Spaces:
Running
Running
Create tools/form_filler_server.py
Browse files- tools/form_filler_server.py +150 -0
tools/form_filler_server.py
ADDED
|
@@ -0,0 +1,150 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import Dict, Any
|
| 2 |
+
import json
|
| 3 |
+
from docx import Document
|
| 4 |
+
from openpyxl import load_workbook
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
async def fill_form(template_path: str, data: Dict[str, Any]) -> dict:
|
| 8 |
+
"""
|
| 9 |
+
Auto-fill form template with extracted data
|
| 10 |
+
|
| 11 |
+
Args:
|
| 12 |
+
template_path: Path to form template (DOCX or XLSX)
|
| 13 |
+
data: Dictionary of field values
|
| 14 |
+
|
| 15 |
+
Returns:
|
| 16 |
+
Dict with filled form path and fields filled
|
| 17 |
+
"""
|
| 18 |
+
try:
|
| 19 |
+
file_ext = Path(template_path).suffix.lower()
|
| 20 |
+
output_path = template_path.replace(file_ext, f'_filled{file_ext}')
|
| 21 |
+
|
| 22 |
+
if file_ext == '.docx':
|
| 23 |
+
return await fill_docx_form(template_path, data, output_path)
|
| 24 |
+
elif file_ext in ['.xlsx', '.xls']:
|
| 25 |
+
return await fill_excel_form(template_path, data, output_path)
|
| 26 |
+
else:
|
| 27 |
+
# Generic text replacement
|
| 28 |
+
return await fill_text_form(template_path, data, output_path)
|
| 29 |
+
|
| 30 |
+
except Exception as e:
|
| 31 |
+
return {'error': str(e), 'success': False}
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
async def fill_docx_form(template_path: str, data: Dict[str, Any], output_path: str) -> dict:
|
| 35 |
+
"""Fill DOCX form template"""
|
| 36 |
+
try:
|
| 37 |
+
doc = Document(template_path)
|
| 38 |
+
fields_filled = []
|
| 39 |
+
|
| 40 |
+
# Replace in paragraphs
|
| 41 |
+
for paragraph in doc.paragraphs:
|
| 42 |
+
for key, value in data.items():
|
| 43 |
+
placeholder = f"{{{{{key}}}}}" # {{field_name}}
|
| 44 |
+
if placeholder in paragraph.text:
|
| 45 |
+
paragraph.text = paragraph.text.replace(placeholder, str(value))
|
| 46 |
+
fields_filled.append(key)
|
| 47 |
+
|
| 48 |
+
# Replace in tables
|
| 49 |
+
for table in doc.tables:
|
| 50 |
+
for row in table.rows:
|
| 51 |
+
for cell in row.cells:
|
| 52 |
+
for key, value in data.items():
|
| 53 |
+
placeholder = f"{{{{{key}}}}}"
|
| 54 |
+
if placeholder in cell.text:
|
| 55 |
+
cell.text = cell.text.replace(placeholder, str(value))
|
| 56 |
+
if key not in fields_filled:
|
| 57 |
+
fields_filled.append(key)
|
| 58 |
+
|
| 59 |
+
doc.save(output_path)
|
| 60 |
+
|
| 61 |
+
return {
|
| 62 |
+
'success': True,
|
| 63 |
+
'output_path': output_path,
|
| 64 |
+
'fields_filled': fields_filled,
|
| 65 |
+
'total_fields': len(fields_filled)
|
| 66 |
+
}
|
| 67 |
+
|
| 68 |
+
except Exception as e:
|
| 69 |
+
return {'error': str(e), 'success': False}
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
async def fill_excel_form(template_path: str, data: Dict[str, Any], output_path: str) -> dict:
|
| 73 |
+
"""Fill Excel form template"""
|
| 74 |
+
try:
|
| 75 |
+
wb = load_workbook(template_path)
|
| 76 |
+
ws = wb.active
|
| 77 |
+
fields_filled = []
|
| 78 |
+
|
| 79 |
+
# Search for placeholders and replace
|
| 80 |
+
for row in ws.iter_rows():
|
| 81 |
+
for cell in row:
|
| 82 |
+
if cell.value and isinstance(cell.value, str):
|
| 83 |
+
for key, value in data.items():
|
| 84 |
+
placeholder = f"{{{{{key}}}}}"
|
| 85 |
+
if placeholder in cell.value:
|
| 86 |
+
cell.value = cell.value.replace(placeholder, str(value))
|
| 87 |
+
fields_filled.append(key)
|
| 88 |
+
|
| 89 |
+
wb.save(output_path)
|
| 90 |
+
|
| 91 |
+
return {
|
| 92 |
+
'success': True,
|
| 93 |
+
'output_path': output_path,
|
| 94 |
+
'fields_filled': list(set(fields_filled)),
|
| 95 |
+
'total_fields': len(set(fields_filled))
|
| 96 |
+
}
|
| 97 |
+
|
| 98 |
+
except Exception as e:
|
| 99 |
+
return {'error': str(e), 'success': False}
|
| 100 |
+
|
| 101 |
+
|
| 102 |
+
async def fill_text_form(template_path: str, data: Dict[str, Any], output_path: str) -> dict:
|
| 103 |
+
"""Fill text-based form"""
|
| 104 |
+
try:
|
| 105 |
+
with open(template_path, 'r', encoding='utf-8') as f:
|
| 106 |
+
content = f.read()
|
| 107 |
+
|
| 108 |
+
fields_filled = []
|
| 109 |
+
for key, value in data.items():
|
| 110 |
+
placeholder = f"{{{{{key}}}}}"
|
| 111 |
+
if placeholder in content:
|
| 112 |
+
content = content.replace(placeholder, str(value))
|
| 113 |
+
fields_filled.append(key)
|
| 114 |
+
|
| 115 |
+
with open(output_path, 'w', encoding='utf-8') as f:
|
| 116 |
+
f.write(content)
|
| 117 |
+
|
| 118 |
+
return {
|
| 119 |
+
'success': True,
|
| 120 |
+
'output_path': output_path,
|
| 121 |
+
'fields_filled': fields_filled,
|
| 122 |
+
'total_fields': len(fields_filled)
|
| 123 |
+
}
|
| 124 |
+
|
| 125 |
+
except Exception as e:
|
| 126 |
+
return {'error': str(e), 'success': False}
|
| 127 |
+
|
| 128 |
+
|
| 129 |
+
async def generate_form_template(fields: list) -> dict:
|
| 130 |
+
"""Generate a blank form template with specified fields"""
|
| 131 |
+
try:
|
| 132 |
+
doc = Document()
|
| 133 |
+
doc.add_heading('Form Template', 0)
|
| 134 |
+
|
| 135 |
+
for field in fields:
|
| 136 |
+
p = doc.add_paragraph()
|
| 137 |
+
p.add_run(f"{field}: ").bold = True
|
| 138 |
+
p.add_run(f"{{{{{field}}}}}")
|
| 139 |
+
|
| 140 |
+
output_path = "data/outputs/form_template.docx"
|
| 141 |
+
doc.save(output_path)
|
| 142 |
+
|
| 143 |
+
return {
|
| 144 |
+
'success': True,
|
| 145 |
+
'output_path': output_path,
|
| 146 |
+
'fields': fields
|
| 147 |
+
}
|
| 148 |
+
|
| 149 |
+
except Exception as e:
|
| 150 |
+
return {'error': str(e), 'success': False}
|