arrr-erp/app/modules/accounting/gst_reconciliation_ui.py

from __future__ import annotations

import calendar
import re
import tempfile
import uuid
import io
import json
import zipfile
from datetime import date, datetime, timedelta, timezone
from pathlib import Path
from urllib.parse import urlencode

import jwt
from fastapi import APIRouter, File, Form, Request, UploadFile
from fastapi.responses import FileResponse, JSONResponse, RedirectResponse, StreamingResponse
from sqlalchemy import select

from app.core.db.common import CommonSessionLocal
from app.core.security.csrf import get_or_create_csrf_token, validate_csrf
from app.core.settings import get_settings
from app.core.templating import templates
from app.modules.accounting.agent_bridge import request_agent_command
from app.modules.accounting.ui import _accounting_storage_payload, _analysis_accounting_storage_payload, _find_visible_client, _require_partner, _visible_clients, _node_online
from app.modules.core.rbac.deps import get_user_permissions, get_user_roles
from app.modules.credential_vault.crypto import decrypt_value
from app.modules.credential_vault.models import CredentialVaultEntry
from app.modules.credential_vault.service import can_view_entry, log_access
from app.modules.documents.services import client_folder_parts, get_active_storage_node_for_branch, sanitize_segment
from app.modules.registrations.models import ClientRegistration, RegistrationType



def _gst_parse_date(value: str) -> str:
    text = str(value or "").strip()
    for fmt in ("%d-%m-%Y", "%d/%m/%Y", "%Y-%m-%d", "%d-%b-%Y", "%d/%m/%y"):
        try:
            return datetime.strptime(text, fmt).date().isoformat()
        except Exception:
            pass
    return text


def _gst_float(value) -> float:
    try:
        return float(value or 0)
    except Exception:
        return 0.0


def _gst_doc_key(value: str) -> str:
    return "".join(ch for ch in str(value or "").upper() if ch.isalnum())


def _gst_walk(obj):
    if isinstance(obj, dict):
        yield obj
        for value in obj.values():
            yield from _gst_walk(value)
    elif isinstance(obj, list):
        for value in obj:
            yield from _gst_walk(value)


def _gst_tax_from_invoice(inv: dict) -> dict[str, float]:
    out = {"taxable": 0.0, "igst": 0.0, "cgst": 0.0, "sgst": 0.0, "cess": 0.0}
    for item in inv.get("itms") or inv.get("items") or []:
        detail = item.get("itm_det") or item.get("item_det") or item
        out["taxable"] += _gst_float(detail.get("txval"))
        out["igst"] += _gst_float(detail.get("iamt"))
        out["cgst"] += _gst_float(detail.get("camt"))
        out["sgst"] += _gst_float(detail.get("samt"))
        out["cess"] += _gst_float(detail.get("csamt"))
    if not any(out.values()):
        out["taxable"] = _gst_float(inv.get("txval") or inv.get("taxable_value"))
        out["igst"] = _gst_float(inv.get("iamt") or inv.get("igst"))
        out["cgst"] = _gst_float(inv.get("camt") or inv.get("cgst"))
        out["sgst"] = _gst_float(inv.get("samt") or inv.get("sgst"))
        out["cess"] = _gst_float(inv.get("csamt") or inv.get("cess"))
    return out


def extract_invoice_rows(data) -> list[dict]:
    """Normalize invoice-like rows from GST JSON without importing workstation runtime code."""
    rows: list[dict] = []
    seen: set[tuple[str, str, str]] = set()
    for obj in _gst_walk(data):
        invno = str(
            obj.get("inum")
            or obj.get("inv_num")
            or obj.get("doc_no")
            or obj.get("invoiceNumber")
            or ""
        ).strip()
        if not invno:
            continue
        gstin = str(
            obj.get("ctin")
            or obj.get("stin")
            or obj.get("supplier_gstin")
            or obj.get("recipientGstin")
            or ""
        ).strip().upper()
        invoice_date = _gst_parse_date(
            str(
                obj.get("idt")
                or obj.get("inv_date")
                or obj.get("doc_date")
                or obj.get("invoiceDate")
                or ""
            )
        )
        key = (gstin, _gst_doc_key(invno), invoice_date)
        if key in seen:
            continue
        seen.add(key)
        tax = _gst_tax_from_invoice(obj)
        rows.append(
            {
                "gstin": gstin,
                "invoice_no": invno,
                "invoice_key": _gst_doc_key(invno),
                "invoice_date": invoice_date,
                "invoice_value": _gst_float(
                    obj.get("val") or obj.get("invoice_value") or obj.get("invoiceValue")
                ),
                **tax,
            }
        )
    return rows


def extract_gstr3b_itc(data) -> dict[str, float]:
    """Normalize GSTR-3B eligible-ITC totals from uploaded portal JSON."""
    totals = {"igst": 0.0, "cgst": 0.0, "sgst": 0.0, "cess": 0.0}
    if not isinstance(data, dict):
        return totals
    itc = data.get("itc_elg") or {}
    for item in itc.get("itc_avl") or []:
        totals["igst"] += _gst_float(item.get("iamt"))
        totals["cgst"] += _gst_float(item.get("camt"))
        totals["sgst"] += _gst_float(item.get("samt"))
        totals["cess"] += _gst_float(item.get("csamt"))
    return totals

router = APIRouter(prefix="/tools/accounting/gst-reconciliation", tags=["accounting-gst-reconciliation-ui"])
_TOKEN_PURPOSE = "gst_lightweight_operator_v3"
_TOKEN_MINUTES = 15
_UPLOAD_ROOT = Path(tempfile.gettempdir()) / "audit_firm_gst_operator_uploads"
_OPERATOR_AGENT_VERSION = "1.6.0"
_OPERATOR_RUNTIME_ROOT = Path(__file__).resolve().parent / "gst_operator_agent_runtime"
_OPERATOR_PACKAGE_FILES = ("gst_operator_agent.py", "requirements.txt", "README.txt", "install_gst_operator_agent.ps1", "uninstall_gst_operator_agent.ps1", "build_gst_operator_agent_exe.ps1")
_OPERATOR_EXE_NAME = f"ARRR_GST_Operator_Agent_Setup_{_OPERATOR_AGENT_VERSION}.exe"
_OPERATOR_EXE_PATH = _OPERATOR_RUNTIME_ROOT / "dist" / _OPERATOR_EXE_NAME


def _fy_bounds(fy: str) -> tuple[date, date]:
    m = re.fullmatch(r"(\d{4})-(\d{2})", str(fy or "").strip())
    if not m:
        raise ValueError("Invalid financial year.")
    y = int(m.group(1))
    return date(y, 4, 1), date(y + 1, 3, 31)


def _periods_for_fy(fy: str) -> list[str]:
    start, _ = _fy_bounds(fy)
    periods=[]
    y=start.year; m=4
    for _ in range(12):
        periods.append(f"{m:02d}{y:04d}")
        m += 1
        if m == 13:
            m = 1; y += 1
    return periods


def _fy_for_period(period: str) -> str:
    digits = re.sub(r"\D", "", period or "")
    if len(digits) != 6:
        raise ValueError("Return period must be MMYYYY.")
    month, year = int(digits[:2]), int(digits[2:])
    if month < 1 or month > 12:
        raise ValueError("Invalid GST return month.")
    sy = year if month >= 4 else year - 1
    return f"{sy}-{str(sy+1)[-2:]}"


def _period_bounds(period: str) -> tuple[str, str]:
    digits = re.sub(r"\D", "", period or "")
    month, year = int(digits[:2]), int(digits[2:])
    last = calendar.monthrange(year, month)[1]
    return date(year, month, 1).isoformat(), date(year, month, last).isoformat()


def _gst_regs(db, tenant_id: int, client_id: int):
    return db.execute(
        select(ClientRegistration, RegistrationType)
        .join(RegistrationType, RegistrationType.id == ClientRegistration.registration_type_id)
        .where(ClientRegistration.tenant_id == tenant_id, ClientRegistration.client_id == client_id)
        .order_by(ClientRegistration.id.asc())
    ).all()


def _is_gstin(reg, typ) -> bool:
    code = str(getattr(typ, "code", "") or "").upper().strip()
    num = re.sub(r"\s+", "", str(getattr(reg, "registration_number", "") or "").upper())
    return code == "GSTIN" and len(num) == 15


def _norm_gstin(value: str | None) -> str:
    return re.sub(r"[^0-9A-Z]", "", str(value or "").upper())


def _credential_is_gst_portal(entry: CredentialVaultEntry) -> bool:
    category = str(entry.category or "").strip().lower()
    title = str(entry.title or "").strip().lower()
    portal_url = str(entry.portal_url or "").strip().lower()
    reference = str(entry.reference_number or "").strip().lower()
    if category == "gst_portal":
        return True
    haystack = " ".join((title, portal_url, reference))
    gst_hint = (
        "gst portal" in haystack or "gst login" in haystack
        or "services.gst.gov.in" in portal_url or "www.gst.gov.in" in portal_url
    )
    if not gst_hint:
        return False
    if category in {"eway_bill", "einvoice", "api_key"}:
        return "gst portal" in title or "gst login" in title
    return True


def _credential_matches_gstin(entry: CredentialVaultEntry, registration_id: int, gstin: str) -> bool:
    if int(entry.registration_id or 0) == int(registration_id):
        return True
    if _norm_gstin(entry.reference_number) == _norm_gstin(gstin):
        return True
    return entry.registration_id is None


def _vault_entries(db, user, request, tenant_id: int, client_id: int, registration_id: int | None = None, gstin: str = ""):
    q = select(CredentialVaultEntry).where(
        CredentialVaultEntry.tenant_id == tenant_id,
        CredentialVaultEntry.client_id == client_id,
        CredentialVaultEntry.status != "archived",
    )
    rows = db.execute(q.order_by(CredentialVaultEntry.title.asc())).scalars().all()
    branch_id = request.session.get("active_branch_id") or getattr(user, "branch_id", None)
    visible = [r for r in rows if can_view_entry(db, user, r, int(branch_id) if branch_id else None)]
    if not registration_id:
        return [r for r in visible if _credential_is_gst_portal(r)]
    eligible = [r for r in visible if _credential_is_gst_portal(r) and _credential_matches_gstin(r, int(registration_id), gstin)]
    def rank(entry: CredentialVaultEntry):
        exact_registration = int(entry.registration_id or 0) == int(registration_id)
        exact_gstin = _norm_gstin(entry.reference_number) == _norm_gstin(gstin)
        client_level = entry.registration_id is None
        return (0 if exact_registration else 1 if exact_gstin else 2 if client_level else 3, str(entry.title or "").lower(), int(entry.id or 0))
    return sorted(eligible, key=rank)


def _storage_payload(client, fy: str, gstin: str, *, db=None, tenant_id: int | None = None, require_registered: bool = False) -> tuple[dict, str]:
    fy_folder = sanitize_segment(f"FY{fy}", "FY")
    letter, client_folder = client_folder_parts(client, int(client.id))
    root = Path(fy_folder) / "Clients" / letter / client_folder
    if require_registered:
        accounting = _analysis_accounting_storage_payload(
            client, fy, db=db, tenant_id=tenant_id
        )
    else:
        accounting = _accounting_storage_payload(client, fy, db=db, tenant_id=tenant_id)
    accounting.setdefault("accounting_relative_dir", (root / "Accounting").as_posix())
    return accounting, (root / "GST" / sanitize_segment(f"GSTIN_{gstin}", "GSTIN")).as_posix()


def _redirect(client_id: int, **params):
    data = {"client_id": client_id, **{k: v for k, v in params.items() if v not in (None, "")}}
    return RedirectResponse("/tools/accounting/gst-reconciliation?" + urlencode(data), status_code=303)


def _encode_operator_token(payload: dict) -> str:
    now = datetime.now(timezone.utc)
    body = {**payload, "purpose": _TOKEN_PURPOSE, "iat": now, "exp": now + timedelta(minutes=_TOKEN_MINUTES)}
    return jwt.encode(body, get_settings().SECRET_KEY, algorithm="HS256")


def _decode_operator_token(token: str) -> dict:
    data = jwt.decode(token, get_settings().SECRET_KEY, algorithms=["HS256"])
    if data.get("purpose") != _TOKEN_PURPOSE:
        raise ValueError("Invalid GST operator token.")
    return data


def _selected_periods(download_mode: str, financial_year: str, period: str) -> list[str]:
    if download_mode == "full_fy":
        return _periods_for_fy(financial_year)
    digits = re.sub(r"\D", "", period or "")
    if len(digits) != 6:
        raise ValueError("Enter a valid MMYYYY period for Single Month mode.")
    return [digits]



GST_LOGIN_URL = "https://services.gst.gov.in/services/login"
_MAX_IMPORT_BYTES = 250 * 1024 * 1024


def _safe_upload_name(name: str) -> str:
    cleaned = re.sub(r"[^A-Za-z0-9._ -]+", "_", str(name or "").strip()).strip(" ._")
    return cleaned or "return.json"


def _period_from_name_or_payload(name: str, payload, fallback: str = "") -> str:
    def valid(v: str) -> str:
        d = re.sub(r"\D", "", str(v or ""))
        if len(d) == 6:
            mm, yy = int(d[:2]), int(d[2:])
            if 1 <= mm <= 12 and 2000 <= yy <= 2100:
                return d
            yy2, mm2 = int(d[:4]), int(d[4:])
            if 2000 <= yy2 <= 2100 and 1 <= mm2 <= 12:
                return f"{mm2:02d}{yy2:04d}"
        return ""

    if isinstance(payload, dict):
        stack = [payload]
        seen = 0
        while stack and seen < 5000:
            obj = stack.pop(); seen += 1
            if not isinstance(obj, dict):
                continue
            for key in ("fp", "rtn_prd", "return_period", "period", "ret_period"):
                if key in obj:
                    found = valid(obj.get(key))
                    if found:
                        return found
            for value in obj.values():
                if isinstance(value, dict):
                    stack.append(value)
                elif isinstance(value, list):
                    stack.extend(x for x in value[:200] if isinstance(x, dict))

    for pattern in (r"(?<!\d)(0[1-9]|1[0-2])(20\d{2})(?!\d)", r"(?<!\d)(20\d{2})(0[1-9]|1[0-2])(?!\d)"):
        m = re.search(pattern, str(name or ""))
        if m:
            if len(m.group(1)) == 2:
                return f"{m.group(1)}{m.group(2)}"
            return f"{m.group(2)}{m.group(1)}"
    return valid(fallback)


def _json_members(upload_name: str, content: bytes) -> list[tuple[str, object, bytes]]:
    lower = str(upload_name or "").lower()
    if lower.endswith(".json"):
        try:
            payload = json.loads(content.decode("utf-8-sig", errors="strict"))
        except Exception as exc:
            raise ValueError(f"{upload_name}: invalid JSON file ({exc}).") from exc
        return [(upload_name, payload, content)]
    if lower.endswith(".zip"):
        out = []
        try:
            with zipfile.ZipFile(io.BytesIO(content), "r") as archive:
                for info in archive.infolist():
                    if info.is_dir() or not info.filename.lower().endswith(".json"):
                        continue
                    raw = archive.read(info)
                    try:
                        payload = json.loads(raw.decode("utf-8-sig", errors="strict"))
                    except Exception:
                        continue
                    out.append((Path(info.filename).name, payload, raw))
        except zipfile.BadZipFile as exc:
            raise ValueError(f"{upload_name}: invalid ZIP file.") from exc
        if not out:
            raise ValueError(f"{upload_name}: ZIP does not contain readable JSON return data.")
        return out
    raise ValueError(f"{upload_name}: only GST JSON or ZIP files are supported.")


def _merge_invoice_rows(payloads: list[object]) -> list[dict]:
    rows=[]; seen=set()
    for payload in payloads:
        for row in extract_invoice_rows(payload):
            key=(str(row.get("gstin") or ""), str(row.get("invoice_key") or ""), str(row.get("invoice_date") or ""))
            if key in seen:
                continue
            seen.add(key); rows.append(row)
    return rows


def _merge_gstr3b_itc(payloads: list[object]) -> dict[str, float]:
    totals={"igst":0.0,"cgst":0.0,"sgst":0.0,"cess":0.0}
    for payload in payloads:
        row=extract_gstr3b_itc(payload)
        for k in totals:
            totals[k]=round(totals[k]+float(row.get(k) or 0),2)
    return totals

@router.get("")
def page(request: Request, client_id: str = "", registration_id: str = "", period: str = "", financial_year: str = "", download_mode: str = "single", message: str = "", error: str = ""):
    db = CommonSessionLocal()
    try:
        user, response = _require_partner(request, db, "accounting.learning.view")
        if response:
            return response
        clients, scope = _visible_clients(db, request, user)
        client_id_text = str(client_id or "").strip()
        registration_id_text = str(registration_id or "").strip()
        client_id_value = int(client_id_text) if client_id_text.isdigit() else None
        registration_id_value = int(registration_id_text) if registration_id_text.isdigit() else None
        selected = next((c for c in clients if client_id_value and int(c.id) == client_id_value), None)
        registrations=[]; selected_reg=None; credentials=[]; node=None
        if not financial_year:
            financial_year = str(request.session.get("active_financial_year") or request.session.get("active_fy") or "2025-26")
        if download_mode not in {"single", "full_fy"}:
            download_mode = "single"
        if selected:
            registrations=[(r,t) for r,t in _gst_regs(db,scope.tenant_id,selected.id) if _is_gstin(r,t)]
            selected_reg=next((r for r,t in registrations if registration_id_value and int(r.id)==registration_id_value),None)
            if not selected_reg and registrations:
                selected_reg=registrations[0][0]
            selected_gstin = _norm_gstin(getattr(selected_reg, "registration_number", "")) if selected_reg else ""
            credentials=_vault_entries(db,user,request,scope.tenant_id,selected.id,int(selected_reg.id) if selected_reg else None, selected_gstin)
            node=get_active_storage_node_for_branch(db,scope.tenant_id,scope.branch_id)
        operator_job = request.session.pop("gst_operator_job", None)
        return templates.TemplateResponse("modules/accounting/templates/accounting/gst_reconciliation.html",{
            "request":request,"current_user":user,"current_user_roles":get_user_roles(db,user.id),"current_user_permissions":get_user_permissions(db,user.id),"csrf_token":get_or_create_csrf_token(request),
            "clients":clients,"selected_client":selected,"registrations":registrations,"selected_registration":selected_reg,"credentials":credentials,"node":node,"node_online":_node_online(node) if node else False,
            "period":period,"financial_year":financial_year,"download_mode":download_mode,"operator_job":operator_job,"gst_login_url":GST_LOGIN_URL,
            "operator_agent_version":_OPERATOR_AGENT_VERSION,
            "message":message,"error":error,"title":"GST Return Reconciliation",
        })
    finally:
        db.close()


@router.get("/operator-agent/download")
def download_operator_agent(request: Request):
    db = CommonSessionLocal()
    try:
        user, response = _require_partner(request, db, "accounting.learning.view")
        if response:
            return response
        if _OPERATOR_EXE_PATH.is_file():
            return FileResponse(
                path=str(_OPERATOR_EXE_PATH),
                media_type="application/vnd.microsoft.portable-executable",
                filename=_OPERATOR_EXE_NAME,
            )
        return JSONResponse(
            {
                "ok": False,
                "error": (
                    "The GST Operator Agent Windows EXE has not been built in this deployment. "
                    "Run app/modules/accounting/gst_operator_agent_runtime/build_gst_operator_agent_exe.ps1 "
                    "on the Windows development repository, commit the generated dist EXE, and redeploy."
                ),
            },
            status_code=503,
        )
    finally:
        db.close()


@router.get("/operator-agent/download-source")
def download_operator_agent_source(request: Request):
    """Preserve the existing source/PowerShell package as a repair fallback."""
    db = CommonSessionLocal()
    try:
        user, response = _require_partner(request, db, "accounting.learning.view")
        if response:
            return response
        if not _OPERATOR_RUNTIME_ROOT.is_dir():
            return JSONResponse({"ok": False, "error": "GST Operator Agent runtime is missing from this ERP build."}, status_code=404)
        memory = io.BytesIO()
        with zipfile.ZipFile(memory, "w", zipfile.ZIP_DEFLATED, compresslevel=6) as archive:
            for name in _OPERATOR_PACKAGE_FILES:
                path = _OPERATOR_RUNTIME_ROOT / name
                if not path.is_file():
                    return JSONResponse({"ok": False, "error": f"GST Operator Agent package file is missing: {name}"}, status_code=500)
                archive.write(path, Path("ARRR_GST_Operator_Agent") / name)
        memory.seek(0)
        headers = {"Content-Disposition": f'attachment; filename="ARRR_GST_Operator_Agent_Source_{_OPERATOR_AGENT_VERSION}.zip"'}
        return StreamingResponse(memory, media_type="application/zip", headers=headers)
    finally:
        db.close()


def _operator_public_base_url(request: Request) -> str:
    proto = str(request.headers.get("x-forwarded-proto") or request.url.scheme or "https").split(",")[0].strip()
    host = str(request.headers.get("x-forwarded-host") or request.headers.get("host") or request.url.netloc).split(",")[0].strip()
    if not host:
        return "https://office.arrrassociates.com"
    return f"{proto}://{host}".rstrip("/")


@router.post("/download/start")
def start_download(
    request: Request,
    client_id: int=Form(...), registration_id: int=Form(...), credential_id: int=Form(...),
    period: str=Form(""), financial_year: str=Form(...), download_mode: str=Form("single"),
    gstr1: str=Form(""), gstr2b: str=Form(""), gstr3b: str=Form(""), gstr2a: str=Form(""),
    csrf_token: str=Form(...),
):
    validate_csrf(request,csrf_token)
    db=CommonSessionLocal()
    try:
        user,response=_require_partner(request,db,"accounting.learning.manage")
        if response: return response
        client,_,scope=_find_visible_client(db,request,user,client_id)
        if not client: return _redirect(client_id,error="Client is not available in your scope.")
        pair=next(((r,t) for r,t in _gst_regs(db,scope.tenant_id,client.id) if int(r.id)==registration_id and _is_gstin(r,t)),None)
        if not pair: return _redirect(client_id,error="Select a valid GSTIN registration.")
        reg,_=pair; gstin=_norm_gstin(reg.registration_number)
        cred=db.get(CredentialVaultEntry,credential_id)
        credential_ok = bool(cred and int(cred.tenant_id or 0)==int(scope.tenant_id) and int(cred.client_id or 0)==int(client.id) and str(cred.status or "").lower()!="archived" and can_view_entry(db,user,cred,scope.branch_id) and _credential_is_gst_portal(cred) and _credential_matches_gstin(cred,int(reg.id),gstin))
        if not credential_ok:
            return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=download_mode,error="Selected GST Portal credential is not available for this client/GSTIN.")
        node=get_active_storage_node_for_branch(db,scope.tenant_id,scope.branch_id)
        if not node or not _node_online(node):
            return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=download_mode,error="Local Storage Agent is offline.")
        periods=_selected_periods(download_mode,financial_year,period)
        return_types=[name for name,flag in (("GSTR1",gstr1),("GSTR2B",gstr2b),("GSTR3B",gstr3b),("GSTR2A",gstr2a)) if flag]
        if not return_types:
            return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=download_mode,error="Select at least one GST return to download.")
        # Ensure FY is consistent in single-month mode.
        if download_mode == "single":
            financial_year=_fy_for_period(periods[0])
        accounting_payload,gst_dir=_storage_payload(client,financial_year,gstin,db=db,tenant_id=scope.tenant_id)
        accounting_dir=str(accounting_payload.get("accounting_relative_dir") or "")
        jti=uuid.uuid4().hex
        token=_encode_operator_token({
            "jti":jti,"user_id":int(user.id),"tenant_id":int(scope.tenant_id),"branch_id":int(scope.branch_id),
            "node_code":str(node.node_code),"client_id":int(client.id),"client_name":str(client.client_name or ""),
            "registration_id":int(reg.id),"credential_id":int(cred.id),"gstin":gstin,"financial_year":financial_year,
            "periods":periods,"return_types":return_types,"gst_relative_dir":gst_dir,"accounting_relative_dir":accounting_dir,
            "erp_base_url":_operator_public_base_url(request),
        })
        log_access(db,request,user,cred,"use_for_gst_download",reason=f"GST returns {financial_year}: {','.join(return_types)}",fields="username,secret",success=True)
        db.commit()
        request.session["gst_operator_job"]={"job_id":jti,"periods":periods,"return_types":return_types,"launch_url":"arrrgst://start?"+urlencode({"token":token})}
        return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=download_mode,message="GST download prepared. Windows will launch the lightweight GST Operator Agent through the ARRR GST protocol. The agent will open the visible GST browser and autofill the selected Credential Vault login. Complete CAPTCHA/OTP there; the agent will enter Return Dashboard through the normal GST portal sequence before downloading return data and transferring it to configured client storage.")
    except Exception as exc:
        db.rollback(); return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=download_mode,error=str(exc))
    finally:
        db.close()


@router.post("/operator/redeem")
async def operator_redeem(request: Request):
    body=await request.json(); token=str(body.get("token") or "")
    try:
        data=_decode_operator_token(token)
        db=CommonSessionLocal()
        try:
            cred=db.get(CredentialVaultEntry,int(data["credential_id"]))
            if not cred or int(cred.tenant_id or 0)!=int(data["tenant_id"]) or int(cred.client_id or 0)!=int(data["client_id"]) or str(cred.status or "").lower()=="archived":
                raise ValueError("GST credential is no longer available.")
            username=decrypt_value(cred.tenant_id,cred.username_encrypted) or ""
            password=decrypt_value(cred.tenant_id,cred.secret_encrypted) or ""
            if not username or not password:
                raise ValueError("GST username/password is missing in Credential Vault.")
        finally:
            db.close()
        return JSONResponse({"ok":True,"payload":{**{k:v for k,v in data.items() if k not in {"iat","exp","purpose"}},"username":username,"password":password,"upload_url":str(data.get("erp_base_url") or "https://office.arrrassociates.com").rstrip("/")+"/tools/accounting/gst-reconciliation/operator/upload","login_timeout_seconds":900,"operator_agent_min_version":_OPERATOR_AGENT_VERSION}})
    except Exception as exc:
        return JSONResponse({"ok":False,"error":str(exc)},status_code=400)


@router.post("/operator/upload")
async def operator_upload(token: str=Form(...), package: UploadFile=File(...)):
    try:
        data=_decode_operator_token(token); jti=str(data.get("jti") or "")
        if not jti:
            raise ValueError("GST operator job id is missing.")
        _UPLOAD_ROOT.mkdir(parents=True,exist_ok=True)
        package_path=_UPLOAD_ROOT/f"{jti}.zip"
        total=0
        with package_path.open("wb") as out:
            while True:
                chunk=await package.read(1024*1024)
                if not chunk: break
                total += len(chunk)
                if total > 250*1024*1024:
                    raise ValueError("GST download package exceeds the 250 MB safety limit.")
                out.write(chunk)
        package_url=str(get_settings().ERP_PUBLIC_BASE_URL).rstrip("/")+f"/tools/accounting/gst-reconciliation/operator/package/{jti}?token={token}"
        result=request_agent_command(str(data["node_code"]),"gst_return_package_store",{
            "client_id":int(data["client_id"]),"gstin":str(data["gstin"]),"financial_year":str(data["financial_year"]),
            "gst_relative_dir":str(data["gst_relative_dir"]),"package_url":package_url,"periods":data.get("periods") or [],"return_types":data.get("return_types") or [],
        },timeout_seconds=120)
        if not result.get("ok"):
            raise RuntimeError(result.get("error") or "Local Storage Agent could not store the GST return package.")
        try: package_path.unlink(missing_ok=True)
        except Exception: pass
        return JSONResponse({"ok":True,"stored":result.get("result") or {}})
    except Exception as exc:
        return JSONResponse({"ok":False,"error":str(exc)},status_code=400)


@router.get("/operator/package/{job_id}")
def operator_package(job_id: str, token: str):
    try:
        data=_decode_operator_token(token)
        if str(data.get("jti") or "") != str(job_id):
            raise ValueError("GST package token does not match the requested job.")
        path=_UPLOAD_ROOT/f"{job_id}.zip"
        if not path.is_file():
            return JSONResponse({"ok":False,"error":"GST package is no longer available."},status_code=404)
        return FileResponse(path,media_type="application/zip",filename=f"gst_returns_{job_id}.zip")
    except Exception as exc:
        return JSONResponse({"ok":False,"error":str(exc)},status_code=400)



@router.post("/import")
async def import_downloaded_returns(
    request: Request,
    client_id: int = Form(...), registration_id: int = Form(...), period: str = Form(""),
    financial_year: str = Form(...), download_mode: str = Form("single"),
    gstr1_files: list[UploadFile] = File(default=[]),
    gstr2b_files: list[UploadFile] = File(default=[]),
    gstr3b_files: list[UploadFile] = File(default=[]),
    gstr2a_files: list[UploadFile] = File(default=[]),
    csrf_token: str = Form(...),
):
    validate_csrf(request, csrf_token)
    db = CommonSessionLocal()
    temp_root = None
    package_path = None
    try:
        user, response = _require_partner(request, db, "accounting.learning.manage")
        if response:
            return response
        client, _, scope = _find_visible_client(db, request, user, client_id)
        if not client:
            return _redirect(client_id, error="Client is not available in your scope.")
        pair = next(((r,t) for r,t in _gst_regs(db, scope.tenant_id, client.id) if int(r.id)==registration_id and _is_gstin(r,t)), None)
        if not pair:
            return _redirect(client_id, error="Select a valid GSTIN registration.")
        reg, _ = pair
        gstin = _norm_gstin(reg.registration_number)
        node = get_active_storage_node_for_branch(db, scope.tenant_id, scope.branch_id)
        if not node or not _node_online(node):
            return _redirect(client_id, registration_id=registration_id, period=period, financial_year=financial_year, download_mode=download_mode, error="Configured Local Storage Agent is offline.")

        allowed_periods = set(_selected_periods(download_mode, financial_year, period))
        if download_mode == "single":
            financial_year = _fy_for_period(next(iter(allowed_periods)))
        accounting_payload, gst_dir = _storage_payload(client, financial_year, gstin, db=db, tenant_id=scope.tenant_id)
        accounting_dir = str(accounting_payload.get("accounting_relative_dir") or "")
        groups = {
            "GSTR1": gstr1_files or [], "GSTR2B": gstr2b_files or [],
            "GSTR3B": gstr3b_files or [], "GSTR2A": gstr2a_files or [],
        }
        if not any(groups.values()):
            raise ValueError("Select at least one GST JSON/ZIP file to import.")

        temp_root = Path(tempfile.mkdtemp(prefix="gst_manual_import_"))
        by_period: dict[str, dict[str, dict]] = {}
        total_bytes = 0
        imported_files = 0
        for rtype, uploads in groups.items():
            for upload in uploads:
                if not upload or not upload.filename:
                    continue
                content = await upload.read()
                total_bytes += len(content)
                if total_bytes > _MAX_IMPORT_BYTES:
                    raise ValueError("GST import exceeds the 250 MB safety limit.")
                for member_name, payload, raw in _json_members(upload.filename, content):
                    p = _period_from_name_or_payload(member_name, payload, period if download_mode == "single" else "")
                    if not p:
                        raise ValueError(f"Could not determine return period for {member_name}. Use GST JSON containing fp/return period, or Single Month mode.")
                    if p not in allowed_periods:
                        raise ValueError(f"{member_name} belongs to {p}, outside the selected {'financial year' if download_mode=='full_fy' else 'month'}.")
                    slot = by_period.setdefault(p, {}).setdefault(rtype, {"payloads": [], "raw": []})
                    slot["payloads"].append(payload)
                    slot["raw"].append((member_name, raw))
                    imported_files += 1

        if not by_period:
            raise ValueError("No readable GST JSON return data was found in the selected files.")

        now = datetime.now(timezone.utc).isoformat()
        return_types = set()
        for p, period_data in sorted(by_period.items()):
            base = temp_root / p
            raw_dir = base / "raw"
            normalized_dir = base / "normalized"
            raw_dir.mkdir(parents=True, exist_ok=True)
            normalized_dir.mkdir(parents=True, exist_ok=True)
            manifest_downloaded=[]; normalized={}
            for rtype, data in period_data.items():
                return_types.add(rtype)
                originals = raw_dir / "original"
                originals.mkdir(parents=True, exist_ok=True)
                payloads = data["payloads"]
                for idx, (member_name, raw) in enumerate(data["raw"], 1):
                    safe = _safe_upload_name(member_name)
                    dest = originals / f"{rtype}_{idx:03d}_{safe}"
                    dest.write_bytes(raw)
                    manifest_downloaded.append({"return_type": rtype, "source_name": member_name, "path": str(dest.relative_to(temp_root).as_posix()), "bytes": len(raw)})
                canonical = payloads[0] if len(payloads) == 1 else {"period": p, "source_files": len(payloads), "payloads": payloads}
                (raw_dir / f"{p}_{rtype}.json").write_text(json.dumps(canonical, ensure_ascii=False, indent=2, default=str), encoding="utf-8")
                if rtype in {"GSTR1", "GSTR2A", "GSTR2B"}:
                    rows = _merge_invoice_rows(payloads)
                    (normalized_dir / f"{p}_{rtype}_invoices.json").write_text(json.dumps({"period":p,"rows":rows}, ensure_ascii=False, indent=2, default=str), encoding="utf-8")
                    normalized[rtype] = {"invoice_rows": len(rows)}
                elif rtype == "GSTR3B":
                    itc = _merge_gstr3b_itc(payloads)
                    (normalized_dir / f"{p}_{rtype}_itc.json").write_text(json.dumps({"period":p,"itc":itc}, ensure_ascii=False, indent=2), encoding="utf-8")
                    normalized[rtype] = {"itc": itc}
            (base / "download_manifest.json").write_text(json.dumps({"gstin":gstin,"period":p,"financial_year":financial_year,"downloaded_at_utc":now,"source":"manual_gst_portal_browser_import","downloaded":manifest_downloaded,"normalized":normalized}, ensure_ascii=False, indent=2), encoding="utf-8")

        jti = uuid.uuid4().hex
        token = _encode_operator_token({"jti":jti,"user_id":int(user.id),"tenant_id":int(scope.tenant_id),"branch_id":int(scope.branch_id),"node_code":str(node.node_code),"client_id":int(client.id),"gstin":gstin,"financial_year":financial_year})
        _UPLOAD_ROOT.mkdir(parents=True, exist_ok=True)
        package_path = _UPLOAD_ROOT / f"{jti}.zip"
        with zipfile.ZipFile(package_path, "w", zipfile.ZIP_DEFLATED, compresslevel=6) as archive:
            for f in sorted(temp_root.rglob("*")):
                if f.is_file():
                    archive.write(f, f.relative_to(temp_root).as_posix())
        package_url = str(get_settings().ERP_PUBLIC_BASE_URL).rstrip("/") + f"/tools/accounting/gst-reconciliation/operator/package/{jti}?token={token}"
        result = request_agent_command(str(node.node_code), "gst_return_package_store", {
            "client_id": int(client.id), "gstin": gstin, "financial_year": financial_year,
            "gst_relative_dir": gst_dir, "package_url": package_url,
            "periods": sorted(by_period), "return_types": sorted(return_types),
        }, timeout_seconds=180)
        if not result.get("ok"):
            raise RuntimeError(result.get("error") or "Local Storage Agent could not store the imported GST return package.")
        return _redirect(client_id, registration_id=registration_id, period=period, financial_year=financial_year, download_mode=download_mode, message=f"Imported {imported_files} GST JSON file(s) for {len(by_period)} period(s) and stored them in the client GST directory. You can now run reconciliation.")
    except Exception as exc:
        db.rollback()
        return _redirect(client_id, registration_id=registration_id, period=period, financial_year=financial_year, download_mode=download_mode, error=str(exc))
    finally:
        db.close()
        if package_path:
            try: package_path.unlink(missing_ok=True)
            except Exception: pass
        if temp_root:
            import shutil
            shutil.rmtree(temp_root, ignore_errors=True)


@router.post("/analyze")
def analyze(request: Request, client_id: int=Form(...), registration_id: int=Form(...), period: str=Form(""), financial_year: str=Form(...), analyze_mode: str=Form("single"), csrf_token: str=Form(...)):
    validate_csrf(request,csrf_token)
    db=CommonSessionLocal()
    try:
        user,response=_require_partner(request,db,"accounting.learning.manage")
        if response: return response
        client,_,scope=_find_visible_client(db,request,user,client_id)
        if not client: return _redirect(client_id,error="Client is not available in your scope.")
        pair=next(((r,t) for r,t in _gst_regs(db,scope.tenant_id,client.id) if int(r.id)==registration_id and _is_gstin(r,t)),None)
        if not pair: return _redirect(client_id,error="GSTIN registration was not found.")
        reg,_=pair; gstin=_norm_gstin(reg.registration_number)
        periods=_periods_for_fy(financial_year) if analyze_mode=="full_fy" else [re.sub(r"\D","",period or "")]
        if any(len(p)!=6 for p in periods): raise ValueError("Enter a valid MMYYYY period.")
        accounting_payload,gst_dir=_storage_payload(client,financial_year,gstin,db=db,tenant_id=scope.tenant_id,require_registered=True)
        accounting_dir=str(accounting_payload.get("accounting_relative_dir") or "")
        node=get_active_storage_node_for_branch(db,scope.tenant_id,scope.branch_id)
        if not node or not _node_online(node): return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,error="Local Storage Agent is offline.")
        results=[]
        for p in periods:
            date_from,date_to=_period_bounds(p)
            res=request_agent_command(node.node_code,"gst_reconciliation_analyze",{**accounting_payload,"client_id":client.id,"gstin":gstin,"financial_year":financial_year,"period":p,"gst_relative_dir":gst_dir,"accounting_relative_dir":accounting_dir,"date_from":date_from,"date_to":date_to},timeout_seconds=30)
            if res.get("ok"):
                results.append((res.get("result") or {}).get("analysis") or {})
        if not results: raise RuntimeError("No stored GST periods could be reconciled.")
        if analyze_mode=="full_fy":
            def sum_counts(section):
                out={}
                for r in results:
                    for k,v in (((r.get(section) or {}).get("counts") or {}).items()): out[k]=out.get(k,0)+int(v or 0)
                return out
            taxes={k:{"gstr2b":0.0,"gstr3b":0.0,"difference":0.0} for k in ("igst","cgst","sgst","cess")}
            for r in results:
                for k,row in (r.get("itc_reconciliation") or {}).items():
                    if k in taxes:
                        for f in taxes[k]: taxes[k][f]=round(taxes[k][f]+float(row.get(f) or 0),2)
            analysis={"period":financial_year,"sales_reconciliation":{"counts":sum_counts("sales_reconciliation")},"purchase_reconciliation":{"counts":sum_counts("purchase_reconciliation")},"itc_reconciliation":taxes}
        else:
            analysis=results[0]
        request.session["gst_reconciliation_result"]=analysis
        return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,download_mode=("full_fy" if analyze_mode=="full_fy" else "single"),message="GST Purchase, Sales and ITC reconciliation completed from stored local return data and Accounting Mirror.")
    except Exception as exc:
        return _redirect(client_id,registration_id=registration_id,period=period,financial_year=financial_year,error=str(exc))
    finally:
        db.close()