SMART_KYC_OCR / app.py
gopichandra's picture
Update app.py
89e89ee verified
raw
history blame
9.91 kB
import os
os.environ["OMP_NUM_THREADS"] = "1"
import re
import json
import datetime as dt
import gradio as gr
from utils import extract_kyc_fields
# ------------------ HARD-CODED SALESFORCE CREDS (as requested) ------------------
SF_USERNAME = "[email protected]"
SF_PASSWORD = "Lic@2025"
SF_SECURITY_TOKEN = "AmmfRcd6IiYaRtSGntBnzNMQU"
SF_DOMAIN = "login" # "login" for prod, "test" for sandbox
# -------------------------------------------------------------------------------
# simple-salesforce + exceptions
try:
from simple_salesforce import Salesforce
from simple_salesforce.exceptions import (
SalesforceAuthenticationFailed,
SalesforceGeneralError,
SalesforceMalformedRequest,
SalesforceExpiredSession,
SalesforceRefusedRequest,
SalesforceMoreThanOneRecord,
SalesforceResourceNotFound
)
SF_AVAILABLE = True
except Exception:
SF_AVAILABLE = False
# ---------- helpers ----------
def _parse_birthdate(dob_text: str):
"""
Normalize DOB to YYYY-MM-DD (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 _fmt_sf_error(err: Exception):
"""Produce a clear, JSON-safe dict from any simple-salesforce error."""
info = {"type": err.__class__.__name__, "message": str(err)}
if isinstance(err, SalesforceAuthenticationFailed):
content = getattr(err, "content", None) or getattr(err, "response", None)
info.update({
"category": "AUTHENTICATION",
"status": getattr(err, "status", None),
"content": getattr(content, "content", None) if hasattr(content, "content") else content,
})
elif isinstance(err, (SalesforceMalformedRequest, SalesforceGeneralError, SalesforceRefusedRequest)):
info.update({
"category": "REQUEST",
"status": getattr(err, "status", None),
"resource": getattr(err, "resource_name", None),
"url": getattr(err, "url", None),
"content": getattr(err, "content", None),
})
elif isinstance(err, SalesforceResourceNotFound):
info.update({"category": "NOT_FOUND"})
elif isinstance(err, SalesforceExpiredSession):
info.update({"category": "AUTH_EXPIRED"})
return info
def sf_connect():
"""
Connect to Salesforce using the hardcoded credentials.
Also runs a quick auth query to surface any auth issues clearly.
"""
if not SF_AVAILABLE:
raise RuntimeError("simple-salesforce is not installed. Add `simple-salesforce` to requirements.txt.")
sf = Salesforce(
username=SF_USERNAME,
password=SF_PASSWORD,
security_token=SF_SECURITY_TOKEN,
domain=SF_DOMAIN,
)
# quick auth sanity check (will raise on bad auth)
sf.query("SELECT Id FROM User LIMIT 1")
return sf
def _raw_create_with_fallback(sf, object_api_name: str, payload: dict):
"""
Fallback path that bypasses simple-salesforce's error wrappers and calls REST directly.
Always returns a dict:
- on success: {"success": True, "id": "...", "url": "...", "status_code": 201}
- on error: {"success": False, "status_code": <int>, "url": "...", "response_json": <json or None>, "response_text": <str>}
"""
url = f"{sf.base_url}sobjects/{object_api_name}"
try:
resp = sf.session.post(url, json=payload, headers=sf.headers, timeout=30)
status = resp.status_code
try:
body = resp.json()
except Exception:
body = None
if 200 <= status < 300:
# Salesforce returns {"id": "...", "success": true, "errors": []}
rec_id = (body or {}).get("id") if isinstance(body, dict) else None
return {"success": True, "id": rec_id, "status_code": status, "url": url, "response_json": body}
else:
# Return raw details so you see the exact field/object error
return {
"success": False,
"status_code": status,
"url": url,
"response_json": body,
"response_text": resp.text,
}
except Exception as e:
# Network or session issues
return {"success": False, "url": url, "exception": _fmt_sf_error(e)}
def sf_push_kyc_record(sf, ocr_results):
"""
Create one KYC_Record__c combining Aadhaar + PAN.
Salesforce custom object 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, # Date field in SF
"PAN_Number__c": pan_number,
"Pan_Name__c": pan_name,
"Pan_DOB__c": pan_dob, # Date field in SF
}
# Remove None keys to avoid nulling non-nullable fields
payload = {k: v for k, v in payload.items() if v is not None}
# First try the nice SDK method
try:
result = sf.KYC_Record__c.create(payload)
return {"success": True, "id": result.get("id"), "payload": payload, "via": "sdk"}
except Exception as e:
# If simple-salesforce throws its unhelpful TypeError, fall back to a raw REST POST
raw = _raw_create_with_fallback(sf, "KYC_Record__c", payload)
if raw.get("success"):
return {"success": True, "id": raw.get("id"), "payload": payload, "via": "raw", "raw": raw}
else:
return {"success": False, "error": _fmt_sf_error(e), "payload": payload, "raw": raw}
# ---------- 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 with robust fallback.
"""
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()
created = sf_push_kyc_record(sf, results)
output["salesforce"] = {"pushed": created.get("success", False), **created}
except Exception as e:
# Even connection/auth errors will be formatted
output["salesforce"] = {"pushed": False, "error": _fmt_sf_error(e)}
return output
# ---------- UI ----------
with gr.Blocks(title="Smart KYC OCR → Salesforce (KYC_Record__c)") as demo:
gr.Markdown(
"""
# 🧾 Smart KYC OCR → Salesforce
Upload **Aadhaar** and **PAN** in separate boxes, then (optional) push one **KYC_Record__c**.
"""
)
with gr.Row():
with gr.Column(scale=1):
aadhaar_uploader = gr.File(
label="📤 Aadhaar Upload",
file_types=[".jpg", ".jpeg", ".png"]
)
with gr.Column(scale=1):
pan_uploader = gr.File(
label="📤 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("🔍 Extract KYC Info", variant="primary")
output_json = gr.JSON(label="📋 Output (OCR + Salesforce)")
submit_btn.click(
fn=process_documents,
inputs=[aadhaar_uploader, pan_uploader, push_to_sf],
outputs=output_json,
)
gr.Markdown("---")
gr.Markdown(
"""
If an error occurs, you'll now see **status_code**, **url**, and Salesforce’s **response_json** for fast debugging.
"""
)
# Important for Spaces: keep `demo` at module level
if __name__ == "__main__":
demo.launch()