roa2web-service-auto/backend/modules/data_entry/schemas/bulk.py

"""Pydantic schemas for bulk upload endpoints."""

from datetime import datetime
from typing import List, Optional

from pydantic import BaseModel, Field


class BulkUploadResponse(BaseModel):
    """Response schema for bulk upload endpoint."""

    batch_id: int = Field(..., description="Unique batch identifier for tracking")
    job_ids: List[str] = Field(..., description="List of OCR job UUIDs created")
    total_files: int = Field(..., description="Number of files in the batch")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "batch_id": 1,
                "job_ids": [
                    "550e8400-e29b-41d4-a716-446655440001",
                    "550e8400-e29b-41d4-a716-446655440002",
                ],
                "total_files": 2,
                "message": "2 files queued for processing"
            }
        }


class BatchJobInfo(BaseModel):
    """Information about a single job in a batch."""

    job_id: str = Field(..., description="OCR job UUID")
    filename: str = Field(..., description="Original filename")
    status: str = Field(..., description="Job status: pending, processing, completed, failed")
    receipt_id: Optional[int] = Field(None, description="Created receipt ID (if completed)")
    error_message: Optional[str] = Field(None, description="Error message (if failed)")


class BatchStatusResponse(BaseModel):
    """Response schema for batch status endpoint."""

    batch_id: int = Field(..., description="Batch identifier")
    status: str = Field(..., description="Overall batch status")
    total_files: int = Field(..., description="Total number of files in batch")
    pending_count: int = Field(..., description="Number of pending jobs")
    processing_count: int = Field(..., description="Number of processing jobs")
    completed_count: int = Field(..., description="Number of completed jobs")
    failed_count: int = Field(..., description="Number of failed jobs")
    jobs: List[BatchJobInfo] = Field(..., description="List of jobs with their status")
    total_amount: Optional[float] = Field(None, description="Sum of all receipt amounts")
    created_at: datetime = Field(..., description="Batch creation timestamp")

    class Config:
        json_schema_extra = {
            "example": {
                "batch_id": 1,
                "status": "processing",
                "total_files": 5,
                "pending_count": 2,
                "processing_count": 1,
                "completed_count": 2,
                "failed_count": 0,
                "jobs": [
                    {"job_id": "abc-123", "filename": "bon1.pdf", "status": "completed", "receipt_id": 15},
                    {"job_id": "def-456", "filename": "bon2.jpg", "status": "processing", "receipt_id": None},
                ],
                "total_amount": 150.50,
                "created_at": "2025-01-09T10:30:00"
            }
        }


class DuplicateFileInfo(BaseModel):
    """Information about a duplicate file detected during upload."""

    filename: str = Field(..., description="Name of the duplicate file")
    error: str = Field(default="duplicate", description="Error type (always 'duplicate')")
    existing_receipt_id: int = Field(..., description="ID of the existing receipt with same file hash")
    message: str = Field(..., description="Human-readable error message")

    class Config:
        json_schema_extra = {
            "example": {
                "filename": "bon_lidl.pdf",
                "error": "duplicate",
                "existing_receipt_id": 123,
                "message": "Fișier duplicat - există deja ca bon #123"
            }
        }


class BulkUploadResponseWithDuplicates(BaseModel):
    """Response schema for bulk upload with partial success (some duplicates)."""

    batch_id: Optional[int] = Field(None, description="Batch ID (None if all files were duplicates)")
    job_ids: List[str] = Field(default_factory=list, description="List of OCR job UUIDs created")
    total_files: int = Field(..., description="Total number of files submitted")
    processed_files: int = Field(..., description="Number of files successfully queued")
    duplicate_files: int = Field(..., description="Number of duplicate files rejected")
    duplicates: List[DuplicateFileInfo] = Field(default_factory=list, description="List of duplicate file details")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "batch_id": 1,
                "job_ids": ["550e8400-e29b-41d4-a716-446655440001"],
                "total_files": 3,
                "processed_files": 1,
                "duplicate_files": 2,
                "duplicates": [
                    {
                        "filename": "bon_lidl.pdf",
                        "error": "duplicate",
                        "existing_receipt_id": 123,
                        "message": "Fișier duplicat - există deja ca bon #123"
                    }
                ],
                "message": "1 fișier în procesare, 2 duplicate ignorate"
            }
        }


class BulkUploadError(BaseModel):
    """Error response for bulk upload validation failures."""

    detail: str = Field(..., description="Error message")
    invalid_files: Optional[List[str]] = Field(None, description="List of invalid filenames")


class RetryResponse(BaseModel):
    """Response schema for retry endpoints."""

    success: bool = Field(..., description="Whether the retry was successful")
    receipt_id: int = Field(..., description="Receipt ID that was retried")
    job_id: Optional[str] = Field(None, description="New OCR job ID created")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "success": True,
                "receipt_id": 123,
                "job_id": "550e8400-e29b-41d4-a716-446655440001",
                "message": "Bon reîncarcat în procesare"
            }
        }


class BatchRetryResponse(BaseModel):
    """Response schema for batch retry endpoint."""

    success: bool = Field(..., description="Whether any retries were successful")
    batch_id: str = Field(..., description="Batch ID that was retried")
    retried_count: int = Field(..., description="Number of receipts successfully retried")
    failed_count: int = Field(..., description="Number of receipts that couldn't be retried")
    errors: List[str] = Field(default_factory=list, description="List of error messages")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "success": True,
                "batch_id": "abc-123",
                "retried_count": 3,
                "failed_count": 0,
                "errors": [],
                "message": "3 bonuri reîncarcate în procesare"
            }
        }


class CancelJobResponse(BaseModel):
    """Response schema for cancel job endpoint."""

    success: bool = Field(..., description="Whether the cancellation was successful")
    job_id: str = Field(..., description="Job ID that was cancelled")
    cancelled_at: datetime = Field(..., description="Timestamp when the job was cancelled")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "success": True,
                "job_id": "550e8400-e29b-41d4-a716-446655440001",
                "cancelled_at": "2025-01-11T15:30:00",
                "message": "Job anulat cu succes"
            }
        }


class CancelBatchResponse(BaseModel):
    """Response schema for cancel batch endpoint."""

    success: bool = Field(..., description="Whether any jobs were cancelled")
    batch_id: int = Field(..., description="Batch ID that was cancelled")
    cancelled_count: int = Field(..., description="Number of jobs successfully cancelled")
    skipped_count: int = Field(..., description="Number of jobs skipped (completed/failed)")
    message: str = Field(..., description="Status message")

    class Config:
        json_schema_extra = {
            "example": {
                "success": True,
                "batch_id": 1,
                "cancelled_count": 3,
                "skipped_count": 2,
                "message": "3 job-uri anulate, 2 ignorate (deja procesate)"
            }
        }