environment variables
(Hugging Face **Space Secrets**).
- Emergency fallback: edit the `SF_DEFAULT_*` constants in `app.py` (NOT recommended).
"""
)
with gr.Row():
with gr.Column(scale=1):
aadhaar_uploader = gr.File(
label="πŸ“€ Aadhaar Upload",
file_types=[".jpg", ".jpeg", ".png"]
)
with gr.Column(scale=1):
pan_uploader = gr.File(
label="πŸ“€ PAN Upload",
file_types=[".jpg", ".jpeg", ".png"]
)
push_to_sf = gr.Checkbox(label="Push to Salesforce (create KYC_Record__c)", value=False)
submit_btn = gr.Button("πŸ” Extract KYC Info", variant="primary")
output_json = gr.JSON(label="πŸ“‹ Output (OCR + Salesforce)")
submit_btn.click(
fn=process_documents,
inputs=[aadhaar_uploader, pan_uploader, push_to_sf],
outputs=output_json,
)
gr.Markdown("---")
gr.Markdown(
"""
πŸ”’ **Security note:** Avoid hardcoding secrets. Prefer Space Secrets / environment variables.
"""
)
if __name__ == "__main__":
demo.launch()