import os
import re
from datetime import datetime
from typing import Optional, Dict, Any
from simple_salesforce import Salesforce
from paddleocr import PaddleOCR

# ------------------------------
# OCR setup (stable on small CPU)
# ------------------------------
os.environ.setdefault("OMP_NUM_THREADS", "1")
_ocr = PaddleOCR(use_angle_cls=True, lang='en')

# Regexes
PAN_REGEX = r'\b[A-Z]{5}[0-9]{4}[A-Z]\b'
AADHAAR_REGEX = r'\b\d{4}[\s-]?\d{4}[\s-]?\d{4}\b'
DOB_REGEXES = [
    r'\b\d{2}[./-]\d{2}[./-]\d{4}\b',
    r'\b\d{4}-\d{2}-\d{2}\b',
    r'\b\d{2}[./-](JAN|FEB|MAR|APR|MAY|JUN|JUL|AUG|SEP|OCT|NOV|DEC)[./-]\d{4}\b',
    r'\b(19|20)\d{2}\b'
]

# ------------------------------
# OCR helpers
# ------------------------------
def extract_kyc_fields(file_path: str, force_type: Optional[str] = None) -> Dict[str, Any]:
    """
    Extracts KYC fields from an Aadhaar or PAN image.
    Returns a dict with keys:
      card_type, name, dob, aadhaar_number, pan_number
    (number keys present only when relevant)
    """
    try:
        result = _ocr.ocr(file_path, cls=True)
        lines = []
        for block in result:
            for line in block:
                text = re.sub(r'\s+', ' ', line[1][0].strip())
                if text:
                    lines.append(text)

        full_text = "\n".join(lines)

        if force_type:
            card_type = force_type.upper()
        else:
            card_type = "UNKNOWN"
            if re.search(PAN_REGEX, full_text):
                card_type = "PAN"
            elif re.search(AADHAAR_REGEX, full_text):
                card_type = "AADHAAR"

        out = {"card_type": card_type}

        if card_type == "PAN":
            out["pan_number"] = _first(PAN_REGEX, full_text) or "Not found"
            out["dob"] = _extract_dob(lines)
            out["name"] = _extract_pan_name(lines)
        elif card_type == "AADHAAR":
            out["aadhaar_number"] = _first(AADHAAR_REGEX, full_text) or "Not found"
            out["dob"] = _extract_dob(lines)
            out["name"] = _extract_aadhaar_name(lines)
        else:
            out["name"] = _extract_generic_name(lines)
            out["dob"] = _extract_dob(lines)
            out["error"] = "Could not identify document as PAN or Aadhaar."
        return out

    except Exception as e:
        return {"error": f"OCR processing failed: {str(e)}", "card_type": "UNKNOWN"}

def _first(pattern: str, text: str, flags: int = 0) -> Optional[str]:
    m = re.search(pattern, text, flags)
    return m.group(0) if m else None

def _extract_dob(lines):
    for line in lines:
        for pattern in DOB_REGEXES[:-1]:
            m = re.search(pattern, line, re.IGNORECASE)
            if m:
                return m.group(0)
    for line in lines:
        m = re.search(DOB_REGEXES[-1], line)
        if m and any(lbl in line.upper() for lbl in ["YOB", "YEAR", "BIRTH", "DOB"]):
            return m.group(0)
    return "Not found"

def _extract_pan_name(lines):
    for i, line in enumerate(lines):
        if "INCOME TAX DEPARTMENT" in line.upper():
            for j in range(i + 1, len(lines)):
                candidate = lines[j].strip()
                if re.match(r'^[A-Z\s.]+$', candidate) and not re.search(r'\d', candidate):
                    if not any(x in candidate.upper() for x in ["INDIA", "GOVT", "DEPARTMENT"]):
                        return candidate
    return "Not found"

def _extract_aadhaar_name(lines):
    for i, line in enumerate(lines):
        if any(re.search(p, line, re.IGNORECASE) for p in DOB_REGEXES):
            if i > 0:
                candidate = lines[i - 1].strip()
                if _looks_like_name(candidate):
                    return candidate
    for line in lines:
        if _looks_like_name(line.strip()):
            return line.strip()
    return "Not found"

def _extract_generic_name(lines):
    for line in lines:
        if _looks_like_name(line.strip()):
            return line.strip()
    return "Not found"

def _looks_like_name(text: str) -> bool:
    if re.search(r'\d', text):
        return False
    if len(text.split()) < 2:
        return False
    banned = ["DOB", "INDIA", "MALE", "FEMALE", "GOVERNMENT"]
    return not any(b in text.upper() for b in banned)

# ------------------------------
# Salesforce helpers
# ------------------------------
def connect_salesforce(
    username: Optional[str] = None,
    password: Optional[str] = None,
    token: Optional[str] = None,
    domain: Optional[str] = None,
):
    """
    Connect using provided args; fallback to environment variables.
    Returns a Salesforce client or raises an Exception with the root cause.
    """
    sf_username = username or os.getenv("SF_USERNAME", "")
    sf_password = password or os.getenv("SF_PASSWORD", "")
    sf_token    = token    or os.getenv("SF_TOKEN", "")
    sf_domain   = (domain or os.getenv("SF_DOMAIN", "login")).strip() or "login"

    if not (sf_username and sf_password and sf_token):
        raise RuntimeError(
            "Missing Salesforce credentials. Provide username/password/token (via UI or env vars)."
        )

    return Salesforce(
        username=sf_username,
        password=sf_password,
        security_token=sf_token,
        domain=sf_domain  # "login"=prod, "test"=sandbox
    )

def create_kyc_record(
    sf,
    aadhaar: Optional[Dict[str, Any]],
    pan: Optional[Dict[str, Any]],
    agent_id: Optional[str] = None
) -> Dict[str, Any]:
    """
    Create a single KYC_Record__c, filling the Aadhaar_* and Pan_* fields
    from the respective OCR outputs. Fields (exact API names):

      Aadhaar_Name__c
      Aadhaar_DOB__c
      Aadhaar_Number__c
      Pan_Name__c
      Pan_DOB__c
      PAN_Number__c
    """
    try:
        def safe(v): 
            if not v or v == "Not found":
                return ""
            return v

        record = {
            "Aadhaar_Name__c":   safe(aadhaar.get("name") if aadhaar else ""),
            "Aadhaar_DOB__c":    safe(aadhaar.get("dob") if aadhaar else ""),
            "Aadhaar_Number__c": safe(aadhaar.get("aadhaar_number") if aadhaar else ""),
            "Pan_Name__c":       safe(pan.get("name") if pan else ""),
            "Pan_DOB__c":        safe(pan.get("dob") if pan else ""),
            "PAN_Number__c":     safe(pan.get("pan_number") if pan else ""),
        }

        if agent_id:
            record["Agent__c"] = agent_id  # only if you have this field in org

        resp = sf.KYC_Record__c.create(record)
        return {"status": "success", "record_id": resp.get("id"), "payload_sent": record}
    except Exception as e:
        return {"status": "error", "message": str(e)}