maddy/OCR
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixed ui and template mapping issue

Browse Source
This commit is contained in:
Narayanan Madaswamy
2026-07-14 12:43:09 +05:30
parent fb0a78a405
commit 460a1c5c51
11 changed files with 383 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

148
docengine/app/api/v1/templates.py
View File

@@ -10,7 +10,13 @@ from app.core.database import get_db
from app.core.dependencies import CurrentUser
from app.core.logging_config import get_logger
from app.repositories.document_repository import DocumentRepository, TemplateMatchRepository
from app.repositories.template_repository import TemplateRepository
from app.repositories.template_repository import (
TemplateRepository,
DocumentRegionRepository,
DocumentCellRepository,
TableFormatRepository,
TableColumnRepository
)
from app.schemas.common import PaginatedResponse, SuccessResponse
from app.schemas.document import TemplateMatchRequest, TemplateMatchResponse
from app.schemas.template import (
@@ -18,6 +24,8 @@ from app.schemas.template import (
TemplateRenderRequest,
TemplateRenderResponse,
TemplateResponse,
TemplateCreateRequest,
TemplateMappingSaveRequest,
)
logger = get_logger(__name__)
@@ -25,6 +33,132 @@ logger = get_logger(__name__)
router = APIRouter(prefix="/templates", tags=["Templates"])
@router.post(
"",
response_model=dict,
summary="Create Template",
description="Create a new template.",
)
def create_template(
payload: TemplateCreateRequest,
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> dict:
"""Create a template."""
template_repo = TemplateRepository(db)
template = template_repo.create_template(
name=payload.template_name,
page_width=1000.0,
page_height=1000.0,
source_document_id=uuid.UUID(payload.source_document_id) if payload.source_document_id else None
)
cell_repo = DocumentCellRepository(db)
saved_fields = []
for index, field in enumerate(payload.fields):
cell = cell_repo.create_cell(
format_id=template.id,
page_number=1,
x=0.0,
y=0.0,
width=0.0,
height=0.0,
data_type=field.field_type,
field_name=field.field_label,
sequence=field.display_order or index,
is_dynamic=True
)
saved_fields.append({
"field_label": cell.field_name,
"field_type": cell.data_type
})
db.commit()
return {
"pk_template_id": str(template.id),
"template_name": template.name,
"fields": saved_fields
}
@router.post(
"/{template_id}/mappings/save",
response_model=SuccessResponse,
summary="Save Template Mappings",
description="Save the field mappings for a template.",
)
def save_mappings(
template_id: uuid.UUID,
payload: list[TemplateMappingSaveRequest],
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> SuccessResponse:
"""Save template mappings."""
template_repo = TemplateRepository(db)
template = template_repo.get_by_id(template_id)
if not template:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"Template '{template_id}' not found",
)
region_repo = DocumentRegionRepository(db)
cell_repo = DocumentCellRepository(db)
table_repo = TableFormatRepository(db)
table_col_repo = TableColumnRepository(db)
cells = cell_repo.get_template_cells(template_id)
cell_map = {c.field_name: c for c in cells}
table_format = None
# Save each mapped node
for mapping in payload:
cell = cell_map.get(mapping.field_name)
is_table_column = cell and cell.data_type == 'TABLE_COLUMN'
for node_idx, node in enumerate(mapping.mapped_nodes):
if is_table_column:
if not table_format:
table_format = table_repo.create_table_format(
format_id=template_id,
page_number=node.page_no or 1,
x=0.0, y=0.0, width=1000.0, height=1000.0,
rows=1, columns=10
)
table_col_repo.create_column(
table_format_id=table_format.id,
column_index=node_idx,
width=node.width,
header_text=mapping.field_name,
data_type="text"
)
region_repo.create_region(
format_id=template_id,
page_number=node.page_no or 1,
region_type="field_mapping",
x=node.x_coordinate,
y=node.y_coordinate,
width=node.width,
height=node.height,
content={
"field_name": mapping.field_name,
"text_value": node.text_value
}
)
db.commit()
# Generate fingerprint now that mappings are populated
from app.services.fingerprint_service import FingerprintService
FingerprintService(db).generate_fingerprint(template)
db.commit()
return SuccessResponse(message="Mappings saved successfully")
@router.get(
"",
response_model=PaginatedResponse[TemplateListResponse],
@@ -32,9 +166,9 @@ router = APIRouter(prefix="/templates", tags=["Templates"])
description="List all active templates with pagination.",
)
def list_templates(
current_user: CurrentUser,
page: int = Query(default=1, ge=1),
page_size: int = Query(default=20, ge=1, le=100),
current_user: CurrentUser = None,
db: Session = Depends(get_db),
) -> PaginatedResponse[TemplateListResponse]:
"""List all active templates."""
@@ -61,7 +195,7 @@ def list_templates(
)
def get_template(
template_id: uuid.UUID,
current_user: CurrentUser = None,
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> TemplateResponse:
"""Get a template by ID."""
@@ -83,7 +217,7 @@ def get_template(
)
def delete_template(
template_id: uuid.UUID,
current_user: CurrentUser = None,
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> SuccessResponse:
"""Soft-delete a template."""
@@ -109,7 +243,7 @@ def delete_template(
)
def match_template(
payload: TemplateMatchRequest,
current_user: CurrentUser = None,
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> list[TemplateMatchResponse]:
"""Match a document against existing templates."""
@@ -162,7 +296,7 @@ def match_template(
)
def render_template(
payload: TemplateRenderRequest,
current_user: CurrentUser = None,
current_user: CurrentUser,
db: Session = Depends(get_db),
) -> TemplateRenderResponse:
"""Render a template to PDF."""
@@ -199,8 +333,8 @@ def render_template(
)
def download_rendered_pdf(
template_id: uuid.UUID,
current_user: CurrentUser,
filename: str = Query(..., description="Filename of the rendered PDF"),
current_user: CurrentUser = None,
db: Session = Depends(get_db),
) -> FileResponse:
"""Download a rendered PDF."""

2
docengine/app/repositories/template_repository.py
View File

@@ -70,6 +70,7 @@ class TemplateRepository(BaseRepository[DocumentFormat]):
fingerprint: dict | None = None,
source_document_id: uuid.UUID | None = None,
created_by: uuid.UUID | None = None,
is_active: bool = True,
) -> DocumentFormat:
"""Create a new template."""
template = DocumentFormat(
@@ -85,6 +86,7 @@ class TemplateRepository(BaseRepository[DocumentFormat]):
fingerprint=fingerprint,
source_document_id=source_document_id,
created_by=created_by,
is_active=is_active,
)
return self.create(template)

29
docengine/app/schemas/template.py
View File

@@ -2,7 +2,7 @@ from __future__ import annotations
import uuid
from datetime import datetime
from typing import Any
from typing import Any, List, Optional, Union
from pydantic import Field
@@ -248,3 +248,30 @@ class TemplateRenderResponse(BaseSchema):
file_size: int
page_count: int
rendered_at: datetime
class TemplateFieldCreate(BaseSchema):
field_label: str
field_type: str = "text"
display_order: int = 0
required_flag: bool = False
class TemplateCreateRequest(BaseSchema):
template_name: str
source_document_id: Optional[str] = None
fields: List[TemplateFieldCreate] = Field(default_factory=list)
class MappingNodeRequest(BaseSchema):
pk_document_data_id: Union[int, str, None] = None
x_coordinate: float
y_coordinate: float
width: float
height: float
text_value: Optional[str] = None
page_width: Optional[float] = None
page_height: Optional[float] = None
page_no: Optional[int] = None
class TemplateMappingSaveRequest(BaseSchema):
field_name: str
mapped_nodes: List[MappingNodeRequest] = Field(default_factory=list)

7
docengine/app/services/document_service.py
View File

@@ -58,11 +58,7 @@ class DocumentProcessingService:
metadata["layout"] = layout_results
document.document_metadata = metadata
# Step 3: Generate template
self.db.refresh(document)
template = self.template_service.generate_template(document)
# Step 4: Update document status
# Step 3: Update document status
self.doc_repo.update_status(document_id, "completed")
self.db.commit()
@@ -70,7 +66,6 @@ class DocumentProcessingService:
"processing_completed",
document_id=str(document_id),
pages=document.page_count,
template_id=str(template.id),
)
return document

2
docengine/app/services/fingerprint_service.py
View File

@@ -187,7 +187,7 @@ class FingerprintService:
serialized = json.dumps(normalized, sort_keys=True, default=str)
return hashlib.sha256(serialized.encode()).hexdigest()
def _normalize_coordinates(self, data: dict[str, Any]) -> dict[str, Any]:
def _normalize_coordinates(self, data: Any) -> Any:
"""Normalize coordinates by rounding to reduce sensitivity to small variations."""
if isinstance(data, dict):
return {k: self._normalize_coordinates(v) for k, v in data.items()}

1
docengine/app/services/template_service.py
View File

@@ -80,6 +80,7 @@ class TemplateService:
description=f"Auto-generated template from {document.original_filename}",
source_document_id=document.id,
created_by=user_id,
is_active=False,
)
# Process each page